Kinetic study of tungsten atoms (a 7S3 and a 5DJ) in the presence of C2H4 and NH3 at room temperature

Article abstract of DOI:10.1246/bcsj.76.1145

The gas-phase reactivity of W (a ⁷S₃ and a ⁵D_J) with C₂H₄ and NH₃ at room temperature was investigated using a time-resolved laser induced fluorescence (LIF) spectroscopy. Tungsten atoms were produced by a 266-nm multiphoton decomposition (MPD) of W(CO)₆. The reactant pressure dependence of the pseudo-first-order depletion rates of W (a ⁷S₃) could allow an estimation of the pseudo-second-order depletion rate constant of W (a ⁷S₃), (4.5 ± 0.5) × 10^-10 cm³ molecule^-1 s^-1 for C₂H₄ and (0.73 ± 0.10) × 10^-10 for NH₃ at 6.0-Torr total pressure with an Ar buffer. A simulation of the transient curves based on a modification of the observed apparent decay rate constants, involving nearby a ⁵D_J states in the presence of C₂H₄ and NH₃, allowed us to separately estimate the contribution of the chemical quenching (W (a ⁷S₃) + R → product(s)) and the physical quenching (W (a ⁷S₃) + R → W (a ⁵D₁) + R) processes. In the case of C₂H₄, chemical quenching appeared to dominate over the physical quenching, while the physical quenching was the main depletion process in the case of NH₃. The large reactivity of the W (a ⁷S₃) state not only for C₂H₄, but also for NH₃, is discussed in terms of the relativistic effects.

Full text of DOI:10.1246/bcsj.76.1145

ꢀ 2003 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 76, 1145–1153 (2003) 1145
7
5
Kinetic Study of Tungsten Atoms (a S3 and a DJ) in the Presence of
C2H4 and NH3 at Room Temperature
ꢀ
Yo-ichi Ishikawa and Yoshitaka Matsumoto
Department of Chemistry and Materials Technology, Kyoto Institute of Technology,
Matsugasaki, Sakyo-ku, Kyoto 606-8585
(Received November 18, 2002)
The gas-phase reactivity of W (a 7S3 and a 5DJ) with C2H4 and NH3 at room temperature was investigated using a
time-resolved laser induced fluorescence (LIF) spectroscopy. Tungsten atoms were produced by a 266-nm multiphoton
decomposition (MPD) of W(CO)6. The reactant pressure dependence of the pseudo-first-order depletion rates of W (a
7S3) could allow an estimation of the pseudo-second-order depletion rate constant of W (a 7S3), ð4:5 ꢁ 0:5Þ ꢂ 10ꢃ10
cm3 moleculeꢃ1 sꢃ1 for C2H4 and ð0:73 ꢁ 0:10Þ ꢂ 10ꢃ10 for NH3 at 6.0-Torr total pressure with an Ar buffer. A simu-
lation of the transient curves based on a modification of the observed apparent decay rate constants, involving nearby a
5DJ states in the presence of C2H4 and NH3, allowed us to separately estimate the contribution of the chemical quench-
ing (W (a 7S3) + R ! product(s)) and the physical quenching (W (a 7S3) + R ! W (a 5D1) + R) processes. In the case
of C2H4, chemical quenching appeared to dominate over the physical quenching, while the physical quenching was the
7
main depletion process in the case of NH3. The large reactivity of the W (a S3) state not only for C2H4, but also for
NH3, is discussed in terms of the relativistic effects.
Many transition-metal atoms have a number of nearby states
derived from the electronic configurations of dnꢃ2 s2 and dnꢃ1
s1, which give a chance to investigate how the electronic con-
figuration affects the reactivity other than the energy. The gas-
phase transition-metal chemistry has so far provided funda-
mental information for metal-ligand interactions.1{3 Those ex-
perimental results would be useful for understanding the cata-
lytic processes on the active sites of transition-metal surfaces
and also for developing theoretical calculations involving a
spin-orbit interaction or a configuration interaction, and so
on.4;5
the s shells, from the inner up to valence shell, and the relativ-
istic expansion and destabilization of the d shell can be expect-
ed to have a certain effect on the chemical and physical inter-
actions of the transition-metal elements with the reactants. An
anomalous gas-phase chemical tendency of Au2 for CO, NH3,
and C2H4, which could not be extrapolated from the trends in
the reactivity from Cu2 to Ag2, was interpreted in terms of re-
lativistic effects.11 The large reactivity of W2 (X) with C2H4
and NH3, compared with Cr2 (X) and Mo2 (X), was also
thought to have originated from relativistic effects.12 W (a
7S3) atoms having naked 5d and 6s electrons would give a
chance to study the relativistic effects in chemical reactions
by a comparison of Cr (a 7S3) and Mo (a 7S3) atoms with
the same electron configuration.
In the gas-phase reactions of W (a 5DJ and a 7S3) with some
7
oxidants, the a S3 (5d5 6s1) state reacts more efficiently with
O2, NO, N2O, and SO2 than the a 5DJ (5d4 6s2) states,6;7 sug-
gesting that the electron configuration plays the dominant role
in the kinetic behavior of gas-phase tungsten. All of these oxi-
dation reactions are reported to be bimolecular, where the
common product is thought to be WO, owing to the indepen-
dence of the reaction rates with respect to the total pressure.
In this study, we measured the transient behavior of the low-
5
7
lying states of a tungsten atom W (a DJ and a S3) upon an
interaction with C2H4 as a typical ꢀ-acceptor, and NH3 as a
typical ꢁ-donor in order to understand the relation between
the electronic structure and the reaction mechanism. A low-ly-
ing energy level diagram of tungsten atom is shown in
5
5
7
The major depletion process for the a D1, a D0 and a S3
states was reported to be chemical quenching, though the con-
tribution of physical quenching could not be excluded com-
5
Fig. 1.13 The a D0 state with the d4 s2 configuration is the
5
ground state, and the other a DJ states (1 ꢄ J ꢄ 4) locate in
5
5
5
pletely for the a D2, a D3, and a D4 states. Because the
quenching rate constants of the a 5DJ states were found to
be characteristically J-dependent for these oxidants, except
for SO2, the oxidation mechanism has been thought to be di-
rect oxygen atom abstraction, but not the electron transfer
mechanism.8 The reactivity of W atoms for other reactants,
except for these oxidants, have not been reported so far.
Relativistic effects may become conspicuous in chemical re-
actions involving heavy elements, such as the third transition
metals.9;10 The relativistic contraction and stabilization of
an energy region lower than 6219 cmꢃ1. The a 7S3 state with
a different electron configuration of d5 s1 has only 2951 cmꢃ1
more energy than the ground state. However, in such a crowd-
ed system, the transient behavior of each state must be affected
by the physical quenching of nearby states as well as by the
chemical quenching, especially in an experimental system
where the multiphoton decomposition of metal carbonyl by fo-
cused laser light is used for the pulse production of transition
atoms. We have tried to discriminate the two processes, che-
mical and physical quenching, by calculation trials of all the

1146 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
Gas-Phase Kinetic Study of Tungsten Atoms
Fig. 1. Energy level diagram of W atom below 8000 cmꢃ1
.
Table 1. Excitation and Monitoring Transitions Used in LIF Spectroscopy
Center wavelength
of monochromator
(20 nm fwhm)
(ꢂ/nm)
Excitation
transition
(ꢂ/nm)
Monitoring
transition
(ꢂ/nm)
Monitoring level
7
5
7
7
z D3 a D4
z D3 ! a S3
5
a D4
430
383
383
384
384
430
(500.76)
(430.33)
5
5
5
5
z P2 a D3
z P2 ! a D2
5
a D3
(407.11)
(383.61)
5
5
5
5
z P2 a D2
z P2 ! a D2
5
a D2
(383.61)
(383.61)
5
5
5
5
z F2 a D1
z F2 ! a D1
5
a D1
(384.73)
(384.73)
5
5
5
5
z F1 a D0
z F1 ! a D0
5
a D0
(384.86)
(384.86)
7
7
7
7
z P2 a S3
z P2 ! a S3
7
a S3
(429.58)
(429.58)
transient figures of the related states (a 5DJ and a 7S3) with one
consistent set of rate constants obtained from a modification of
the apparent decay rates. In a laser vaporization/flow tube/LIF
rate synchronized with the photolysis light by a digital delay gen-
erator (SRS DG535). The electronic transition used to monitor
each state and the associated wavelength of the monochromator
for fluorescence observation are given in Table 1. The fluores-
cence was detected by a photomultiplier (Hamamatsu R-928) set
behind an exit slit of the monochromator (Spex Minimate;
f ¼ 20 cm, reciprocal dispersion = 4 nm/mm), the output of
which was subsequently sent to a digital oscilloscope (Tektronix
TD320). The digital output was stored and analyzed by a PC.
The typical average laser shots were 256 to obtain an LIF intensity
corresponding to W atom concentration at each delay time.
W(CO)6 (Aldrich, 98%) was degassed by several freeze-thaw cy-
cles at ꢃ10ꢅC and used without further purification. A sample
mixture of W(CO)6 (.10 mTorr) and Ar as a carrier gas was al-
lowed to flow slowly through a brass cell with two suprasil
windows. The flow quantity of each gas was controlled by a ca-
librated mass-flow controller (STEC-400MARK2 or 3: full scale,
10 sccm; accuracy, < ꢁ0:2 sccm) prior to admission to the reac-
tion cell. The mass-flow controller was used as purchased without
any further calibration. The total mass-flow rates were between
5–10 sccm, and the reactant (C2H4 and NH3) flow rates were
0.02–1.0 sccm. The total pressure was kept constant at 6.0 Torr
with an Ar buffer, which was monitored with a capacitance man-
ometer (MKS 220BH: full scale, 10 Torr; accuracy, < ꢁ0:015
7
experiment developed by Lian et al., Mo (a S3) atoms have
been reported to react with C2H4, but not to react with
NH3.14 Such a specific reactivity was discussed in terms of
the extent for reduction of the electronic repulsive interaction
as an early barrier. Another object of this study was to see
whether the trend of reactivity can persist in W atoms of the
same group f concerning relativistic effects.
Experimental
The concentration of each low-lying electronic state of tungsten
atom was pursued by using a time-resolved LIF technique in a
mass-flow-controlled cell at room temperature. Our time-resolved
LIF apparatus was very similar to a typical one.15;16 A photolysis
266-nm light source was a pulsed Nd:YAG laser (Spectron SL
803, ꢂ ¼ 266 nm). The laser beam, after passing through a Teflon
iris (6 mm in diameter), was focused onto the center of a reaction
cell using a quartz lens (f ¼ 15 cm). As a source of resonance
light for LIF, a Nd:YAG laser (Spectra-Physics GCR290-10)
pumped dye laser (Lumonics HD-500) with a dye of Stilbene
420 (2 ꢂ 10ꢃ4 M in CH3OH) was used with a 10 Hz repetition

Y. Ishikawa et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1147
Fig. 3. (a) NH3 addition effects on the transient LIF curves
7
Fig. 2. (a) C2H4 addition effects on the transient LIF curves
of W (a 7S3) at 6.0-Torr total pressure with Ar buffer gas.
C2H4 pressures are 0.0 ( ), 0.05 ( ), 0.1 ( ), 0.2 ( ),
and 0.3 Torr ( ). Each solid line is a single-exponential
decay fit obtained with an appropriate decay rate
constant. (b) Determination of the apparent second-order
depletion rate constant. Solid line is a linear fit.
of W (a S3) at 6.0-Torr total pressure with Ar buffer gas.
NH3 pressures are 0.0 ( ), 0.1 ( ), 0.2 ( ), 0.3 ( ), and
0.4 Torr ( ). Each solid line is a single-exponential decay
fit obtained with an appropriate decay rate constant. (b)
Determination of the apparent second-order depletion rate
constant. Solid line is a linear fit.
Torr) attached directly to the center of the flow cell.
fering from Cr and Mo, and the transient of this state must
be affected by physical quenching from/to nearby states. Thus,
the transient behaviors of some low-lying states of tungsten (a
Results
Typical depletion plots of W (a 7S3) in the presence of C2H4
at 6.0-Torr total pressure using an Ar buffer are shown in
Fig. 2(a), where approximate single-exponential decays were
observed. The solid lines through data are single-exponential
decay fits. These pseudo-first-order rates estimated from the
single-exponential fitting increased proportionally with in-
creasing the C2H4 pressure (Fig. 2(b)), giving an apparent sec-
ond-order depletion rate constant of ð4:5 ꢁ 0:5Þ ꢂ 10ꢃ10 cm3
5DJ) in addition to the a S3 state were also investigated as a
7
function of the reactant pressure (C2H4 or NH3) in order to dis-
tinguish between the chemical quenching and the physical
quenching to nearby states.
Figures 4 and 5 show the addition effects of C2H4 and NH3
on the transients of W (a 5DJ) together with W (a 7S3) at a total
pressure of 6.0 Torr with an Ar buffer. The apparent depletion
rate constants estimated from the additive concentration de-
pendence of the decay rates are summarized in the third col-
umn of Tables 2 and 3. Each solid line in the figures is a cal-
culated curve based on the assumption of an appropriate
mechanism involving physical and chemical quenching pro-
cesses, as mentioned below. Without any additives, except
for Ar buffer (6.0 Torr) and a trace of W(CO)6 (.10 mTorr)
moleculeꢃ1 ꢃ1. Similar plots obtained in the case of NH3 ad-
s
dition are shown in Fig. 3. The apparent second-order deple-
tion rate constant of ð0:73 ꢁ 0:10Þ ꢂ 10ꢃ10 cm3 moleculeꢃ1
sꢃ1 was estimated from the NH3 pressure dependence of the
pseudo-first order decay rates (Fig. 3(b)). Each error was
7
one statistical deviation (ꢁꢁ) in the linear fitting. W (a S3,
d5 s1) is more interactive with C2H4 than with NH3 in a similar
as precursor of W atoms, each electronic state (a DJ and a
5
7
manner as Mo (a S3, d5 s1).
7S3) shows a characteristic transient behavior, as shown by
the open square symbols in each figure. The apparent growth
in the a 5D4, a 5D3, a 5D1 and a 5D0 states indicates the occur-
rence of the physical quenching of the upper states correlating
to these states, probably by Ar. The lack of growth in the a
Because a multiphoton decomposition (MPD) of W(CO)6
was used for the pulse production of metal species in our ex-
periment, many electronic states other than the a 7S3 state were
7
inevitably produced. The a S3 state is metastable in W, dif-

1148 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
Gas-Phase Kinetic Study of Tungsten Atoms
Fig. 4. C2H4 addition effects on the transient LIF curves of W (a 7S3) (a) and W (a 5DJ ; J ¼ 4 (b), 3 (c), 2 (d), 1 (e), and 0 (f)) at
6.0-Torr total pressure with Ar buffer gas. C2H4 pressures are 0.0 ( ), 0.05 ( ), 0.1 ( ), 0.2 ( ), and 0.3 Torr ( ). Each solid
line is a calculation curve on the assumption of an appropriate mechanism involving physical and chemical quenching processes.
See discussion in the text.

Y. Ishikawa et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1149
Fig. 5. (a) NH3 addition effects on the transient LIF curves of W (a 7S3) (a) and W (a 5DJ ; J ¼ 4 (b), 3 (c), 2 (d), 1 (e), and 0 (f)) at
6.0-Torr total pressure with Ar buffer gas. NH3 pressures are 0.0 ( ), 0.1 ( ), 0.2 ( ), 0.3 ( ), and 0.4 Torr ( ). Each solid line
is a calculation curve on the assumption of an appropriate mechanism involving physical and chemical quenching processes. See
discussion in the text.

1150 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
Gas-Phase Kinetic Study of Tungsten Atoms
Table 2. Second-Order Rate Constants (kð2Þapparent, kð2Þphysical, and kð2Þchemical/10ꢃ10 cm3
5
7
moleculeꢃ1 ꢃ1) of W (a DJ and a S3) with C2H4 at Room Temperature
s
Level energy
/cmꢃ1
kð2Þ
kð2Þ
physical
chemical
b)
State (R.P.)a)
kð2Þ
apparent
ꢀ 2
ꢀ 2 b)
(ꢁ/A )
(ꢁ/A )
ꢆ0:013d)
0.40
(0.074)
5
a D4 (1.0)
6219.33
4830.00
0:42 ꢁ 0:04c)
0:28 ꢁ 0:02
(0.002)
ꢆ0:013
0.35
(0.065)
5
a D3 (1.5)
(0.002)
7
ꢆ0:05 (! a S3)
(0.009)
ꢆ0:013
5
a D2 (5.0)
3325.53
2951.29
0:11 ꢁ 0:02
4:5 ꢁ 0:5
5
(0.002)
0.20 (! a D1)
(0.037)
5
ꢆ0:05 (! a D2)
(0.009)
5.2
(0.96)
7
a S3 (9.0)
5
0.10 (! a D1)
(0.018)
ꢆ0:013
0.20
(0.037)
5
a D1 (8.0)
1670.30
0
0:14 ꢁ 0:01
(0.002)
0.013
(0.002)
5
a D0 (20.0)
0:013 ꢁ 0:002
—
a) Presumed relative initial population for kinetic simulation. b) The absolute cross section
1=2
ꢁ (¼ k=<ꢃ>); <ꢃ> ¼ ð8kT=ꢀꢄÞ ¼ 541 m sꢃ1. c) One standard deviation of the linear
fitting. d) A swung mark (ꢆ) means that the simulation is not sensitive to the value prob-
ably including ꢁ200–300% error.
Table 3. Second-Order Rate Constants (kð2Þapparent, kð2Þphysical, and kð2Þchemical/10ꢃ10 cm3
5
7
moleculeꢃ1 ꢃ1) of W (a DJ and a S3) with NH3 at Room Temperature
s
Level energy
/cmꢃ1
kð2Þ
kð2Þ
physical
chemical
ꢀ 2 b)
State (R.P.)a)
kð2Þ
apparent
ꢀ 2 b)
(ꢁ/A )
(ꢁ/A )
ꢆ0:0015d)
1.8
(0.27)
5
a D4 (1.0)
6219.33
4830.00
0:84 ꢁ 0:02c)
0:50 ꢁ 0:03
(0.0002)
ꢆ0:0015
1.6
(0.24)
5
a D3 (1.5)
(0.0002)
7
ꢆ0:30 (! a S3)
(0.045)
ꢆ0:0015
5
a D2 (5.0)
3325.53
2951.29
0:57 ꢁ 0:05
0:73 ꢁ 0:10
5
(0.0002)
0.90 (! a D1)
(0.14)
5
ꢆ0:20 (! a D2)
(0.045)
0.20
(0.03)
7
a S3 (9.0)
5
0.70 (! a D1)
(0.11)
ꢆ0:0015
0.52
(0.078)
5
a D1 (8.0)
1670.30
0
0:37 ꢁ 0:03
(0.0002)
0.0015
(0.0002)
5
a D0 (20.0)
0:0015 ꢁ 0:0004
—
a) Presumed relative initial population for kinetic simulation. b) The absolute cross section
1=2
ꢁ (¼ k=<ꢃ>); <ꢃ> ¼ ð8kT=ꢀꢄÞ ¼ 663 m sꢃ1. c) One standard deviation of the linear
fitting. d) A swung mark (ꢆ) means that the simulation is not sensitive to the value prob-
ably including ꢁ200–300% error.
7
5
5D2 state may be interpreted by a faster mixing with the a S3
the ground state (a D0) is chemically less reactive for both
5
state, owing to the small energy gap of 370 cmꢃ1, or the larger
initial concentration compared with the upper a 5D3 and a 5D4
states, so that the contribution of physical quenching to this
state may be obscured at a glance. The concentration of the
reactants. The absence of a fast production of the a D1 state
7
synchronized with the fast decay of the a S3 state by the ad-
dition of C2H4 suggests that the main depletion process of the
a 7S3 state is chemical quenching. Compared with the case of
C2H4 (Fig. 4), there seem to be significant growths in the a
5D3, a 5D1, and a 5D0 states by the addition of NH3 in
Fig. 5, suggesting that physical quenching is the main deple-
tion process for the upper states of these states.
5
ground state of a D0 increases until about 100 ms, and then
decreases over a fairly long time (ꢆms). Upon the addition
7
of reactants, the transient behaviors show that the a S3 state
5
is more interactive for C2H4 than any a DJ states, and that

Y. Ishikawa et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1151
physical quenching between the a 7S3 state and the a 5D2 state
was introduced to the calculation owing to its small energy gap
of 374 cmꢃ1. At first, the transients of these states (a 7S3 and a
5DJ) observed under no reactant addition were simulated. This
simulation fixed the relative initial concentrations of these six
states and some rate constants unrelated to the reactant. The
presumed initial concentration ratios are given in Tables 2
and 3, which are roughly similar to the Boltzmann distribution
at 1800 K ([5D4]:[5D3]:[5D2]:[7S3]:[5D1]:[5D0] = 1.0:2.4:5.7:
10.7:12.7:16.1) when the statistical weight (2J þ 1) is taken
into consideration. The difference in the initial concentrations
between two systems of C2H4 and NH3 might be due to the ex-
perimental scatter. Three states of Wx, Wy, and Wz had to be
introduced into the reaction mechanism in order to simulate
5
the rise behaviors of the corresponding three states: a D4, a
7
5D3, and a S3. The possible candidates for these Wx, Wy,
5
7
and Wz states are the upper states of a DJ or a S3, as shown
in Fig. 1. The initial relative concentrations of the Wx, Wy,
and Wz states were assumed to be 3.0, 1.5, and 2.0 when [a
5D4] at t ¼ 0 was set to 1.0. The residual parameters are
the chemical and physical quenching rate constants of the re-
5
5
5
7
5
actant with Wx, Wy, Wz, a D4, a D3, a D2, a S3, a D1,
and a 5D0 states, where the a 5D2 and a 7S3 states had two phy-
sical rate constants quenched to the a 7S3 and a 5D1 states and
to the a 5D2 and a 5D1 states, respectively. The disappearance
of W atoms due to an escape from an observation zone was
postulated to have the same rate constant kd (¼ 2:5 ꢂ 103
Fig. 6. An assumed quenching processes of W (a7S3, and
a5DJ) in the presence of C2H4 or NH3 for the calculation
of the transient concentrations of W (a7S3 and a5DJ). In
the kinetic schemes, kc’s are the chemical quenching con-
stants and kp’s are the physical quenching constants.
s
ꢃ1) for all states. Because there is not any physical quenching
process for the lowest state, a 5D0, the slope of the decay rates
vs reactant pressure plots was thought to give a chemical
quenching rate constant, kc ¼ ð1:3 ꢁ 0:2Þ ꢂ 10ꢃ12 cm3 sꢃ1
for C2H4 and kc ¼ ð1:5 ꢁ00:4Þ ꢂ 10ꢃ13 cm3 sꢃ1 for NH3.
Discussion
0
The addition of C2H4 or NH3 on the transients of W (a 5DJ)
together with W (a 7S3) at a total pressure of 6.0 Torr with an
Ar buffer resulted in the following qualitative kinetic beha-
viors: a) W (a 7S3) was more interactive with C2H4 than
NH3, while W (a 5DJ) was less interactive with C2H4 than
In a simulation, it was assumed that these chemical quenching
5
rate constants were adaptable for every a DJ state. The phy-
sical quenching rate constants of the a 5DJ states were estima-
ted from the difference between the apparent total quenching
rate constant and the chemical quenching rate constant, the va-
lues of which were used as the initial values in the trial
calculation. Because the chemical quenching rate constants
7
NH3. b) The important quenching process for W (a S3) by
C2H4 seemed to be chemical, because fast production of the
5
7
5
lower state (a D1) corresponding to the decay of W (a S3)
was apparently not observed. c) The depression efficiency
of the a 7S3 state by NH3 seemed to be comparable with those
for the a 5DJ states, except for the a 5D0 state. d) There
seemed to be some upper states which are physically quenched
of the a DJ states were fairly small compared to the apparent
quenching process in both cases of C2H4 and NH3, the altera-
tion within one order in the chemical quenching rate constant
did not affect these calculation curves significantly. It was
therefore impossible to recognize the characteristically J-de-
pendent of the chemical quenching rate constants of the a
5DJ states, as observed for some oxidants.6 The observed tran-
sient data without C2H4/NH3, especially the three data sets in
Figs. 4a, 4d, and 5d, show a substantial disagreement with the
simulation results compared to the case with the reactant. This
may suggest that there are some minor processes not included
in our mechanism, such as the physical and chemical quench-
ing by coordinatively unsaturated tungsten carbonyls or CO
produced in photolysis. By the addition of the reactant, these
minor processes may have less effect on the total reaction me-
chanism, owing to the appearance of the main processes with
the reactant.
5
5
to the a D4, a 5D3, a 5D1 and a D0 states, mainly by Ar buf-
fer.
Some numerical calculations were carried out in order to de-
duce semi-quantitative information about the kinetics of the a
5
7S3 state and the a DJ states with C2H4 and NH3 from their
transient data. The presumed kinetic mechanism used for
the numerical calculation is shown in Fig. 6. In the kinetic
schemes, kc is a pseudo second-order rate constant for chemi-
i
cal quenching of the i-state, and kp is a physical quenching
rate constant of the i-state to the j-state. Because the energy
ij
5
differences between two states of ꢁJ ¼ 1 in a DJ multiplet
were larger than 1390 cmꢃ1, which are fairly large compared
to the thermal energy at room temperature (ꢆ200 cmꢃ1),
any energy transfer from a lower state to an upper state in
The rate constants estimated from the visual fittings for W
5
7
(a DJ and a S3) are summarized in Table 2 (for C2H4) and
Table 3 (for NH3) together with the apparent second-order
5
the a DJ multiplet was ignored in the calculation, while the

1152 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
Gas-Phase Kinetic Study of Tungsten Atoms
Table 4. Second Order Rate Constants kð2Þ for Chemical
quenching rate constants. Some noticeable tendencies can be
seen in their reactivity, though the resultant rate constants were
not definitive because of the arbitrary kinetic model and of the
arbitrary optimization of the fitting. The apparent second-
7
5
Quenching of W (a S3, a DJ) with Some Simple Mole-
cules at ꢆ300 K (kð2Þ/10ꢃ12 cm3 moleculeꢃ1 sꢃ1
)
a)
b)
O2
NOb)
N2Ob)
SO2
C2H4
NH3
7
order quenching rate constants of W (a S3) with C2H4 and
NH3 of ð4:5 ꢁ 0:5Þ ꢂ 10ꢃ10 and ð0:73 ꢁ 0:10Þ ꢂ 10ꢃ10 cm3
5
a D4
5
1.8
2.5
1.3
7.1
15
9.9
1.5
1.9
1.1
36
22
16
a D3
5
ꢆ1:3
ꢆ0:15
moleculeꢃ1 ꢃ1, estimated simply from the reactant pressure
s
a D2
dependence of the pseudo-first-order depletion rates of W (a
7S3), are slightly smaller than the sum of two rate constants
for the physical and chemical processes. It seems to be reason-
able that the state, which has an upper state with a tendency to
be physically quenched, has a larger total quenching rate con-
stant than the apparent depletion rate constant.
7
a S3
180
190
58
400
520
ꢆ20
5
a D1
5
11
1.6
49
28
0.14
0.003
21
6.0
ꢆ1:3
ꢆ0:15
a D0
1.3
0.15
7
a) Ref. 6. b) Ref. 7.
The a S3 state is chemically more reactive with C2H4 than
NH3. Such a characteristic reactivity for the ꢀ-acceptor and
ꢁ-donor molecules has been observed in the ground state (a
7S3) of the Mo atom, where the second-order rate constant
for the Mo + C2H4 at 296 K was 2:3 ꢂ 10ꢃ13 cm3 sꢃ1 (at
7.8 Torr He), while that for the Mo + NH3 was too small to
quintet. This coordination may subsequently be assisted by
a ꢀ-back donation, because of the relativistic extraction of
the 5d orbital. The addition product of W(C2H4) is expected
to have a bent structure with a triplet (or quintet) electronic
be evaluated, <5 ꢂ 10ꢃ15 cm3 sꢃ1 14
.
Less reactivity of Cr
state with a loose C C bond, the interaction of which may
7
ꢇꢇꢇ
(a 7S3) for NH3 and C2H4 has been observed at room tempera-
ture in the gas phase.15;16 A significant reactivity of W (a 7S3),
be similar to the Cr (a S3) + O2.5
The second-order rate constants for the chemical quenching
of W (a 7S3) for C2H4 and NH3 at room temperature are sum-
marized in Table 4 together with the presumed constants of W
7
even for a ꢁ-donor (NH3), contrary to the cases of Cr (a S3)
and Mo (a S3), may be interpreted as indicating a larger po-
7
5
7
larizability of the tungsten atom caused by 6s/5d hybridization,
though the reported values of the static average electric dipole
polarizabilities for ground state these atoms (Cr; 11.6 (ꢁ2.9),
Mo; 12.8 (ꢁ3.2), W; 11.1 (ꢁ2.8))17 do not show such a
tendency. A 6s-orbit contraction and stabilization as a result
of relativistic effect may make this hybridization easier. How-
ever, some of the resultant collision complexes, W(NH3), can
be expected to be unstable and to dissociate spontaneously.
This dissociation process would contribute to the physical
quenching to the lower electronic state, while the stabilization
process for a collision complex would contribute to the chemi-
cal quenching.
(a DJ). C2H4 is chemically reactive for the a S3 (5d5 6s1)
state as much as the oxidants,6;7 and the electron configuration
also plays an important role in the coordination reaction of
C2H4 in the same manner as the oxidation reactions.
There seems to be slight difference in the efficiency of phy-
5
sical quenching between C2H4 and NH3 for the a DJ states,
where NH3 is observed to physically quench these states more
effectively than C2H4. The degenerate deformation vibration
frequency of NH3 is 1627.5 cmꢃ1 and the C=C stretching vi-
bration frequency of C2H4 is 1623.3 cmꢃ1, which are approxi-
5
mately resonant with the energy gaps of the a DJ manifold
(J ¼ 0 to J ¼ 1, 1670.3 cmꢃ1; 1 to 2, 1655.2; 2 to 3,
1504.5; 3 to 4, 1389.3). The facts that the physical quenching
efficiency of NH3 is about a few times as large as that of C2H4,
and that the physical quenching efficiency monotonously in-
creases with increasing the level energy, are not consistent
with the direct energy-transfer mechanism. These facts sug-
gest that NH3 is relatively easier to form a collision complex
7
The a S3 state (5d5 6s1) is chemically more reactive with
5
C2H4 than any states of a DJ (5d4 6s2) by a factor of 500
7
(Table 2). The bond formation of W (a S3) with C2H4 can
be interpreted by the Dewar-Chatt-Duncanson mechanism
7
based on the analogy of the MO calculation of Mo (a S3)–
7
C2H4.4;18;19 The chemical reaction of W (a S3) with C2H4
5
could be treated as a second-order reaction in kinetic simula-
tions, which may suggest that this addition pathway reaches
a high-pressure limit, even at 6.0 Torr Ar buffer gas, and that
with a DJ than C2H4 because the early barrier, probably for
sd hybridization, may be overcome by the large dipole moment
of NH3. However, the resultant complex, W(NH3), seems not
to be stable at room temperature and to dissociate, where the a
5DJ state may be physically quenched to the lower spin-orbit
state. On the other hand, it is fairly difficult for C2H4 to over-
come the early barrier because the electron density on C2H4
approaching the W atom is not high, resulting in the small
quenching efficiency. The cross sections of intramultiplet mix-
ing between the a 5DJ states with ꢁE ꢈ 1500 cmꢃ1 are on the
ꢀ
the adduct complex W(C2H4) has a rather long lifetime. In
the s1 configuration, the singly occupied d orbital of tungsten
ꢀ
might overlap effectively with the vacant ꢀ orbital of C2H4
owing to the smaller electronic repulsion, weakening the C–
C bond to form three-membered ring complex having some in-
ternal coordinates easy to flow into for the excess energy in the
ꢀ
association coordinate. Such a long-lived M(alkene) com-
ꢀ 2
plex has been expected by a statistical unimolecular rate theory
(RRKM theory), which assesses the plausibility of a saturated
termolecular mechanism, even at 1 Torr He.4 The reduction of
the 6s electron repulsion for the ꢀ electron of C2H4 can be in-
terpreted in terms of the formation of 6s/5d hybridization. The
first step is a weak adsorption of the C2H4 on the W atom,
which is connected with a spin flip from a septet to a
order of 0.1 A , which are smaller than those observed in the
Cr (7PJ) states with ꢁE ꢈ 100 cmꢃ1 by two orders.20
Summary
The a 7S3 state of the tungsten atom is chemically more re-
active with C2H4 than with NH3, though it does react chemi-
cally with NH3 differing from the other group-6 elements

Y. Ishikawa et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1153
(Cr, Mo). Both reactions, W (a 7S3) + C2H4 and W (a 7S3) +
NH3, could be analyzed as being bimolecular, implying that
even at a low total pressure of 6-Torr Ar the adduct complexes,
6
M. L. Campbell and R. E. McClean, J. Chem. Soc., Fara-
day Trans., 91, 3787 (1995).
J. S. S. Harter, M. L. Campbell, and R. E. McClean, Int. J.
Chem. Kinet., 29, 367 (1997).
K. Honma, M. Nakamura, D. E. Clemmer, and I. Koyano,
J. Phys. Chem., 98, 13286 (1994).
K. S. Pitzer, Acc. Chem. Res., 12, 271 (1997).
7
ꢀ
ꢀ
W(C2H4) or W(NH3) , had a relatively long lifetime, enough
for deactivation. The a 7S3 state (5d5 6s1) is chemically more
reactive with C2H4 than any states of a 5DJ (5d4 6s2) by a fac-
tor of 500 (Table 2). The physical quenching appeared to
8
9
10 P. Pyykko, Chem. Rev., 88, 563 (1988).
11 L. Lian, P. A. Hackett, and D. M. Rayner, J. Chem. Phys.,
99, 2583 (1993).
12 Y. Ishikawa and Y. Matsumoto, Chem. Phys. Lett., 352,
209 (2002).
13 ‘‘Atomic Energy Levels, Vol. c,’’ ed by C. E. Moore,
NSRDS-NBS, 35, U.S. Government Printing Office, Washington
D.C. (1971), p. 156.
14 L. Lian, S. A. Mitchell, and D. M. Rayner, J. Phys. Chem.,
98, 11637 (1994).
15 J. M. Parnis, S. A. Mitchell, and P. A. Hackett, J. Phys.
Chem., 94, 8152 (1990).
¨
5
dominate over the chemical quenching for W (a DJ) in both
cases of C2H4 and NH3. The large interaction between W
(a 7S3) and W (a 5DJ), not only with C2H4, but also with
NH3, is consistent with the relativistic contraction of the 6s
shell and the relativistic expansion of the 5d shell.
We are grateful to Professor K. Honma (Himeji Institute of
Technology) for useful comments on the kinetic interpretation.
This work was supported by a Grant-in-Aid for Scientific Re-
search (No. 11640508) from the Ministry of Education, Cul-
ture, Sports, Science and Technology.
16 Y. Ishikawa, H. Ogawa, and T. Iwaki, Bull. Chem. Soc.
Jpn., 75, 493 (2002).
17 ‘‘CRC Handbook of Chemistry and Physics, 70th Edition,’’
ed by R. C. Weast, CRC Press, Inc., Boca Raton, Florida (1989–
1990), E-72.
References
1 Y. Wen, A. Yethirai, and J. C. Weisshaar, J. Chem. Phys.,
106, 5509 (1997).
2
3
4
M. L. Campbell, J. Phys. Chem. A, 102, 892 (1998).
K. Honma, J. Phys. Chem. A, 103, 1809 (1999).
J. J. Carroll, K. L. Haug, J. C. Weisshaar, M. R. A.
18 M. R. A. Blomberg, P. E. M. Siegbahn, and M. Svensson,
J. Phys. Chem., 96, 9794 (1992).
19 J. J. Carroll, K. L. Haug, and J. C. Weisshaar, J. Am. Chem.
Soc., 115, 6962 (1993).
20 J. M. Parson and T. Ishikawa, J. Chem. Phys., 80, 3137
(1984).
Blomberg, P. E. M. Siegbahn, and M. Svensson, J. Phys. Chem.,
99, 13955 (1995).
5
A. Martinez, J. Phys. Chem. A, 102, 1381 (1998).