Internal Energy Distribution of the NCO Fragment from Near-Threshold Photolysis of Isocyanic Acid, HNCO

Article abstract of DOI:10.1021/jp952667r

We report the first measurement of the vibrational- and rotational-state distributions in the NCO fragment resulting from photolysis of HNCO.Recent studies have drawn conclusions about the photochemistry of HNCO and the vibrational distribution in the NCO fragment from observations of the kinetic energy distribution of the H atom produced in this photolysis; however, there has been no direct observation of the NCO fragment itself.We use laser-induced fluorescence to detect the nascent NCO photoproducts and spectral simulations to extract vibrational-state populations.The rotational distributions, where we can measure them, show little excitation, and the vibrational energy preferentially appears in the bending mode.The vibrational-state distribution results directly from the excited-state geometry of the HNCO parent, in which the NCO group is bent.The dissociation proceeds from this bent NCO group to a linear NCO fragment, strongly exciting the bending mode.We find about 65percent of the total energy in relative translation of the fragments, while 30percent goes into vibration and 5percent into rotation of NCO.

Full text of DOI:10.1021/jp952667r

7948
J. Phys. Chem. 1996, 100, 7948-7955
Internal Energy Distribution of the NCO Fragment from Near-Threshold Photolysis of
Isocyanic Acid, HNCO
Steven S. Brown, H. Laine Berghout, and F. Fleming Crim*
Department of Chemistry, UniVersity of WisconsinsMadison, Madison, Wisconsin 53706
ReceiVed: September 12, 1995X
We report the first measurement of the vibrational- and rotational-state distributions in the NCO fragment
resulting from photolysis of HNCO. Recent studies have drawn conclusions about the photochemistry of
HNCO and the vibrational distribution in the NCO fragment from observations of the kinetic energy distribution
of the H atom produced in this photolysis; however, there has been no direct observation of the NCO fragment
itself. We use laser-induced fluorescence to detect the nascent NCO photoproducts and spectral simulations
to extract vibrational-state populations. The rotational distributions, where we can measure them, show little
excitation, and the vibrational energy preferentially appears in the bending mode. The vibrational-state
distribution results directly from the excited-state geometry of the HNCO parent, in which the NCO group is
bent. The dissociation proceeds from this bent NCO group to a linear NCO fragment, strongly exciting the
bending mode. We find about 65% of the total energy in relative translation of the fragments, while 30%
goes into vibration and 5% into rotation of NCO.
I. Introduction
have measured the rotational, vibrational, and translational
energy disposals of both the NH and CO fragments at a variety
of photolysis wavelengths ranging from 230 to 193.3 nm,8,19-24
and all experiments are consistent with an impulsive dissociation
on a steeply repulsive potential energy surface in which the NCO
skeleton is bent rather than having its ground-state linear
geometry. The other major photochemical channel in HNCO
has received less experimental attention, despite the accessibility
of the NCO fragment via LIF. A number of studies have
focused on the measurement of the branching ratio between the
two channels. Using kinetic measurements, Spiglanin et al.
determined an upper limit to the quantum yield for NCO at
193.3 nm of ΦNCO e 0.1,8 and Yi and Bersohn later confirmed
this value in measurements of H atom LIF subsequent to HNCO
photodissociation at 193.3 and 212.6 nm. They found quantum
yields of ΦNCO ) 0.05 ( 0.006 and ΦNCO ) 0.13 ( 0.01,
respectively.25 Earlier studies that measured the concentrations
of final products from the reactions of the radicals produced in
the initial photolysis step also found that the production of NH
(a1∆) dominates at short photolysis wavelengths but that the
importance of the NCO channel increases at longer wave-
lengths.5 Our own study of this photochemistry, which mea-
sured the relative LIF signals of NCO and NH (a1∆) fragments,
found that the NCO channel dominates near the energetic
threshold for production of NH (a1∆) (240.8 nm) but that the
quantum yield for NCO drops dramatically between 235 and
225 nm.10 A very recent study of the translational energy of
the H atom produced in the photolysis of HNCO with 193.3-
nm light determined that 70% of the available energy went into
relative translation, 26% went into vibration of the NCO
fragment, with a long progression in the bending mode, and
4% went into NCO rotation. These observations, like those
for the NH + CO channel, are consistent with a rapid
dissociation from an excited state with a bent NCO geometry.26
The chemistry of isocyanic acid, HNCO, is both practically
and fundamentally interesting. For example, HNCO removes
deleterious NOx compounds from exhaust gases after combus-
tion, a process called RAPRENOx (RAPid REduction of NOx).1
The NCO radical, produced by abstraction of an H atom from
HNCO by OH radicals present in the exhaust gas, seems to
play a major role in the reduction of NOx.2 This chemistry has
implications for dealing with issues such as photochemical smog
and acid rain. The photochemistry of isocyanic acid is also
fundamentally interesting because it has several excited-state
decomposition pathways that involve different product electronic
states and cleave different bonds. In particular, irradiation of
HNCO at wavelengths below 242 nm produces either H + NCO
or NH + CO photoproducts,3-6 where the NH radical is
produced in its lowest singlet state rather in the ground triplet
state,
NH (a1∆) + CO (X1Σ+)
HNCO (1A′′)
H (X2S) + NCO (X2Π)
The N-H dissociation threshold is 37 630 ( 420 cm-1,7 and
that for C-N singlet dissociation lies slightly higher at 41 530
( 150 cm-1 8
.
Recently, we reported the bond-selective photodissociation
of isocyanic acid.9,10 In these experiments, we used a two-step
excitation to the dissociative electronic surface (vibrationally
mediated photodissociation)11,12 in which we first prepared a
vibrationally excited N-H stretching state containing 3 or 4
quanta of excitation. Photodissociation of this state changed
the branching ratio between the two channels, ΦNCO/ΦNH, by a
factor of at least 17 in favor of the NCO channel relative to
photodissociation of the ground vibrational state. In this
experiment, we used laser-induced fluorescence (LIF) detection
of both NH (a1∆)13-15 and NCO (X2Π)16-18 fragments and
compared the relative intensities of each.
The ultraviolet absorption in HNCO at photon energies that
dissociate the molecule comes from a transition from the ground
to the first excited singlet state, 1A′′ r 1A′ in Cs symmetry.27-29
The ground (1A′) state is planar and has a nearly linear N-C-O
angle of 172° and an H-N-C angle of 124°.30 The excited
(1A′′) state results from the promotion of an electron from the
The NH (a1∆) + CO (X1Σ+) channel in the photodissociation
of HNCO has received considerable attention. Several authors
X Abstract published in AdVance ACS Abstracts, April 1, 1996.
S0022-3654(95)02667-0 CCC: $12.00 © 1996 American Chemical Society

Internal Energy Distribution of the NCO Fragment
J. Phys. Chem., Vol. 100, No. 19, 1996 7949
2a′′ orbital, which is primarily a combination of N and O atom
p orbitals, to the antibonding 10 a′ orbital. Because the energy
of the 2a′′ orbital is smallest for a linear NCO group and the
minimum energy of the 10 a′ orbital occurs for a bent NCO
group,29 the geometry of HNCO changes significantly upon
excitation. Analysis of rotational structure at the low-energy
end of the UV absorption spectrum suggests a bent NCO group
in the excited state with a N-C-O angle of 119° for a trans
configuration.27
We directly observe the quantum-state distribution of the
NCO fragment from the photolysis of HNCO at several different
photolysis energies near the threshold for production of this
fragment. There are a few recent studies of the gas-phase
quantum-state distribution of the NCO fragment from other
sources, including the reaction of fluorine atoms with isocyanic
acid (F + HNCO),16,31 cyanide radicals with molecular oxygen
(CN + O2),18,32 and a flame of CH4 and N2O.17 The spectrum
of NCO (X2Π) provides one of the few examples of the
Renner-Teller effect, which couples the vibrational angular
momentum from the degenerate bend in this linear triatomic
molecule to the orbital angular momentum of the Π electronic
state.33-37 Indeed, previous studies have found extensive
excitation of the bend in NCO17,18 and have characterized both
the Renner interaction and the Fermi resonance that occur
between the states (V1, V2, V3) and (V1 + 1, V2 - 2, V3), where
ν1 is the symmetric stretch, ν2 is the doubly degenerate bend,
and ν3 is the antisymmetric stretch.38-40
We are interested in determining the rotational and vibrational
quantum-state populations in the NCO fragment from the
photolysis of HNCO. A measurement of the nascent popula-
tions of the NCO photofragments helps reveal the dynamics
that occur on the excited-state surface of HNCO. In our
vibrationally mediated photodissociation experiments,10,41 we
wish to measure the effect of exciting particular vibrational states
in the parent HNCO on the dynamics of the photodissociation
into the NCO channel, and we must first establish the quantum-
state distributions that result from photodissociation out of the
ground vibrational state.
Figure 1. Photolysis yield spectrum of NCO (X2Π) fragments from
HNCO photolysis in three different vibrational states: (000), (010), and
(020). Vibrational and rotational assignments of the bands from Dixon
(ref 27) appear over the spectra and the inset, respectively. The arrows
on the lower trace mark the literature value for the N-H dissociation
threshold (ref 7) and the photolysis energies used in this experiment.
We measured the effects of saturation by recording the
intensities of NCO rotational lines in the spectrum of the ground
vibrational state in a rotationally thermalized sample, which we
obtained by introducing several microseconds of delay between
the photolysis and probe lasers and increasing the sample
pressure to 200 mTorr. Taking into account the effect of
saturation on the anticipated line intensities using a simple three-
state kinetic model,42 we adjusted the value of the spectral power
density, F, in our simulations to fit the experimentally observed
intensity distribution, assuming that it was a Boltzmann distribu-
tion with a temperature of 300 K. We then used this value of
F to simulate the nascent distributions of rotational states using
the expression for partially saturated rotational line strengths.32,42
The photolysis and probe beams counterpropagated through
a cylindrical fluorescence cell in which the pressure of isocyanic
acid was 100 mTorr. The HNCO vapor flowed through this
cell from a reservoir of liquid HNCO at -78 °C (dry ice/acetone
bath) with a vapor pressure slightly under 1 Torr.43 We verified
that the photoproducts from the preceding laser shot did not
interfere with our signal by placing the probe laser pulse slightly
prior to the photolysis laser pulse and observing that there was
no NCO LIF signal. For experiments measuring the nascent
distribution of NCO fragments, the photolysis-probe delay was
30 ns, conditions under which we estimate that fewer than 3%
of the NCO photofragments have a hard-sphere collision before
detection. An f/1 optical system, photomultiplier tube, gated
integrator, digital-to-analog converter, and laboratory computer
collected and stored the fluorescence signals from NCO A(2Σ+)
We synthesized HNCO by the dropwise addition of a saturated
solution of potassium cyanate (KOCN) to concentrated phos-
phoric acid (H3PO4).44 After condensing HNCO vapor in a trap
at -196 °C, we distilled it twice between traps at -30 and -196
°C to remove water.
II. Experimental Approach
The experimental apparatus is similar to one described
previously for vibrationally mediated photodissociation of
isocyanic acid,9 except that here we use only two lasers: one
for photolysis of HNCO and one for detection of the NCO
fragment by laser-induced fluorescence. The photolysis laser
is a Nd:YAG laser pumped dye laser operating on a range of
coumarin dyes with a frequency-doubled output in the range
240-260 nm and a pulse energy of e1 mJ. The probe laser is
an XeCl excimer laser pumped dye laser that uses a series of
stilbene and coumarin dyes to detect various vibrational states
of NCO (X2Π) by laser-induced fluorescence. The energy of
the probe laser is on the order of 100 µJ in a 10-ns pulse focused
to a spot approximately 1 mm in diameter.
The largest source of uncertainty is the power drift in both
the photolysis and probe lasers due to the short lifetime of the
coumarin dyes. The problem was particularly pronounced for
the photolysis laser, which had a higher overall pulse energy.
We compensated for the resulting signal drift by monitoring
the pulse energy of both lasers using photodiodes and normal-
izing the signal. We checked for consistency by scanning the
NCO spectrum several times, scanning the spectral regions in
a different order each time. The fluctuations in the normalized
signal level were approximately 15-20% using this procedure,
and we used this as the uncertainty in our measured vibrational-
state populations of NCO.
III. Results
A. Yield Spectra. Figure 1 shows the photofragment yield
spectrum of NCO (X2Π) produced from HNCO photolysis. Each
trace of this figure has the probe laser fixed on a transition
originating in a different vibrational state in NCO (X2Π) and
the photolysis laser scanned through the region of the dissocia-
tion threshold. Superimposed above the spectra are a series of

7950 J. Phys. Chem., Vol. 100, No. 19, 1996
Brown et al.
combs that mark the vibrational progressions that Dixon and
Kirby27 identified in the UV absorption spectrum of HNCO.
They do not assign these progressions to any particular
vibrations in the electronically excited state of HNCO but rather
label them as A, B, C, and D. Because the A and B progressions
occur primarily at lower energies in the UV absorption spectrum,
below the threshold for dissociation of the N-H bond, we do
not observe them in our yield spectrum. Progressions C and
D, which begin near the N-H dissociation threshold, are clearly
apparent in our spectra. The spacing between the bands in each
of these progressions is approximately 500 cm-1, suggesting
that they are progressions in bending vibrations on the excited
surface, consistent with the excited state having a minimum for
a bent NCO geometry. The Franck-Condon factors should
favor excitation of bending modes on such a surface. Dixon
and Kirby suggested that the interleaving of two vibrational
progressions might reflect the presence of cis and trans isomers
in the excited state. The structure in each of the spectra probing
different vibrational states in the NCO fragment is identical,
except for the increase in the appearance thresholds for different
vibrational states with increasing excitation in the NCO bending
mode. The arrow on the left-hand side of the lower trace marks
the measured threshold for NCO production7 and its uncertainty.
Dixon and Kirby’s analysis of the UV spectrum included
rotational as well as vibrational assignments. The rotational
structure is quite diffuse, and they identified it primarily from
the more well-defined structure at the low energy end of the
UV spectrum below the NCO appearance threshold. The inset
in Figure 1 is an expanded view of one of the vibrational bands
in our photolysis yield spectrum of NCO (000), with the
rotational assignments for this band27 superimposed over the
top. The rotational structure is not extremely clear because the
bands are diffuse at this energy and there is a large amount of
rotational congestion. Isocyanic acid in its ground state is a
near-prolate top, with an asymmetry parameter κ ) 0.9996.30
The ground-state A rotational constant, 30.637 cm-1, is quite
large, and the Ka structure dominates the rotational bands in
the spectrum since the energy spacing of different Ka states is
intermediate between the bending vibrations and that of the J
and Kc states.45 Based on the observed rotational structure,
Dixon27 assigned the UV bands of HNCO to a perpendicular
transition with prolate-top selection rules of ∆K ) (1. The
comb in the inset of Figure 1 marks the position of the Q branch
transitions for different values of K. The notation of the markers
is ∆K∆JK. Because the A rotational constant is so large, only a
small number of K states are populated in our room-temperature
sample. The measured change in the magnitude of the A
rotational constant from Dixon and Kirby’s work is ∆A ) -26
cm-1, which is consistent with the equilibrium geometry of the
NCO group changing from nearly linear in the ground state to
bent in the excited state.
B. NCO Spectroscopy. Renner-Teller interactions33,34
strongly influence the vibrational energy level structure of NCO
(X2Π). Hougen describes this interaction,35 and here we give
only a brief summary of the remarkably complex vibrational
energy level structure of NCO. For a particular value of the
bending quantum number, V2, the quantum number for the
projection of the vibrational angular momentum along the
molecular axis is l, which takes on the values l ) V2, V2 - 2, V2
- 4, ..., -V2. The Renner-Teller interaction couples the
vibrational angular momentum to the orbital angular momentum,
generating a new angular momentum quantum number, Κ )
Λ + l. For example, with V2 ) 2, l ) 2, 0, and -2, and Λ )
(1 in the Π electronic state, there are six different combinations
Figure 2. Vibronic energy level diagram of NCO (X2Π) showing the
Renner-Teller and spin-orbit interactions.
of l and Λ that give the values Κ ) 3, 1, 1, -1, -1, and -3.
These six states divide into three sets of two. There are two Π
states, which are labeled µ and κ such that the κ states lie above
the µ states in energy. There is also one unique Φ state
corresponding to Κ ) V2 + 1. The spin-orbit interaction further
splits these states into Ρ ) Κ + Σ components, where Σ ) 1/2
for the doublet spin state. The energy level diagram in Figure
2 illustrates these interactions and the energy ordering of the
states for NCO, which has a negative spin-orbit parameter.36
C. Product-State Distributions. We measured the product-
state distribution of the NCO photofragments at five different
photolysis energies near the threshold for cleavage of the N-H
bond. The arrows on the lower trace in Figure 1 mark the
different photolysis wavelengths, for which the available energy
ranges from 890 to 3030 cm-1, and which excite both C and D
vibrational bands in the excited state of HNCO. In each case,
we fix the photolysis laser wavelength on the blue edge of the
vibrational band in order to excite the K′′ ) 0, RQ0 band. This
assures consistency in the rotational states excited for different
bands since the sharp edge on the blue side of each vibrational
band is the most prominent and easily identified feature, as the
inset in Figure 1 shows.
The laser-induced fluorescence spectrum from nascent NCO
fragments at the lowest photolysis energy (hνDiss ) 38 516 cm-1
or Eavail ) 890 cm-1 above threshold) is the lower trace of Figure
3, and a simulation of the bands is the upper trace. We used
the simulation to determine the populations of different vibra-
tional states in the experimental spectrum using the prescription
described by Sauder et al.32 Table 1 lists the vibrational-state
populations. There is little enough congestion in this spectrum

Internal Energy Distribution of the NCO Fragment
J. Phys. Chem., Vol. 100, No. 19, 1996 7951
TABLE 1: NCO (X2Π) Vibronic-State Populations in the
Photolysis of HNCO
populationsb for excess photolysis energy, cm-1
vibronic energy,a
state
cm-1
890
1135
1950
2250
3030
(000)
2Π3/2
2Π1/2
0
94.8
1
0.46
1
0.62
1
0.96
1
0.94
1
0.82
(010)
µ2Σ+
488.7
534.5
628.3
669.6
0.04
0.24
0.11
0.03
0.03
0.24
0.14
0.02
0.16
0.81
0.68
0.13
0.26
0.81
0.46
0.16
0.25
1.04
0.75
0.17
2
∆
5/2
2
∆
3/2
κ2Σ-
(020)
µ2Π
989.5
0.02
0.03
0.08
0.03
0.08
0.11
0.25
0.52
0.48
0.25
0.26
0.58
0.59
0.20
0.60
0.89
0.76
0.68
2Φ7/2 1051.0
2Φ5/2 1143.6
κ2Π
(030)
1220.2
<0.01
<0.01
µ2Σ- 1466.7
0.01
0.04
0.10
0.10
0.02
0.08
0.08
0.09
0.16
0.13
0.08
0.20
0.29
0.27
0.59
0.44
0.32
0.13
µ2∆
2Γ9/2
2Γ7/2
κ2∆
1495.3
1571.7
1661.4
1763.5
Figure 3. Laser-induced fluorescence spectrum of NCO (X2Π) from
HNCO photolysis at Eavail ) 890 cm-1. The upper trace is a simulation
of the spectrum that was used to extract vibrational populations from
the experimental data. The combs over the top mark the positions of
the P2 and Q1 bandheads belonging to the unique Κ ) V2 + 1 vibronic
states that occur in the center of each ν2 manifold.
κ2Σ+ 1777.9
(040)
2Η11/2 2091.6
2Ηg/2 2179.0
0.05
0.04
0.35
0.22
that it is relatively easy to pick out the individual rotational
structure, particularly in the intense (000) bands, which are
clearly the most populated at this low photolysis energy. The
combs over the top of the spectrum mark the intense bandheads,
which occur for the P2 and Q1 branches that belong to the unique
Κ ) V2 + 1 states that occur for each value of V2.18 There is
also a great deal of rotational structure, including numerous
bandheads, arising from states with Κ * V2 + 1 that is more
apparent in spectra taken at higher photolysis energies, where
there is greater population in states with V2 > 0. However, we
use the Κ ) V2 + 1 bandheads as convenient markers for each
ν2 state since they are very intense and occur in the center of
the manifold.
Using the rotational branches in the (000) manifold that are
not badly overlapped, we fit the line intensities to a Boltzmann
distribution to determine a rotational temperature. Such a
distribution is not necessarily appropriate for the unrelaxed
photolysis products, but the observed intensities fit this distribu-
tion remarkably well. We also considered an unbiased prior
rotational distribution, taking into account only the rotational
and translational densities of states for a given excess photolysis
energy.46,47
(050)
2Ι3/2
2Ι11/2
(100)
2609.2
2693.2
0.09
0.07
2Π3/2 1273.0
2Π1/2 1362.4
0.06
0.05
0.08
0.05
0.28
0.12
(110)
2
2
∆
∆
1828.2
1909.0
<0.01
<0.01
0.01
0.03
0.11
0.14
5/2
3/2
(120)
2Φ7/2 2364.4
2Φ5/2 2437.5
0.05
0.03
(001)
2Π3/2 1920.6
2Π1/2 2017.7
0.05c
0.02c
(011)
2
2
∆
∆
2440.0
2534.8
0.02c
5/2
0.005c
3/2
a Reference 18. b Uncertainty is (20%. c Estimate based on an
assumed value for the Franck-Condon factor.
the more flexible Boltzmann distribution does quite well. Even
if we allow the maximum available energy to vary as a fit
parameter, the prior distribution does not fit the data as well as
the Boltzmann distribution. We used the Boltzmann distribution
in all of our analysis, and we report rotational populations for
all NCO bands in terms of temperature.
1/2
P0(J) (2J + 1) (Erot(J) - Evib - Eavail
)
In this expression, Erot is the rotational energy as a function of
J, Evib is the vibrational energy, and Eavail is the total available
excess energy from the photolysis. Figure 4 is a plot of the
rotational populations obtained from the P12, Q2 + R12, and R2
branches in the (000) 2Π1/2(F2) state as a function of J for two
different photolysis energies. The solid lines are fits to a
Boltzmann distribution with rotational temperatures of 225 K
for the lower energy photolysis and 415 K for the higher energy
photolysis, and the dashed lines are the unbiased prior distribu-
tions, which have no adjustable parameters other than the scale
factor. We have weighted the prior distributions according to
the population of HNCO rotational states present in our thermal
gas sample, assuming that HNCO is a prolate symmetric top
and that we excite exclusively the K ) 0 f 1 transition and all
possible J states within the K ) 0 sublevel. The prior
distribution is clearly too energetic to describe our data, but
Determining reliable rotational temperatures for the higher
lying vibrational bands is more difficult. The spectral conges-
tion, even at very modest photolysis energies, makes extracting
rotational distributions difficult even for the (010) states. Rather
than trying to fit the data for these bands to specific rotational
temperatures, we estimated the dependence of the best fit
rotational temperature on the available energy. The only band
which has little enough congestion from overlap with other
bands for us to observe intensities of individual lines is the (000)
2Π1/2 band at the red end of the A(0V20) r X(0V20) spectrum,
where we can measure rotational populations even at higher
photolysis energies. We measured the rotational temperatures
in this state and fit them to a linear dependence on the available
energy. These rotational temperatures appear at the bottom of

7952 J. Phys. Chem., Vol. 100, No. 19, 1996
Brown et al.
Figure 5. Experimental LIF spectra of NCO fragments produced from
photolysis of HNCO in three separate D bands at Eavail ) 890, 1950,
and 3030 cm-1. The increase in excitation of the bending mode with
increasing photolysis energy is apparent.
diagonal and off-diagonal, makes assignments very difficult.
For photolysis energies that populate these two vibrational states,
we simulated only the most intense structure of the Κ ) V2 +
1 states. The resulting vibrational state populations are in Table
1.
The photolysis does not populate the stretching modes, ν1
and ν3, of the NCO fragment as efficiently as the bend. Figure
6 is the LIF spectrum of the NCO fragment from the photolysis
at 3030 cm-1, displaying the A(0V20) r X(1V20) and A(0V20)
r X(0V21) bands as well as the A(0V20) r X(0V20) bands. The
Franck-Condon factors for the transitions from V1′′ and V3′′ )
1 to V1′ and V3′ ) 0 on the left-hand side of the spectrum are
not as strong as the diagonal transitions between V1′′ and V3′′ )
0 and V1′ and V3′ ) 0 on the right-hand side. We have taken
into account the decrease in our partially saturated line
strengths42 for the A(0V20) r X(1V20) structure based on the
calculated Franck-Condon factor for this transition from Sauder
et al.32 The magnitude of the vibrational overlap integral from
V1′′ ) 1 to V1′ ) 0 is 0.435, compared to a value of 0.890 for
the diagonal transition.32 The A(0V20) r X(1V20) structure in
the middle of Figure 6 includes this difference such that its
intensity reflects the population in X(1V20) when compared to
the diagonal A(0V20) r X(0V20) bands. The Franck-Condon
factors for A(0V20) r X(0V21) are not available.32 Unless these
factors are anomalously small, the low intensity of the A(0V20)
r X(0V21) transition on the left side of Figure 6 reflects a
significantly smaller population in the V3′′ ) 1 states than in
excited bending states at comparable energies. For the sake of
comparison, we use the same scale factor for the V3′′ ) 1 bands
Figure 4. Average intensities of rotational lines from three different
branches in the (000) 2Π1/2 state of NCO produced at Eavail ) 890 and
2250 cm-1. The solid lines are fits to Boltzmann distributions with a
temperature of 225 K for the 890-cm-1 data and 415 K for the 2250-
cm-1 data. The dashed lines are the predictions of a purely statistical
prior distribution.
Table 2. We extrapolated the rotational temperature for the
higher lying vibrational states at each photolysis energy from
the rotational temperature determined for the (000) 2Π1/2 state,
the available energy for each vibrational state, and the depen-
dence of the rotational temperature on available energy. The
resulting simulations appear visually correct, and changing the
predicted rotational temperature by (50-100 K makes the
simulated spectra noticeably worse.
The rapid increase in the vibrational excitation of the NCO
fragment with increasing photolysis energy is evident in Figure
5, which displays the A(0V20) r X(0V20) region of the NCO
LIF spectrum for excitation of three different D vibrational bands
in the photolysis of HNCO at 860, 1950, and 3030 cm-1 above
the dissociation threshold. The strongest excitation in all of
the distributions is the bending mode, with excitation evident
up to V2 ) 5. For the V2 ) 4 and V2 ) 5 states, the bands are
quite weak, and the congestion from numerous bands, both
TABLE 2: Energy Partitioning in NCO(X2Π) vs Excess Photolysis Energya
energy partitioning
excess photolysis energy, Eavail, cm-1
890
0.24
1135
0.21
1950
0.30
2250
0.28
3030
0.31
14 100b
impulsive limitc
0.025e
statistical limit (3030 cm-1
0.310
)
f
fbend
0.26e
fstretch
frot
ftrans
0
0.10
0.66
0
0.07
0.72
0.01
0.04
0.65
0.01
0.05
0.66
0.02
0.04
0.63
0.105
0.400
0.185
0.04
0.70
0.020
0.955
T
rot (2Π3/2),d K
225
225
325
415
480
a Typical uncertainty in the fractions is (20%. b Reference 26. Measured by H atom center-of-mass kinetic energy at 193.3-nm photolysis.
c Reference 49. d Best fit rotational temperature in the NCO (000) 2Π3/2 state. e Sum of vibrational excitation in both stretching and bending modes.
f
fi ) Ei /Eavail
.

Internal Energy Distribution of the NCO Fragment
J. Phys. Chem., Vol. 100, No. 19, 1996 7953
Figure 6. Comparison of the LIF spectra of NCO fragments probing
the (0V20), (1V20), and (0V21) bands. The A(0V20) r X(1V20), (0V21)
bands are scaled to account for the difference in the Franck-Condon
factors between the A(0V20) r X(1V20) and A(0V20) r X(0V20) bands.
The Franck-Condon factor for the A(0V20) r X(0V21) bands is
unknown, and we took them to be the same as those for the A(0V20)
r X(1V20) bands for the sake of comparison.
as for the V1′′ ) 1 bands in Figure 6, but because the actual
Franck-Condon factors for these bands are unknown, our
analysis does not accurately determine the populations for the
(0V21) states. As with the higher lying (0V20) bands, we have
simulated only the structure belonging to the Κ ) V2 + 1
vibronic states for excitation of ν1 and ν3, and we present the
vibrational-state populations in Table 1.
Figure 7. Vibrational-state populations in the NCO bending mode
summed over all of the states in a given ν2 manifold for three different
photolysis energies. The solid lines are the predictions of an unbiased
prior distribution.
IV. Discussion
The internal energy distribution of the NCO fragment contains
information about the photodissociation dynamics and about the
nature of the excited electronic state in the parent HNCO
molecule. The simplest quantity characterizing the dissociation
is the amount of available energy after the photolysis, which in
turn depends on the actual value of the threshold for dissociation
of the N-H bond, a quantity having an uncertainty of several
hundred wavenumbers (Figure 1).7 The appearance threshold
for NCO fragments in our experiment is consistent with the
reported value of the dissociation energy, but we cannot reduce
the uncertainty in this quantity. Our room-temperature sample
has significant rotational excitation, amounting to approximately
100 cm-1 on average, which is comparable to the measured
uncertainty in the value of the bond dissociation energy. There
is also a strong likelihood that there is a photolysis signal from
vibrational hot band transitions in our photolysis yield spectra.
The lowest vibrationally excited state in HNCO is the 577-cm-1
cis bend,48 which has a population of about 6% compared to
the ground state in our 300 K sample. Because of the geometry
change upon excitation, hot band transitions from an excited
bending state are likely to have favorable Franck-Condon
factors and contribute significantly to the photolysis signal. In
fact, we find population in states of the NCO fragment that lie
above the total excess photolysis energy as measured by the
difference between the photon energy and the literature value
of the dissociation energy, suggesting the participation of hot
bands.
reflect the lifetime of the excited state, and, consistent with such
lifetime broadening, increases with photolysis energy. Because
the extensive congestion in the spectrum prevents the measure-
ment of the line width of an isolated rotational transition, we
can establish only an upper limit to the lifetime of the excited
state of HNCO. In separate double-resonance measurements
of widths of single rotational lines in the NCO photolysis yield
spectrum,41 we obtain a width of about 1.8 cm-1 in the energy
region of the inset of Figure 1, corresponding to an excited-
state lifetime of 3 ps.
The change in the geometry of the NCO group from linear
in the ground state to bent in the excited state produces a
significant force along the bending coordinate during the
dissociation. The comparison between the populations of
excited stretching states in NCO compared to isoenergetic
excited bending states illustrates this point. For example, Table
1 shows that in the dissociation at Eavail ) 3030 cm-1, the
population of the (020) Φ states is more than 4 times the
population of the corresponding Π states in the (100) manifold.
Likewise, the ratio of the populations in (030) Γ and (110) ∆ is
also approximately 4, and the ratio of (040) Η to (120) Φ is
more than 7. A purely statistical dissociation would populate
each of these sets of states nearly equally since they have the
same degeneracy and roughly the same energy.
Figure 7 is a plot of the distribution of population in the
bending states as a function of V2 for the three highest photolysis
energies. The populations are sums over all of the states in a
single ν2 manifold relative to the population in V2 ) 0. For V2
) 4 and 5, where we have only measured the population in the
The rather diffuse spectra in Figure 1 display both vibrational
and coarse rotational structure. Our photolysis laser line width,
which is on the order of 0.1 cm-1, does not limit the width of
the features in the spectrum. The observed width seems to

7954 J. Phys. Chem., Vol. 100, No. 19, 1996
Brown et al.
Κ ) V2 + 1 states, we estimate the population summed over
all of the states in the manifold from the amount of signal in
the states that we measured and the amount of the total
population represented by these states in V2 ) 0, 1, 2, and 3.
The error bars on the points are (20% of the measured values,
based on the uncertainty measurement described above. The
smooth lines in these plots are the predictions of the unbiased
prior distribution, considering excitation only in the bending
mode. We approximate this distribution as a function of the
vibrational energy by treating both the vibrational and rotational
energies as continuous variables and integrating over the
rotational distributions. The result is that the vibrational-state
population depends on the difference between the available
energy and the vibrational energy,46
For comparison, the energy partitioning in the statistical limit
for Eavail ) 3030 cm-1 also appears in Table 2. Unlike the
impulsive model, the partitioning in the prior distribution
depends on the amount of available energy, with the fraction
in NCO internal energy increasing steadily with available
energy. As we noted above, the statistical calculation predicts
the excitation in the NCO bend quite well but fails to accurately
predict the amount of energy in other vibrations, NCO rotations,
or relative translation. Clearly, an accurate description of the
dissociation requires a more complicated dynamical model that
includes the forces in the exit channel. Except for the bend,
the experiment finds more energy in NCO internal degrees of
freedom than the impulsive model predicts and less than the
prediction of the prior distribution. The impulsive model
describes state distributions well if the excited-state potential
energy surface is purely repulsive, and the prior distribution
does well for a long-lived excited electronic state with no
interaction as the products separate. Our results show that
HNCO is predissociative at energies near the N-H dissociation
threshold, indicating the presence of a well on the excited-state
surface. The partitioning of large amounts of energy into the
NCO bending mode and the relative translation of H and NCO
fragments suggests that there are strong forces along both the
N-H and the NCO bending coordinates on the dissociative
surface.
P0(V2) 2(V2 + 1)(Eavail - Evib)3/2
The 2(V2 + 1) factor is the degeneracy of the vibrational state.
The distributions are quite close to the statistical prediction,
especially for the highest energy photolysis with Eavail ) 3030
cm-1. For the other two energies, the actual distributions are
slightly less energetic than the prior distribution. The statistical
calculations do a much poorer job of predicting the amount of
vibrational excitation in the symmetric stretch, overestimating
the population in this mode by a factor of 2-4. It is clear that
the vibrational excitation results from the forces present on the
dissociative potential energy surface, which should be much
stronger in the bending coordinate than in the symmetric or
antisymmetric NCO stretches. The rotational state distributions
in Figure 4 are also less energetic than a purely statistical model
predicts. Both the vibrational and rotational distributions
suggest that the force of the dissociation, rather than a statistical
partitioning of energy, determines the internal energy content
of the NCO fragment.
Table 2 summarizes the energy partitioning at the five
different photolysis energies as a function of the total energy.
For comparison, Table 2 also contains the energy partitioning
data from the work of Zhang et al.,26 who photolyzed HNCO
at 193.3 nm in a molecular beam and measured the kinetic
energy distribution of the H atom fragment. Our results at lower
photolysis energies are generally consistent with theirs. We
find approximately 35% of the energy in internal degrees of
freedom at Eavail ) 1950, 2250, and 3030 cm-1, with ap-
proximately 30% in vibration and 5% in rotation. There is
proportionally less energy in vibration and more in rotation
closer to threshold. The major of the vibrational energy appears
in the bend, and less than 2% appears in excited stretching states.
The results of the 193.3-nm photolysis show a slightly larger
fraction of the total energy, 70%, in translation.
V. Summary
We report the direct observation of the internal energy
distribution of the NCO fragment from photolysis of HNCO at
five different energies near the threshold for cleavage of the
N-H bond. We probe the NCO fragment by LIF on its A r
X transition.18,36,37 The photolysis yield spectrum, in which we
fix the probe laser wavelength on a transition in the NCO
fragment and scan the photolysis laser through the UV absorp-
tion bands of the parent HNCO, shows clear rotational and
vibrational structure, particularly at lower photolysis energies.
In separate double-resonance experiments,41 we have measured
the lower limit to this lifetime to be 3 ps at 38 500 cm-1. Here,
we excite five different vibrational bands up to 3030 cm-1
excess photolysis energy. A Boltzmann distribution fits the
rotational-state populations well, and the overall rotational
excitation is much less than predicted by a simple statistical
model. Vibrational energy appears primarily in the bending
mode of the NCO fragment, with the population in excited
bending states exceeding that in nearly isoenergetic excited
stretching states by a factor of 6 or more. The small population
in the symmetric stretch shows that the dissociation is very
nonstatistical, and the preference for partitioning of energy into
the NCO bending mode is a result of the dissociation from a
bent NCO group in the A′′ state of HNCO to a linear NCO
fragment where the force in the NCO bending coordinate
strongly excites the bend. Overall, about 65% of the total energy
appears in relative translation of the fragments at all of these
near-threshold photolysis energies, while approximately 30%
goes into vibration (dominated by the bending excitation) and
5% into rotation.
Table 2 also presents a classical calculation49 of the energy
partitioning in the impulsive limit. Because this model does
not account for the forces along the NCO bending coordinate
during the dissociation, it predicts little vibrational excitation.
Indeed, we find very little energy in the stretching vibrations
but substantial amounts in the bending vibration. The calculated
fraction of rotational excitation in NCO is small (2% of the
available energy) as a result of the small torque that the light H
atom exerts on the NCO fragment. The experimental rotational
excitation is also small for large excess photolysis energies but
never decreases to the impulsive limit even at very high energies,
where the limiting experimental value is 4%. Our results, which
are for a lower available energy than those of Zhang et al.,26
have a slightly smaller fraction of energy in translation and larger
fraction in rotation, consistent with the dissociation being
somewhat less strongly impulsive from the region of the
dissociative potential that our excitation reaches.
Acknowledgment. We thank the Division of Chemical
Sciences of the Office of Basic Energy Sciences of the
Department of Energy for supporting this work. In addition,
we thank Professor Paul Dagdigian for graciously supplying
us with his programs to simulate the LIF spectrum of the NCO
fragment. These programs, along with the prescription for
extracting vibrational populations from the spectral simulations,
were essential in our analysis.

Internal Energy Distribution of the NCO Fragment
J. Phys. Chem., Vol. 100, No. 19, 1996 7955
(25) Yi, W.; Bersohn, R. Chem. Phys. Lett. 1993, 206.
References and Notes
(26) Zhang, J.; Dulligan, M.; Wittig, C. J. Chem. Phys. 1995.
(27) Dixon, R. N.; Kirby, G. H. Trans. Faraday Soc. 1968, 64, 2002.
(28) Rabalais, J. W.; McDonald, J. R.; McGlynn, S. P. J. Chem. Phys.
1969, 51, 5103.
(1) Perry, R. A.; Siebers, D. L. Nature 1986, 324, 657.
(2) Miller, J. A.; Bowman, C. T. Int. J. Chem. Kinet. 1991, 23, 289.
(3) Mui, J. Y. P.; Back, R. A. Can. J. Chem. 1963, 41, 826.
(4) Wooley, W. D.; Back, R. A. Can. J. Chem. 1967, 46, 295.
(5) Bradley, J. N.; Gilbert, J. R.; Svejda, P. Trans. Faraday Soc. 1968,
64, 911.
(6) Okabe, H. J. Chem. Phys. 1970, 53, 3507.
(7) Russic, B.; Berkowitz, J. J. Chem. Phys. 1994, 100, 4498.
(8) Spiglanin, T. A.; Perry, R. A.; Chandler, D. W. J. Phys. Chem.
1986, 90, 6184.
(9) Brown, S. S.; Berghout, H. L.; Crim, F. F. J. Chem. Phys. 1995,
102, 8440.
(10) Brown, S. S.; Metz, R. B.; Berghout, H. L.; Crim, F. F. J. Chem.
Phys., to be submitted.
(11) Likar, M. D.; Baggot, J. E.; Sinha, A.; Ticich, T. M.; Vander Wal,
R. L.; Crim, F. F. J. Chem. Soc., Faraday Trans. 2 1988, 84, 1483.
(12) Crim, F. F. Annu. ReV. Phys. Chem. 1993, 44, 397.
(13) Ram, R. S.; Bernath, P. F. J. Opt. Soc. Am. B 1986, 3, 1170.
(14) Ubachs, W.; Meyer, G.; Ter Meullen, J. J.; Dymanus, A. J. Mol.
Spectrosc. 1986, 115, 88.
(15) Hack, W.; Mill, T. J. Mol. Spectrosc. 1990, 144, 358.
(16) Charlton, T. R.; Okamura, T.; Thrush, B. A. Chem. Phys. Lett. 1982,
89, 98.
(17) Copeland, R. A.; Crosley, D. R. Can. J. Phys. 1984, 62, 1488.
(18) Patel-Misra, D.; Sauder, D. G.; Dagdigian, P. J. J. Chem. Phys.
1990, 93, 5448.
(19) Drozdoski, W. S.; Baronovski, A. P.; McDonald, J. R. Chem. Phys.
Lett. 1979, 64, 421.
(29) Rabalais, J. W.; McDonald, J. M.; Scherr, V.; McGlynn, S. P. Chem.
ReV. 1971, 71, 73.
(30) Yamada, K. J. Mol. Spectrosc. 1980, 79, 323.
(31) Du, K. Y.; Setser, D. W. 1988.
(32) Sauder, D. G.; Patel-Misra, D.; Dagdigian, P. J. J. Chem. Phys.
1991, 95, 1696.
(33) Herzberg, G.; Teller, E. Z. Phys. Chem. B 1993, 21, 410.
(34) Renner, R. Z Phys. 1934, 92, 172.
(35) Hougen, J. T. J. Chem. Phys. 1961, 36, 519.
(36) Dixon, R. N. Phil. Trans. R. Soc. London, Ser. A 1960, 252, 165.
(37) Bolman, P. S. H.; Browm, J. M.; Carrington, A.; Kopp, I.; Ramsay,
D. A. Proc. R. Soc. London, Ser. A 1975, 343, 17.
(38) Woodward, D. R.; Fletcher, D. A.; Brown, J. M. Mol. Phys. 1987,
62, 517.
(39) Wu, M.; Northrup, F. J.; Sears, T. J. J. Chem. Phys. 1992, 97, 4583.
(40) Hemmerling, B.; Vervloet, M. Mol. Phys. 1993, 78, 1423.
(41) Brown, S. S.; Berghout, H. L.; Crim, F. F. J. Chem. Phys., in
preparation.
(42) Silver, J. A.; Dimpfl, W. L.; Brophy, J. H.; Kinsey, J. L. J. Chem.
Phys. 1976, 65, 1811.
(43) Linhard, M. Z. Anorg. Allg. Chem. 1938, 236, 200.
(44) Ashby, R. A.; Werner, R. L. J. Mol. Spectrosc. 1965, 18, 184.
(45) East, A. L. L.; Johnson, C. S.; Allen, W. D. J. Chem. Phys. 1993,
98, 1299.
(46) Levine, R. D.; Bernstein, R. B. Acc. Chem. Res. 1974, 7, 393.
(47) Levine, R. D.; Bernstein, R. B. Molecular Reaction Dynamics and
Chemical ReactiVity; Oxford University: New York, 1987.
(48) Steiner, D. A.; Wishah, K. A.; Polo, S. R.; McCubbin, T. K. J.
Mol. Spectrosc. 1979, 76, 341.
(20) Fujimoto, G. T.; Umstead, M. E.; Lin, M. C. Chem. Phys. 1982,
65, 197.
(21) Spiglanin, T. A.; Perry, R. A.; Chandler, D. W. J. Chem. Phys.
1987, 87, 1568.
(22) Spiglanin, T. A.; Chandler, D. W. J. Chem. Phys. 1987, 87, 1577.
(23) Spiglanin, T. A.; Chandler, D. W. Chem. Phys. Lett. 1987, 141,
428.
(49) Tuck, A. F. J. Chem. Soc., Faraday Trans. 2 1977, 73, 689.
(24) Bohn, B.; Stuhl, F. J. Phys. Chem. 1993, 97, 4891.
JP952667R