Relaxation and vibrational fluorescence of WO in solid neon

Article abstract of DOI:10.1016/S0009-2614(98)00605-8

Studies of laser-induced fluorescence of tungsten oxides in solid neon at 6 K provide interesting insights into the mechanisms of WO relaxation. The relaxation occurs by interelectronic cascade between the vibrational manifolds of excited electronic state

Full text of DOI:10.1016/S0009-2614(98)00605-8

17 July 1998
Ž
.
Chemical Physics Letters 291 1998 291–299
Relaxation and vibrational fluorescence of WO in solid neon
)
Martin Lorenz, Jurgen Agreiter, Nicola Caspary, Vladimir E. Bondybey
¨
Institut fur Physikalische und Theoretische Chemie, Technische UniÕersitat Munchen, Lichtenbergstraße 4, D-85748 Garching, Germany
¨
¨
¨
Received 28 April 1998
Abstract
Studies of laser-induced fluorescence of tungsten oxides in solid neon at 6 K provide interesting insights into the
mechanisms of WO relaxation. The relaxation occurs by interelectronic cascade between the vibrational manifolds of excited
electronic states, eventually reaching the ÕYs7 of the ground electronic state. Remarkably, for the heavy WO oxide, the
non-radiative processes are surprisingly inefficient, with relaxation in the ground state apparently occurring predominantly
by sequential vibrational fluorescence. The relaxation pathways of WO are discussed and compared with relaxation in
lighter, first-row molecules exemplified by CN. q 1998 Elsevier Science B.V. All rights reserved.
1. Introduction
ation and for which intense vibrational fluorescence
w
x
is observed 11 .
Low-temperature rare-gas solids are a useful
medium for studies of molecular relaxation and
Relaxation for molecules involving heavier ele-
ments is usually efficient, and even in electronic
spectra one mostly sees only vibrationally relaxed
fluorescence or phosphorescence. Particularly in
transition metals, with their dense manifold of elec-
tronic states, fast relaxation is usually the rule and
only in relatively few cases extensive ‘hot’ fluores-
cence is detected. We have recently examined the
spectroscopy and photophysics of the oxides of the
w
x
non-radiative processes 1–4 . While relaxation in
the condensed phase is usually fast, often proceeding
on a picosecond or even subpicosecond timescale, at
low temperatures in solid rare gases it often slows
down remarkably. In small molecules often vibra-
tionally unrelaxed ‘hot’ electronic fluorescence is
observed and in particular in the ground states of
light molecules the non-radiative relaxation may be
slow. In a number of light molecules, mostly involv-
ing hydrides or simple compounds of first-row ele-
ments, vibrational fluorescence was also observed
w
x
third-row transition metal tungsten 12 . Rather sur-
prisingly we have observed that, in spite of the
essential differences between the two molecules, the
energy relaxation in WO parallels the processes pre-
w
x
w
x
and investigated 5–10 . The diatomic CN is an
example of a molecule which exhibits a slow relax-
viously detected in CN 13,14 . Extensive vibra-
tionally unrelaxed emission from several electronic
states is observed and even more surprisingly, quite
intense ground-state vibrational fluorescence is eas-
ily detectable. Our observations of the WO relax-
ation behavior are the subject of this Letter.
)
Corresponding author.
E-mail: bondybey@verona.phys.chemie.tu-muenchen.de
0009-2614r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 98 00605-8

(
)
292
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
2. Experimental
current interest and we have therefore decided to
re-examine the WO spectroscopy using absorption
and fluorescence techniques, in order to gain addi-
tional spectral information and characterize possible
transitions in the infrared.
Tungsten and its oxides are high-melting refrac-
tory materials and their vaporization requires temper-
atures ranging up to 3000 K. Here we employ a
technique developed in our laboratory more than a
The solid neon matrices, generated by laser vapor-
izing tungsten in the presence of a small amount of
oxygen, contained diatomic tungsten oxide, as evi-
denced by infrared absorptions. Not only the oxygen,
but also the tungsten isotopes are clearly resolved.
Thus absorptions due to 184 W16O and 184 W18 O oc-
cur at 1056.49 and 1001.25 cmy1, respectively. In
agreement with Weltner and McLeod, we observe in
the visible the states he denotes A–F, as well as
several new states. In addition, in the near infrared
we have detected a group of states. Several progres-
sions occur with origins between 7403.8 and 8007.6
cmy1, with by far the strongest n00 absorption oc-
w
x
decade ago 15 ; laser vaporization of the metal in
the presence of a large excess of cold carrier gas,
neon gas being used to grow the matrix. The fre-
Ž
.
quency-doubled pulses of an Nd-YAG laser 30 mJ
used to vaporize a tungsten target are synchronized
with the opening of a fast valve admitting pulses of
the carrier gas. The valve, with opening times of 0.1
ms, is operated with a backing pressure of 5 bar of a
18
Ž
.
500:1 mixture of neon with O2 or O2 . The neon
gas with the products and oxides formed in the
vaporization fixture are then condensed on a 6 K
silver-coated copper substrate.
18
y1
curring at 7527.76 cmy1 7529.08 cm for W O .
While some of the weaker bands appearing here are
due to WO in ‘perturbed’ matrix sites, there are
clearly at least three absorbing states present in this
region. They may be components of the 7P state
predicted by Nelin and Bauschlicher. We will report
Ž
.
The deposits are examined spectroscopically with
the help of a Bruker IFS 120 HR FT instrument with
a spectral range from 500 to 30000 cmy1. Spectra
have been recorded with resolutions of 0.5 cmy1 and
0.06 cmy1 IR . To excite the sample fluorescence,
either an argon ion laser pumped dye laser, or a
Ti-Sapphire laser were used. A PMT detector was
used in the visible and UV range, while germanium
and indium antimonide detectors were available for
studies in the near infrared range.
Ž
.
Table 1
Molecular constants of 184 W16O in solid neon
Statea
Te
ve
ve xe
y1
y1
y1
Ž
cm
.
Ž
cm
.
Ž
.
cm
X5P
0.0
1064.49
957.9
959.6
958.7
956.1
1022.3
970.5
1084.3
1051.2
929.5
932.6
996.8
998.7
971.3
955.4
934.2
931.6
3.998b
3.27
3.76
3.55
3.26
c
Ž7
.
a
a
b
c
P?, red site
P?
P?
P?
7454.3
7578.6
7761.1
8062.1
3. Results and discussion
Ž7
Ž7
Ž7
.
.
.
c
c
c
3.1. WO electronic structure
dc
ec
fc
A
B
10938.7
12075.4
14132.7
14116.9
17336.9
19267.5
20799.8
20821.4
21538.2
23439.4
23852.7
27037.3
Tungsten oxides were studied in solid rare gases
more than 30 years ago by Weltner and McLeod
Ž5
.
S or 7 S?
10.84
y1.07
2.76
8.22
8.51
8.49
9.52
y5.1
13.15
w
x
16 , who observed a number of WO electronic states
in the visible region, and some of these states were
C
Ž
.
D red site
some 10 years later also investigated by Samoilova
D
E
F
w
x
et al. in the gas phase 17 . Nelin and Bauschlicher
have more recently investigated WO by ab initio
techniques, which suggested that the ground state is
G
5
formally a P state and predicted the excited 5S and
HXc,d
7 S states in the visible region. Their calculations
yielded another, low lying 7P state around 7400
a
w
x
Symmetries as proposed by Nelin and Bauschlicher 18 .
b
ve ye sy0.0034 cmy1 and for the X5P ground state.
cmy1 in the near infrared 18 . The electronic struc-
ture of transition metals and the ability of ab initio
methods to predict it is a topic of considerable
w
x
c Previously not observed.
d
Not identical with the H-state T00 s30446 cmy1 of Samoilova
Ž
.
w
x
et al. 17 .

(
)
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
293
others were conspicuously absent, indicative of a
complex isotopically dependent pattern of relaxation.
This relaxation process is quite complex and in-
volves vibrational manifolds of the numerous elec-
tronic states present in this region, both those ob-
served in absorption, and probably of other, ‘dark’
states. Rates of such interelectronic relaxation pro-
cesses are known to depend sensitively upon the
quantum mechanical coupling between the different
electronic states, but also on the ‘energy gap’ be-
tween the relaxing levels and are often governed by
the presence of accidental resonances or near reso-
nances.
When the lower energy levels at the red end of
the spectrum reported by Weltner and McLeod are
excited, the fluorescence is relatively simple and
easy to understand. Thus exciting the sample in the
Õs0 level of the W18 O in the B-state of Weltner
and McLeod at 17273.9 cmy1, one observes emis-
sion from the excited level into ÕY s0, 1 and 2
levels of the ground electronic state. A further pro-
gression originating from a level at 16980.4 cmy1
into the ÕY s1–4 ground-state levels can also easily
be identified as being due to the ÕX s3 A-state level,
293.5 cmy1 lower in energy. The 16980.44 cmy1 is
also seen weakly in absorption, although not previ-
ously observed by Weltner and McLeod. Excitation
Fig. 1. W16O potential curves and vibrational levels in the range
of 0–22000 cmy1. The res are obtained by Franck–Condon
w
x
analysis or are taken, where available, from Samoilova et al. 17 .
the details of our spectroscopic observations else-
w
x
where 12 and the present Letter will focus mainly
on the WO electronic and vibrational relaxation. We
summarize the information about the observed elec-
tronic states and their molecular constants necessary
for discussing the relaxation processes in Table 1
and show the energy level structure of WO schemati-
cally in Fig. 1.
of the corresponding ÕX s0 B-state level of W16O
y1
Ž
.
17280.5 cm , on the other hand, produces almost
no B-state emission, but mainly weak emissions
from the ÕX s3 and 2 A-state levels. Owing to the
isotopic shift, in the W16O the ÕX s3 A-level is only
150 cmy1 below the ÕX s0 B-level and the smaller
energy gap results in a more efficient depopulation
of the B-state.
3.2. Fluorescence and relaxation of the excited WO
electronic states in the Õisible
As one proceeds towards higher energies, the
fluorescence gets progressively more complex. For
X
y1
Ž
.
Fluorescence spectra can often provide a valuable
complementary contribution to absorption studies.
As explained in the experimental section, we have
used our tunable laser sources to excite the neon
matrix absorptions due to WO. Such an excitation at
a variety of frequencies ranging from about 10000 to
25000 cmy1 resulted in a fairly strong visible and
near infrared fluorescence, whose intensity and spec-
tral distribution depended sensitively upon the exci-
tation wavelength and upon the isotopic species ex-
cited. While some levels were emitting efficiently,
instance, excitation of the Õ s2 D 22635.9 cm
,
produces not only a weak resonant emission from the
excited level, but extensive relaxed fluorescence from
a number of other levels. No emission is seen from
ÕY s1 D, but several bands originating from ÕX s0
y1
Ž
.
E 21493.0 cm
and a fairly strong progression
X
y1
Ž
.
.
from Õ s0 D 20789.3 cm
and a weaker one
are seen. Another
X
y1
Ž
from Õ s0 B 17274.1 cm
rather long progression originates from the A ÕX s6
y1
Ž
.
and extends down to 11867.9
level 19675.5 cm
cmy1 into the ÕY s8 ground state. In addition to the

(
)
294
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
clearly identifiable progressions, numerous very
weak, sharp bands are observed confirming that in
addition to states previously detected by Weltner and
McLeod, several other ‘dark’ or very weakly absorb-
ing states must be present in this region and partici-
pate in the relaxation process.
the emission appearance is essentially independent of
which of the vibronic bands of the visible or near IR
progressions is excited by the laser and from the
multiple ‘systems’ detected in the absorption spec-
trum, only the strongest one with its origin at 7527.6
cmy1 appears. As already noted above, it may be
due to the ‘spin forbidden’ phosphorescence from
It may be noted that some weak bands appearing
in the absorption spectrum are due to ‘site effects’.
The bands due to ‘sites’ are, however, easily identi-
fied, since unlike excitation of WO in the ‘main’
site, their excitation results in spectrally shifted emis-
sion bands. Thus all the bands of the D progression
exhibit a much weaker satellite some 24 cmy1 lower
in energy, whose excitation produces fluorescence
7
the P state, although L is probably not well de-
fined in a molecule as heavy as WO. In contrast with
the visible fluorescence, the near infrared emission is
relatively simple and easy to analyse to yield the
molecular constants of the ‘a’-state. Interestingly, the
infrared spectrum is at least one order of magnitude
more intense in the heavier W18 O isotopic molecule,
which appears with appreciable intensity even in the
natural isotopic abundance samples.
bands uniformly shifted by the same 24 cmy1
.
3.3. Infrared phosphorescence
The spectra show that vibrational relaxation of
WO in neon is relatively slow even in the electroni-
cally excited ‘a’-state, since not only ÕX s0, but also
emission originating from vibrationally unrelaxed
levels is observed. In the case of the stronger W18 O
fluorescence, levels up to ÕX s3 are detected in
emission and the spectra can be followed with an
InSb detector down to about 3300 cmy1 in the
infrared, with the progressions in the ground-state
When the visible bands are excited, one observes
not only visible fluorescence, but also a strong emis-
sion in the near infrared range, clearly due to the
band system also observed in absorption. Representa-
tive spectra obtained with the Ge detector by tuning
the exciting laser to the B–X 0–0 transition at 17281
and 17274 cmy1 for W16O and W18 O, respectively,
are shown in Fig. 2. Unlike the visible fluorescence,
Fig. 2. NIR absorption and phosphorescence of tungsten oxide. The upper trace shows progressions of several NIR states, with the strongest
5
Ž7
.
absorption, denoted a P? §X P, occurring at T00 s7527.8 cmy1. The lower two traces display the phosphorescence originating from
Ž7
the ‘a’ P? state of both tungsten oxides, obtained after excitation of the B Õs0 state.
.

(
)
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
295
vibrational frequency extending to ÕY s5. The relax-
ation of the lighter W16O is apparently more efficient
and the corresponding emission appreciably weaker.
In this case, besides the vibrationless level only
ÕX s1 emission is observed.
The emission bands are readily identified as vibra-
tional overtone fluorescence of diatomic WO, with
the eight strong bands being the 9–7, 8–6, 7–5, 6–4,
5–3, 4–2, 3–1 and 2–0 transitions and in addition
also the very weak 10–8 bands are detected. This
assignment is quite unambiguous and is further con-
firmed by the clearly resolved tungsten isotopic
structure. In an experiment with 18-oxygen, the ap-
propriately shifted 7–5, 6–4, . . . , 2–0 bands are
observed. A careful examination of the region around
3000 cmy1 revealed also the DÕs3 sequence. Fi-
nally, as one can see in Fig. 4, the DÕs1 transitions
could easily be observed between 1000 and 1100
cmy1 using a mercury–cadmium telluride detector.
As noted above, the visible bands of WO exhibit,
in agreement with Weltner and McLeod, strong in-
terstate interactions and perturbations, so that only
the W18 O spectrum can be reasonably fitted to obtain
the spectroscopic constants of the excited states.
However, employing the appropriate isotopic rela-
tionships, the ground-state vibrational data including
some 80 transitions for eight different isotopic species
of WO and involving levels up to ÕY s9, can be
almost perfectly fitted. A Morse potential with ve s
3.4. Vibrational fluorescence of WO
At the red end of the spectrum obtained with the
indium antimonide detector one can observe, starting
at around 3000 cmy1, a slowly rising background,
which reaches a maximum near 1900 cmy1 and then
drops to zero due to a cutoff of the InSb detector
Ž
.
see Fig. 3 . This broad maximum is clearly due to
the high energy end of the ‘black-body’ radiation
from the room temperature apparatus walls. Super-
imposed over this maximum one can see in the
experiment employing natural abundance oxygen
three strong absorptions at 2312.4, 2330.6 and 2347.4
cmy1, as well as a group of eight moderately strong
emission bands at 1991.5, 2007.8, 2024.1, 2040.4,
2056.6, 2072.7, 2088.8 and 2104.9 cmy1. The three
absorption bands are clearly due to three isotopic
species of carbon dioxide, C18 O2 , 16OC18 O and
C16O2 probably formed due to the oxygen reaction
with a minor carbon impurity in the tungsten target.
1064.65 cmy1 and ve xe s4.046 cmy1 for the
Ž
184
16
.
W O isotopomer reproduces the observed levels
Fig. 3. Vibrational overtone fluorescence of W16O. The main trace displays the DÕs2 and the weaker DÕs3 transitions of ground state
W16O in solid neon. Superposed by black-body radiation, the 9–7, 8–6, . . . , 2–0 bands are easily observed. The 8–5, 7–4, 6–3 and 5–2
transitions appear near 3000 cmy1, with about one order of magnitude weaker intensity. The insert shows the strong 8–6 transition with
tungsten isotopical resolution. Peaks denoted with an asterisk are probably due to impurities.

(
)
296
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
Fig. 4. Vibrational fluorescence of tungsten oxide in neon. These two graphs show the DÕs1 ground-state emission of W16O and W18
O
after excitation at B Õs0. The spectra, measured with a mercury cadmium telluride detector, are background corrected by subtracting the
dominating black-body radiation.
with a RMS deviation of -0.09 cmy1. While a
measurement of the vibrational lifetimes. Estimating
absolute emission quantum yields is always difficult,
but semiquantitative consideration of the observed
band intensities suggests that it is high and that for
every excited molecule several IR photons may be
emitted. Such a high quantum yield of vibrational
emission indicates that the non-radiative relaxation
of the WO ground-state vibrational levels must be
exceedingly slow. The fact that life times are long in
the WO case is, however, also demonstrated by the
fact that absorptions originating in excited vibra-
tional levels can easily be detected. An inefficient
energy transfer between the light neon atoms and the
much heavier WO molecule might be a contributing
factor. In many molecules where strong vibrational
fluorescence is observed, e.g. OH, NH, CH3F or
hydrogen halides, it has been demonstrated that the
relaxation involves a V–R process, with the rota-
tional modes of the guest molecule being the primary
y1
Ž
.
somewhat better fit RMS-0.03 cm
is obtained
with ve ye sy0.0034 cmy1 the other constants
change slightly to ve s1064.49 cmy1 and ve xe s
Ž
3.998 cmy1 , the inclusion of ve ze or higher terms
.
results in no further improvement.
3.5. X-state Õibrational relaxation of WO
The observation of ground-state vibrational fluo-
rescence of WO is rather surprising for at least two
reasons. In the first place, while one occasionally
observes vibrationally unrelaxed fluorescence in
electronic transitions of similar matrix isolated small
metal containing molecules, such observations in the
infrared are quite uncommon. In view of the much
lower oscillator strengths and longer radiative life-
times associated with vibrational transitions, vibra-
tional fluorescence is usually observed only for hy-
drides and light molecules involving first row atoms,
with high vibrational frequencies usually above
1500–2000 cmy1. In the second place, the ground-
state potential of WO is relatively harmonic and
therefore one might not expect strong overtone tran-
sitions.
w
x
acceptors of the vibrational energy 19 . The rota-
Ž
tional constants of WO are much smaller Be f0.415
cmy1 and the rotations probably contribute little,
.
since high rotational levels would have to be in-
volved.
The appearance of the overtones in the vibrational
spectrum will, of course, be aided by the n 4 depen-
dence of the coefficient of spontaneous emission,
Unfortunately, we are not equipped for the direct

(
)
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
297
which will favor the DÕs2 and 3 transitions by
factors of 16 and 81, respectively, and thus counter-
act the harmonic oscillator selection rules. Even
though the mechanical anharmonicity of WO is small,
the overtones can be favored by electric anharmonic-
ity. The WO molecule may be fairly ionic, but at
larger distances it dissociates into neutral atoms and
its dipole is surely not a linear function of r.
The observed relaxation processes, where absorp-
tion of a single visible photon is followed by emis-
sion of a series of IR photons seems thus far to be
unique. It may be noted that similar processes, that is
UV pumping, followed by an internal conversion and
ground-state vibrational fluorescence are thought to
be responsible for the so-called unidentified infrared
state in the CN or the ‘a’ WO state is reached, the
process slows down further and in both molecules
extensive vibrationally unrelaxed, ‘hot’ emission, is
w
x
observed 13 . In CN it was shown that rather than
proceeding directly down the vibrational manifold of
the A-state, further relaxation occurs exclusively by
repeated crossings between the vibrational levels of
the A2 P state and the nearest X2 S ground-state
levels. Similarly it is almost certain that also in the
Ž7
.
lowest a P? excited state of WO, the relaxation
proceeds by an interelectronic mechanism. In the
light CN molecule, the A2 P spin–orbit constant is
y1
Ž
.
relatively small Asy52.64 cm , with relax-
ation between the components being extremely fast
and the splitting has little effect upon the relaxation
in the matrix. While in WO the spin–orbit effects are
surely much larger and the components of the widely
separated ‘septet’ state could be involved in the
relaxation, it appears likely that here ground-state
levels are involved, too.
Ž
.
emission bands UIE , with polycyclic aromatic hy-
Ž
.
drocarbons PAH being the suggested carriers
20,21 . The search for emission of these species in a
w
x
neon matrix might be of interest. The excellent
signal to noise ratio and clearly high quantum yields
of infrared fluorescence observed in WO raise the
question if such a behavior might not be much more
widespread, and if studies of vibrational emission
could not provide a useful general tool for charac-
terizing small metal oxides, carbides or nitrides.
After the relaxation reaches the vibrationless ÕX s
0 level of the lowest electronically excited state, only
one intramolecular non-radiative process is possible:
crossing into one of the nearby ground-state levels.
From that point on, the interelectronic channel is
closed and the molecule can only relax intrastate,
either non-radiatively, or by emitting infrared pho-
tons. Thus in the case of CN, the ÕY s4 X2 S
ground-state level is populated from ÕX s0 A2 P.
Further relaxation proceeds by sequential infrared
fluorescence and the ground-state ÕY s1–4 levels
appear strongly in emission. In W18 O the vibration-
3.6. Pathways of relaxation of WO in solid neon
One of the molecules for which strong infrared
emission has been detected, and whose relaxation in
matrices has been most extensively investigated is
probably the CN radical. Even though the properties
and electronic structures of the two molecules could
hardly be much more different and while WO has a
vibrational frequency about half of that of CN, the
relaxation pathways of the two molecules in the
matrix are remarkably similar. In the higher energy
region, the relaxation proceeds in both molecules by
an interelectronic process involving the vibrational
manifolds of several excited electronic states. Thus
Ž7
.
X
y1
less Õ s0 a P? level occurs 7529.39 cm above
the X ÕY s0 level, with the nearest ground-state
vibrational level being at 6857.7 cmy1, 671.69 cmy1
lower in energy. Quite analogous to the CN relax-
ation, as discussed above, the highest level from
which vibrational infrared fluorescence is observed
is ÕY s7. Clearly also here further relaxation appar-
ently occurs by sequential vibrational fluorescence.
As noted above, in contrast with the heavier
isotope, in W16O the a-state emission is weaker and
conversely, the vibrational emission considerably
more efficient. An explanation is easily found by
examining the vibrational level structures of the two
oxygen isotopic species, which also provides further
as in CN, the excitation zig-zags between the B2 S
2
w
x
and A P electronic states 14 , also WO cris-crosses
in the visible between the several electronic states
available at that energy. The non-radiative interelec-
tronic transitions are dominated by accidental level
degeneracies and the rates are governed by Franck–
Condon factors and the energy gap law.
Ž7
.
support for the relaxation in the lowest a P? state
When the lowest excited electronic state, the A2 P
proceeding, as in CN, via the ground-state vibra-

(
)
298
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
tional levels. While in W18 O, the ÕX s0–3 are lo-
cated 671, 623, 576 and 526 cmy1 above the closest
ground-state level, for W16O the numbers are 300,
changes much more drastically when an electronic
transition takes place, for instance from a S- to a
P-state. This again is usually evidenced by the much
more prominent phonon sidebands associated with
electronic transitions. Thus in the optical A2 Pl
X2 S transition in matrices, rather strong phonon
sidebands are observable. In the case of WO, the
phonon wings are weaker, but clearly observable.
These same Franck Condon factors which channel
energy into the lattice phonons during an optical
transition between the two electronic states also fa-
cilitate the corresponding non-radiative relaxation
process.
252, 205 and 158 cmy1 see Fig. 1 . The smaller
energy gaps in W16O result in a much more efficient
relaxation into the ground-state vibrational manifold,
while conversely the larger gaps in W18 O lead to
enhanced phosphorescence, bypassing the excited
ground-state levels.
Ž
.
The involvement of the ground state in a-state
relaxation is then directly evidenced by the observa-
Ž7
.
X
tion of emission from levels above Õ s0 a P? .
As noted above, also ÕX s8, 9 and very weakly also
ÕX s10 X-levels, located 779, 740 and 701 cmy1
above the ÕX s2, 1 and 0 a-state levels, are seen in
emission in W16O. These relatively large ‘energy
gaps’, combined with poor Franck–Condon factors,
make the reverse transition into the a-state once the
crossing into the X-state has occurred inefficient and
vibrational emission becomes competitive. On the
other hand, in the heavier isotope the corresponding
gaps are only 360, 320 and 280 cmy1, respectively.
This makes the reverse crossing into the a-state more
efficient compared with direct sequential radiative
vibrational relaxation within the ground state.
4. Conclusion
Pathways of relaxation of electronically excited
WO molecules in solid neon at 6 K were investi-
gated. In spite of the much lower vibrational fre-
quency and entirely dissimilar molecular properties
and structure, the pattern of relaxation processes
resembles those found in the much lighter CN radi-
cal. The relaxation occurs by interelectronic cascade
between the vibrational manifolds of excited elec-
tronic states, reaching eventually the ÕY s7 of the
ground electronic state. The non-radiative processes
in the ground state are surprisingly inefficient, with
relaxation occurring predominantly by sequential vi-
brational fluorescence.
3.7. Intrastate Õs. interstate Õibrational relaxation
As described above, the relaxation in the quite
dissimilar molecules CN and WO proceeds in both
cases via the same interelectronic mechanism, and
similar observations have been made for a variety of
other systems. In fact, even when the energy gaps for
the two processes are comparable, the interstate pro-
cess is in general much more efficient. As explained
previously, the physical reasons are to be found in
Franck–Condon arguments applied to the local
phonon modes and to the response of the lattice to
the change of state of the guest. The nearby lattice
atoms experience little change when the guest
molecule changes its vibrational state. This is evi-
denced by the vibrational spectra of the guest
molecules, which mostly exhibit only strong ‘zero
phonon lines’ and little evidence of phonon side-
bands.
Acknowledgements
Financial support of this research by the Deutsche
Forschungsgemeinschaft and the Fonds der Chemis-
chen Industrie is gratefully acknowledged. The au-
Ž
thors are also indebted to NATO Grant No. CRG
.
960009 for additional support.
References
w x
1
Ž
.
M. Berkowitz, R.B. Gerber, Chem. Phys. Lett. 49 1977
260.
On the other hand, the electron density distribu-
tion in the neighborhood of the guest molecule
w x
Ž .
2
w x
3
F. Legay, Chem. Biochem. Appl. Lasers 2 1977 43–86.
Ž .
H. Kono, S.H. Lin, J. Chem. Phys. 79 1983 2748.

(
)
M. Lorenz et al.rChemical Physics Letters 291 1998 291–299
299
w x
Ž
.
w
x Ž .
13 V.E. Bondybey, A. Nitzan, Phys. Rev. Lett. 38 1977 889.
4
w x
5
V.E. Bondybey, L.E. Brus, Adv. Chem. Phys. 41 1980 269.
Ž
.
w
x
14 D.H. Katayama, T.A. Miller, V.E. Bondybey, J. Chem. Phys.
P.F. Zittel, C.B. Moore, J. Chem. Phys. 59 1973 6636.
w x
Ž
.
Ž .
71 1979 1662.
6
V.E. Bondybey, L.E. Brus, J. Chem. Phys. 63 1975 794.
C. Blindauer, M. Winter, O. Sild, G. Jansen, B.A. Hess, U.
w x
w
x
Ž
.
7
15 V.E. Bondybey, J.H. English, J. Chem. Phys. 74 1981
Ž
.
Schurath, J. Phys. Chem. 97 1993 10002.
6978.
w x
w
x Ž .
16 W. Weltner, D. McLeod, J. Mol. Spectrosc. 17 1965 276.
8
N. Caspary, B.E. Wurfel, A.M. Smith, V.E. Bondybey,
Ž
.
w
x
17 A.N. Samoilova, Y.M. Efremov, L.V. Gurvich, J. Mol. Spec-
Chem. Phys. 220 1997 241.
w x
Ž
.
Ž .
trosc. 86 1981 1–15.
9
H. Dubost, R. Charneau, Chem. Phys. 12 1976 407.
w
w
x
w
x
Ž
.
10 L. Abouaf-Marguin, B. Gauthier-Roy, J. Dupre, C. Meyer, J.
18 C.F. Nelin, C.W. Bauschlicher, Chem. Phys. Lett. 118 1985
Ž
.
Mol. Spectrosc. 110 1985 347.
221.
x
w
w
w
x
Ž
.
11 B.E. Wurfel, G. Schallmoser, G.M. Lask, J. Agreiter, A.
19 L. Young, C.B. Moore, J. Chem. Phys. 81 1984 3137.
x
Ž
.
Thoma, R. Schlachta, V.E. Bondybey, Chem. Phys. 174
20 A. Leger, J.L. Puget, Astron. Astrophys. 137 1984 L5.
Ž
.
x
1993 255.
21 L.J. Allamandola, A.G. Tielens, J.R. Barker, Astrophys. J.
w
x
Ž .
290 1985 L25.
12 M. Lorenz, V.E. Bondybey, manuscript in preparation.