Photodissociation dynamics of the singlet and triplet states of the NCN radical

Article abstract of DOI:10.1063/1.479751

The spectroscopy and photodissociation dynamics of the NCN radical have been investigated by fast beam photofragment translational spectroscopy. The B³∑_u^-←X ³∑_g^-, c¹Π_u,←a¹Δ _g, and d¹Δ_u ←-a¹Δ_g transitions were examined. The major dissociation products for the B³∑_u^- and c¹Π_u states are N₂(X¹∑_g⁺) + C(³P), while the d¹Δ_u state dissociates to N₂(X¹∑_g^-+)+ C(¹D). The dissociation channel, N(⁴S) + CN(X²∑¹) is observed for the B3∑_u^- state at photon energies greater than 4.9 eV, where it comprises ≈25±10% of the total signal. At all photon energies, the photofragment translational energy distributions show a resolved progression corresponding to the vibrational excitation of the N₂ photofragment. The rotational distributions of the molecular fragments suggest that the dissociation pathway for the N₂ loss channel involves a bent transition state while the N+CN photofragments are produced via a linear dissociation mechanism. The P(E_T) distributions provide bond dissociation energies of 2.54±0.030 and 4.56±0.040 eV for the N₂ and CN loss channels, respectively, yielding ΔH_{f,0 K}(NCN)=4.83±0.030 eV.

Full text of DOI:10.1063/1.479751

Photodissociation dynamics of the singlet and triplet states of the NCN radical
Ryan T. Bise, Hyeon Choi, and Daniel M. Neumark
Citation: The Journal of Chemical Physics 111, 4923 (1999); doi: 10.1063/1.479751
View online: http://dx.doi.org/10.1063/1.479751
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/111/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
State-selected imaging of HCCO radical photodissociation dynamics
J. Chem. Phys. 128, 134301 (2008); 10.1063/1.2831788
Photodissociation dynamics of the HCNN radical
J. Chem. Phys. 124, 204303 (2006); 10.1063/1.2196890
Ultraviolet photodissociation dynamics of the SH radical
J. Chem. Phys. 123, 054330 (2005); 10.1063/1.1961565
193-nm photodissociation of acryloyl chloride to probe the unimolecular dissociation of CH 2 CHCO radicals and
CH 2 CCO
J. Chem. Phys. 120, 4223 (2004); 10.1063/1.1644096
Photodissociation of triplet and singlet states of the CCO radical
J. Chem. Phys. 108, 4070 (1998); 10.1063/1.475839
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

JOURNAL OF CHEMICAL PHYSICS
VOLUME 111, NUMBER 11
15 SEPTEMBER 1999
Photodissociation dynamics of the singlet and triplet states
of the NCN radical
Ryan T. Bise, Hyeon Choi, and Daniel M. Neumarka)
Department of Chemistry, University of California, Berkeley, California 94720 and Chemical Sciences
Division, Lawrence Berkeley National Laboratories, Berkeley, California 94720
͑Received 22 April 1999; accepted 18 June 1999͒
The spectroscopy and photodissociation dynamics of the NCN radical have been investigated by fast
Ϫ
Ϫ
3
3
1
1
˜
˜
˜
beam photofragment translational spectroscopy. The B ⌺u —X ⌺g ,
˜
c 1⌸u—
˜a ⌬g , and d ⌬u
Ϫ
1
3
c 1⌸u states
˜
—
˜
a ⌬g transitions were examined. The major dissociation products for the B ⌺u and
˜
ϩ
ϩ
1
3
1
1
1
˜
˜
˜
are N2(X ⌺g )ϩC( P), while the d ⌬u state dissociates to N2(X ⌺g )ϩC( D). The dissociation
channel, N( S)ϩCN(X ⌺ ) is observed for the B ⌺Ϫu state at photon energies greater than 4.9 eV,
where it comprises Ϸ25Ϯ10% of the total signal. At all photon energies, the photofragment
translational energy distributions show a resolved progression corresponding to the vibrational
excitation of the N2 photofragment. The rotational distributions of the molecular fragments suggest
that the dissociation pathway for the N2 loss channel involves a bent transition state while the
NϩCN photofragments are produced via a linear dissociation mechanism. The P(ET) distributions
provide bond dissociation energies of 2.54Ϯ0.030 and 4.56Ϯ0.040 eV for the N2 and CN loss
channels, respectively, yielding ⌬Hf,0 K͑NCN͒ϭ4.83Ϯ0.030 eV. © 1999 American Institute of
Physics. ͓S0021-9606͑99͒00535-8͔
4
2
ϩ
3
˜
˜
I. INTRODUCTION
Fig. 1 with the corresponding dominant electronic configu-
rations given below:
Photodissociation of small polyatomic molecules pro-
vides a rich probe of the potential energy surfaces involved
in bond rupture. Detailed measurements such as photofrag-
ment yield spectra, translational energy distributions, and
product branching ratios allow for a careful examination of
the dissociation mechanism.1 While there have been numer-
ous photodissociation studies of stable closed-shell mol-
ecules, far fewer studies of open-shell radicals have been
performed. Our laboratory has demonstrated the ability to
generate a well-characterized source of radicals via photode-
tachment of negative ions, allowing us to perform photodis-
sociation experiments on reactive open-shell species. Here,
we report the photodissociation dynamics for triplet and sin-
glet electronic states of the NCN radical.
The NCN free radical has been proposed as an important
combustion intermediate since its emission was observed in
hydrocarbon flames by Jennings and Linnett.2 It has been
suggested by Smith et al.3 that the NCN radical plays a sig-
nificant role in the combustion of nitramine propellant mol-
ecules, given the large extent of C–N bonding in these mol-
ecules. Additionally, ultraviolet emission studies of the
comet Brorosen–Metcalf suggest that NCN is present in
comets and is a source of CN radicals.4
Ϫ
ϩ
2
4
2
3
3
3
1
1
˜
˜
... 3␴ ͒ 1␲ ͒ 1␲ ͒
X ⌺g ,˜a ⌬g ,b ⌺g
͑
͑
͑
u
u
g
1
4
3
c 1⌸u
˜
... 3␴ ͒ 1␲ ͒ 1␲ ͒
A ⌸u ,
˜
͑
͑
͑
u
u
g
Ϫ
2
3
3
1
˜
˜
... 3␴ ͒ 1␲ ͒ 1␲ ͒
B ⌺u ,d ⌬u .
͑
͑
͑
u
u
g
According to Walsh’s rules,5 all of these electronic states are
predicted to be linear.
Rotationally resolved ultraviolet absorption spectra were
6
3
3
˜
˜
measured by Travis and Herzberg for the A ⌸u—X ⌺g
band, and by Kroto7,8 for the
˜
c 1⌸u—
˜
a 1⌬g band, confirming
that these states are linear. Kroto observed a number of
Ϫ
Ϫ
3
3
˜
˜
higher lying transitions attributed to the B ⌺u —X ⌺g and
˜
a 1⌬g bands.9 Milligan, Jacox, and Bass observed
1
˜
d ⌬u—
Ϫ
Ϫ
3
3
˜
˜
this B ⌺u —X ⌺g band as well as a number of infrared
absorptions in matrix isolation studies.10,11
In the matrix work, depletion of the NCN radical was
observed at wavelengths shorter than 280 nm ͑4.42 eV͒; on
the basis of secondary reaction products within the matrix,
ϩ
1
˜
the authors proposed that NCN dissociates to N2(X ⌺g )
ϩC(3P) in this wavelength range, which overlaps the
Ϫ
Ϫ
B ⌺u —X ⌺g band. Smith et al.3 measured the radiative
3
3
˜
˜
lifetime for the vibrationless level and a number of vibronic
3
The NCN radical has attracted a great deal of attention
from both spectroscopists and theorists who have been par-
ticularly concerned with its bonding and geometry. The
ground and relevant excited states for this study are shown in
˜
bands of the A ⌸u state, determining the radiative lifetime
of the vibrationless level to be 183 ns. The radiative lifetimes
show very little variation with increased vibrational excita-
tion, implying that this state does not dissociate.
More recently, McNaughton and co-workers12 measured
a high resolution Fourier transform infrared spectrum of the
NCN antisymmetric stretch band in the gas phase. Brown
a͒
Author to whom correspondence should be addressed. Electronic mail:
dan@radon.cchem.berkeley.edu
0021-9606/99/111(11)/4923/10/$15.00
4923
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

4924
J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Bise, Choi, and Neumark
FIG. 2. Fast radical beam translational spectrometer. The dotted line sepa-
rates the radical production section on the left from the photodissociation
experiment on the right.
spin multiplicity.31 A photodetachment energy slightly above
the electron affinity produces the NCN radical in the ground
Ϫ
3
˜
X ⌺g state exclusively, allowing us to examine the spec-
troscopy and dynamics of triplet dissociative electronic
states. A higher photodetachment energy can then be used to
FIG. 1. Energy level and correlation diagram for the NCN radical along a
Ϫ
3
populate both the X ⌺g and ˜
a 1⌬g states, the latter provid-
˜
ϩ
1
1
˜
˜
C2v dissociation coordinate. The energies of the
˜
a
⌬g , b
⌺
, A 3⌸u ,
g
ing access to the dissociative states in the singlet manifold.
Overall, we have examined the photodissociation dynamics
Ϫ
3
1
1
˜
˜
⌸u , and d ⌬
B
⌺
,
˜c
states are based upon experimental work dis-
u
u
cussed in the text. The 1A1 , 3A2 , and 3B2 states are from ab initio calcu-
lations ͑Ref. 24͒ and the product state energies are from JANAF thermo-
chemical tables ͑Ref. 42͒.
Ϫ
Ϫ
3
3
1
1
˜
˜
˜
of the B ⌺u —X ⌺g ,
˜
c 1⌸u—
˜
a ⌬g , and d ⌬u—
˜
a 1⌬g
electronic bands obtaining structured photodissociation
cross-sections, product branching ratios, and detailed internal
energy distributions of the photofragments.
and co-workers13–16 used laser-induced fluorescence and la-
ser magnetic resonance to refine the spectroscopic constants
II. EXPERIMENT
˜
˜
of the X and A states and obtain estimates of the bend fre-
quencies in the two states.
The fast beam photofragment translational spectrometer,
Fig. 2, has been described in detail elsewhere;32–34 only a
brief description will follow. In this experiment, we generate
a clean source of neutral radicals by mass-selectively photo-
detaching a beam of stable negative ions. The neutral radi-
cals are then photodissociated by a second laser.
The photoelectron spectrum of NCNϪ measured by Clif-
ford et al.17 showed the photodetachment process to be ver-
tical, giving rise to no vibrational excitation of the neutral
radical upon photodetachment. Higher energy photoelectron
spectroscopy studies performed by Taylor et al.18 located the
To generate a sufficient number density of NCNϪ ions,
we made a slight modification to our pulsed electric dis-
charge source35 by inserting a reservoir containing cyana-
mide (H2NCN) between the pulsed molecular beam valve
and the pulsed electric discharge. Our adaptation of this
source has been described in detail previously.18 Neat O2 at a
stagnation pressure of ϳ3 atm is expanded through a pulsed
molecular beam valve into the reservoir containing the cy-
anamide and finally through the discharge, generating CNϪ,
NCNϪ, HNCNϪ, and NCOϪ ions. Analysis of photofrag-
ment yield spectra obtained in this study indicates that this
source produces anions with rotational and vibrational tem-
peratures of 50 and 200 K, respectively.
ϩ
Ϫ
1
3
˜
˜
˜
a 1⌬g and b ⌺g states relative to the X ⌺g ground state,
thereby determining the singlet–triplet splitting, E(
˜
a 1⌬g
Ϫ
3
˜
ϪX ⌺g )ϭ1.010Ϯ0.010 eV.
Theoretical studies of the geometry and bonding of the
NCN radical showed the ground state to be linear with the
central carbon doubly bonded to both nitrogen atoms.17,19–23
Martin et al. have performed ab initio studies on both the
NCN and CNN radicals for linear, cyclic, and bent
geometries24 finding the barrier to isomerization to lie ap-
proximately 3 eV above the NCN ground state. This is con-
sistent with the experimental observation of both isomers as
distinct species.25–30
The negative ions generated in the source region are ac-
celerated to 8 keV and separated temporally by a Bakker
time-of-flight ͑TOF͒ mass spectrometer, which induces very
little kinetic energy spread of the ions.36,37 The ion of interest
is selectively photodetached by a pulsed laser. Based upon
the photoelectron spectrum of Taylor et al.,18 an excimer-
pumped dye laser operating at 2.82 eV was used to photode-
The NCN radical provides as rich system for photodis-
sociation studies. It has several low-lying singlet and triplet
excited states above the dissociation limit as well as multiple
energetically accessible product states, as shown in Fig. 1.
Furthermore, the photodecomposition of NCN to products
Ϫ
Ϫ
3
3
3
˜
˜
N2ϩC( P) from the linear B ⌺u —X ⌺g transition ob-
served by Milligan and co-workers10 implies that the disso-
ciation mechanism is complex, possibly involving bent or
cyclic electronic states.
Ϫ
3
tach NCNϪ to produce the neutral X ⌺g state exclusively,
˜
while 4.03 eV light ͑308 nm͒ from a XeCl excimer laser was
Ϫ
3
In this study, a mass selected beam of NCN radicals is
generated from the photodetachment of NCNϪ ions and sub-
sequently photodissociated. We recently demonstrated the
ability to probe transitions not only from the ground elec-
tronic state but also from low-lying excited states of different
a 1⌬g and X ⌺g states of the
˜
used to populate both the ˜
neutral. Any undetached ions are deflected out of the beam
path.
In the dissociation region, the neutrals are intersected by
the frequency-doubled output of an excimer-pumped dye la-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Photodissociation dynamics of NCN
4925
ser with a bandwidth of 0.3 cmϪ1. A fraction of the neutrals
absorb and dissociate yielding photofragments detected di-
rectly by either the TOF or TPS ͑time and position sensing͒
microchannel plate detector assemblies in Fig. 2. An alumi-
num strip is positioned at the center of each detector to pro-
hibit any undissociated radicals from impacting the detector,
so that any observed signal is entirely from the recoiling
photofragments.
TABLE I. Peak energies, assignments, progression spacings, and widths for
Ϫ
Ϫ
3
3
˜
˜
u —X ⌺
the observed transitions in the B
⌺
PFY spectrum.
g
Transition energy
(n)Ϫ (nϪ1)
FWHM
v
v
͑cmϪ1
͒
Assignment
000
͑cmϪ1
͒
͑cmϪ1
168
116
48
͒
33 488
101
34 555
35 597
1067
1042
102
Two types of experiments are performed. First, the spec-
troscopy of the dissociative electronic states is examined by
measuring the total flux of photofragments arriving at the
retractable TOF detector, located at 0.68 m from the disso-
ciation laser, as a function of photon energy. The resulting
photofragment yield ͑PFY͒ spectra is complementary to ab-
sorption and fluorescence measurements. The current study
examined the photolysis of NCN between 30 000 to 43 500
211120
103
35 714
36 738
37 740
38 759
39 774
40 793
41 801
36 634
37 657
38 654
39 698
40 707
41 762
42 736
1037
1023
997
86
40
30
21
20
¯
20
211130
104
1024
1002
1019
1015
1019
1008
211140
105
cmϪ1
.
Once the spectroscopy of the dissociative states has been
examined, a second type of experiment, which probes the
dissociation dynamics, can be performed. In this detection
scheme, both photofragments from a single parent radical are
detected in coincidence using a time-and-position sensitive
͑TPS͒ detector of the type developed by de Bruijn and Los.38
Our implementation of this detection scheme has been de-
scribed in detail elsewhere.32,33 The TPS detector records the
positions and difference in arrival time of the two photofrag-
ments from a single dissociation event. This information is
then used to determine the masses of the fragments, their
relative translational energy ET , and the scattering angle ␪
between the relative velocity vector and the electric vector of
the polarized dissociation laser ͑parallel to the ion beam
axis͒. The photofragment mass resolution is m/⌬mϷ10
while the translational energy resolution for these experi-
ments is ⌬ET /ETϭ3.0%.
211150
106
1044
1009
1055
974
211160
107
211170
108
211180
109
from reaching the detector, while high-energy recoil frag-
ments with values of ␪ close to 90° miss the detector. The
raw translation energy distributions are therefore normalized
by the calculated detector acceptance function, D(ET ,␪),
which has been described in detail by Continetti et al.39
Due to the geometry of the TPS detector, which is 40
mm in diameter with an 8 mm beam block located at the
center, the detection efficiency of photofragments depends
upon their values of ␪ and ET . The beam block, in addition
to blocking undissociated radicals, prevents fragments of low
translational energy or with values of ␪ close to 0° to 180°
III. RESULTS
؊
؊
3
3
˜
˜
A. Photofragment yield spectra B ⌺u —X ⌺g
transitions
Ϫ
Ϫ
3
3
˜
˜
The PFY signal for the B ⌺u —X ⌺g band is shown in
Fig. 3, covering 33 000 to 43 000 cmϪ1. The spectrum shows
a progression of predissociative resonances spaced by ap-
proximately 1050 cmϪ1. The positions, assignments, and
widths of these transitions are listed in Table I. This progres-
sion was observed previously in the ultraviolet absorption
studies of Kroto9 in the gas phase, and by Milligan, Jacox,
and Bass10 in matrix isolation studies, and was assigned to
the symmetric stretch (1n0) progression. The extended pro-
gression and change in frequency from 1259 cmϪ1 in the
ground state40 to 1050 cmϪ1 for the excited state is consistent
with promotion of an electron from the 1␲ to the 1␲ or-
u
g
bital.
A weaker 1050 cmϪ1 progression is observed Ϸ100
cmϪ1 to the blue of the 1n0 progression. Travis and Herzberg6
have shown for second row triatomics that the bending fre-
Ϫ
Ϫ
g
3
3
˜
˜
⌺
FIG. 3. Photofragment yield spectrum of the B
⌺
u —X
band obtained
at a photodetachment energy of 2.82 eV. The vibrational combs denote
symmetric stretch 10n and sequence band 10n 211 progressions.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

4926
J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Bise, Choi, and Neumark
FIG. 5. PFY spectrum from 36 300 to 37 800 cmϪ1. The dashed line indi-
cates features observed at a photodetachment energy of 2.82 eV. The solid
line denotes additional features seen at a photodetachment energy of 4.03
1
FIG. 4. Photofragment yield ͑PFY͒ spectrum of the
band.
˜c
1⌸u—
˜a
⌬
origin
g
eV. These features a, b, and c are assigned to the d 1⌬u—
˜a
⌬
band.
1
˜
g
quency increases with the number of electrons in the ␲
g
3
˜
orbital. The bending frequency of the A ⌸u state, which
yields minimal vibrational excitation of the neutral radical,
allowing the anion to be photodetached well above threshold
while maintaining a cold neutral vibrational distribution.
Any new features in the PFY spectrum are therefore a result
also has three electrons in the outermost ␲ orbital, was
g
found to be 96 cmϪ1 larger than the ground state bend fre-
quency, estimated to be 437.7 cmϪ1 14
.
On this basis, we
assign this progression to the 1n0 211 sequence band and esti-
of transitions from the ˜
a 1⌬g electronic state.
The intense transition centered around 30 045 cmϪ1
agrees well with the rotationally resolved spectrum obtained
by Kroto,7 which he assigns to the ͑000–000͒ transition of
Ϫ
3
˜
mate the bend frequency of the B ⌺u state to be 540
Ϯ20 cmϪ1
.
Neither our photodissociation studies nor the previously
mentioned absorption studies have been able to locate any
the
˜
c 1⌸u—
˜
a 1⌬g band. Additional weaker features seen at
30 025, 30 045, and 30 130 cmϪ1 are consistent with band-
Ϫ
Ϫ
3
3
˜
˜
transitions of the B ⌺u —X ⌺g band below the apparent
origin at 33 488 cmϪ1. In agreement with the observations of
Kroto,9 the two lowest energy peaks were found to be re-
markably broad, ϳ168 and 116 cmϪ1 in width. The peak
widths gradually narrow as the photon energy is increased.
The 130 transition is anomalously broad showing essentially
three peaks with the most intense feature located to the blue.
Although we performed scans with a laser bandwidth of
0.2 cmϪ1 and a step size of 0.1 cmϪ1, we were unable to
heads observed by Kroto and are assigned to the ͑100–100͒
ϩ
1
1
1
1⌸u— ⌬g , ͑010–010͒ 1⌺g — ⌸u , and ͑010–010͒
⌬
g
1
— ⌽u vibronic transitions, respectively. The rotational con-
tour of the ͑000–000͒ band is characteristic of a perpendicu-
lar transition moment with an intense Q-branch and weaker
P- and R-branches. The Q-branches of the ͑100–100͒ 1⌸u
ϩ
— ⌬g and ͑010–010͒ 1⌺g — ⌸u bands are heavily over-
lapped by the P- and R-branches of the ͑000–000͒ transition,
respectively. Using the previously determined rotational con-
stants, the best fit to this rotational contour is achieved with
a rotational temperature of 50 K.
1
1
˜ ˜
resolve any rotational structure for the B—X band and there-
fore cannot determine the excited state lifetime from line-
width measurements. The relative intensities of the vibra-
tional features in the PFY spectra are similar to those for the
matrix absorption experiments of Milligan et al.,10 suggest-
ing that the quantum yield for photodissociation is Ϸ1.
Figure 5 shows the PFY spectrum for the wavelength
region between 36 200 to 37 800 cmϪ1 obtained at 4.03 eV
Ϫ
Ϫ
3
3
˜
˜
photodetachment energy. In addition to the B ⌺u —X ⌺g
transitions seen in Fig. 3, a number of new transitions appear
labeled a, b, and c. These transitions have been observed
previously in the gas phase absorption studies of Kroto and
3
3
˜
˜
Despite extensive efforts to observe the A ⌸u—X ⌺g
transitions, we were unable to detect photodissociation signal
from the origin or higher energy vibronic transitions associ-
ated with this band.
9
1
˜
co-workers and were assigned to the d ⌬u—
˜
a 1⌬g band
based on kinetic information and complicated vibrational
structure. The most intense features, a and c, belong to a
progression of Ϸ1020 cmϪ1 which is attributed to the sym-
metric stretch. The weaker feature b most likely involves
bend excitation. Our photodetachment technique confirms
1
1
B. Photofragment yield spectra, c˜ ⌸u—a˜ ⌬g and
1
1
˜
d ⌬u—a˜ ⌬g transitions
that these transitions originate from the ˜
a 1⌬g state.
Figure 4 shows the PFY spectrum from 30 000 to 30 150
cmϪ1 when the photodetachment energy is 4.03 eV. The re-
cent photoelectron spectrum taken by Taylor et al.18 showed
that a photodetachment energy of 4.03 eV will populate both
C. Translational energy distributions
Ϫ
3
the X ⌺g and ˜
a 1⌬g electronic states with an intensity ratio
˜
of approximately 2:1. Since the anion and neutral have es-
At the photon energies explored in this study, three pos-
sible dissociation channels are accessible:
sentially the same geometry, photodetachment of NCNϪ
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Photodissociation dynamics of NCN
4927
ϩ
g
1
3
˜
͑I͒
N X ⌺ ͒ϩC P͒ ⌬Hrxn,0ϭ2.541Ϯ0.030 eV,
͑
2
͑
h␷
Ϫ
g
ϩ
g
3
1
1
˜
˜
͑II͒
͑III͒
NCN X ⌺ ͒˜ N2͑X ⌺ ͒ϩC͑ D͒ ⌬Hrxn,0ϭ3.801Ϯ0.030 eV,
͑
ͭ
2
ϩ
4
˜
CN X ⌺ ͒ϩN S͒ ⌬Hrxn,0ϭ4.563Ϯ0.050 eV.
͑
͑
The above heats of reaction ͑at 0 K͒ represent a refinement
of previous values and were determined directly from this
work ͑see below͒.
All three channels are accessible following excitation of
the 1n0 , nу4 transitions in Fig. 3. For each dissociation
event, the photofragment mass ratios are calculated from the
relative recoil distance of the coincident photofragments
from the center of the neutral beam. This enables us to
largely but not completely distinguish CϩN2 from NϩCN
products, since our product mass resolution is only ϳ10. For
the resulting N2 vibrational distributions listed in Table II.
While the P(ET) distributions show highly structured fea-
tures at high translational energy, the low kinetic energy re-
gions (ETϽ0.8 eV) of these spectra are relatively broad and
unresolved.
From our product mass analysis, of which Fig. 6 is an
example, we expect significant NϩCN contributions at low
n
˜ ˜
ET for the 10 , nу6 B—X transitions. The NϩCN P(ET)
distributions for the low ET regions of the 160, 107, and 109
transitions ͑the first two are from data at a flight length of 2
m͒ are shown in Fig. 8. The 160, 107, and 109 transitions yield
P(ET) distributions with sharp narrow peaks (FWHM
Ϸ100 meV) with maxima located at 0.19, 0.31, and 0.56 eV,
respectively. The peak positions shift with increased excita-
tion energy; consistent with opening of the N(4S)
˜ ˜
data collected at a 1 m flight length, the B—X transitions
with photon energies less than 5.1 eV produce a mass ratio
of 12:28, indicating CϩN2 products. The observed mass dis-
tribution is significantly broadened at translational energies
less than 0.7 eV for the 190 transition, implying that the
NϩCN channel contributes at low ET . Increasing the flight
length to 2 m improves the detection efficiency of low trans-
lational energy fragments. At this flight length, the 160 and 107
7
2
ϩ
˜
ϩCN(X ⌺ ) channel. The P(ET) distributions for the 10
and 190 transitions display additional peaks at Ϸ250 meV
lower ET ; these are attributed to vibrational excitation of the
CN fragment. There is signal extending to higher transla-
tional energies than energetically accessible for NϩCN prod-
ucts. This is caused by our limited photofragment mass reso-
lution that cannot completely exclude N2ϩC events from the
NϩCN distribution.
transitions exhibit
a
photofragment mass ratio of
14:26 ͑NϩCN͒ at translational energies Ͻ0.3 and Ͻ0.45 eV,
respectively and a mass ratio of 12:28 at higher translational
energies, as shown in Fig. 6. The photofragment mass ratio is
1
1
˜
12:28 for the
all ET .
˜
c 1⌸u—
˜
a ⌬g and d ⌬u—
˜
a 1⌬g transitions at
The NϩCN photofragments have relatively low recoil
energies ͑Ͻ0.6 eV͒ and are therefore preferentially detected
P(ET) distributions for the CϩN2 channel from several
˜ ˜
B—X transitions are shown in Fig. 7; these distributions
were taken at a flight length of 1 m. The distributions display
a sharp onset at high translational energy and well-resolved
structure with peaks separated by approximately 290 meV,
corresponding to excitation of the N2 product vibration. The
full width at half maximum ͑FWHM͒ of each peak is ap-
proximately 150 meV with an asymmetric tail extending to-
ward low translational energy. The assignment of the struc-
ture of the P(ET) distributions is presented in Sec. IV with
FIG. 7. P(ET) distributions for the N2ϩC(3P) channel resulting from
FIG. 6. Fragment mass distribution for ETϽ0.35 eV for the 170 transition
͑open circles͒ and ETϾ0.5 for the 106 transition ͑open triangles͒. The data
was collected with a 2 m flight length.
Ϫ
Ϫ
B
⌺
transitions. The dashed vertical lines denote EmT ax for prod-
3
3
˜
˜
u —X ⌺
g
ucts. Data are shown with open circles, while the results of a fit to the
product internal energy distribution are shown with a solid line.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

4928
J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Bise, Choi, and Neumark
TABLE II. Vibrational distributions of the N2 and CN photofragments obtained from a fit of the P(ET)
Ϫ
Ϫ
g
3
3
1
1
, d ⌬u—a 1⌬g , and
˜c
1⌸u—
˜a ⌬ tran-
g
˜
˜
˜
˜
distributions, Figs.7–10, resulting from excitation of B
sitions of NCN,
⌺
u —X
⌺
N2 vibrational distributon
ϭ0 ϭ1 ϭ2 ϭ3 ϭ4 ϭ5 ϭ6
CN vibrational distribution
ϭ0 ϭ1
Photon
Energy ͑eV͒ ͑%͒ ͑%͒ ͑%͒ ͑%͒ ͑%͒ ͑%͒ ͑%͒
v
v
v
v
v
v
v
v
v
Transition
͑%͒
͑%͒
Ϫ
Ϫ
g
3
3
˜
B
˜
⌺
⌺
u —X
000
4.153
4.283
4.414
4.542
4.670
4.790
4.927
5.049
5.299
22
22
22
26
11
15
15
9
22
19
13
20
17
9
11
10
12
26
31
13
12
12
9
16
14
14
30
28
30
12
12
9
10
12
9
101
102
12
24
33
24
17
12
19
103
6
15
32
18
13
17
104
105
106
12
30
15
100
60
60
¯
40
40
107
109
14
1
1
˜
d
1⌬u—
˜a
⌬
⌬
g
c
4.646
18
16
17
14
21
14
1⌸u—
˜a
˜
c
g
000
3.725
3.735
17
13
19
12
17
19
10
17
22
19
15
20
211
We have also obtained P(ET) distributions for ˜
c 1⌸u
over the more energetic N2ϩC products when a 2 m flight
distance is used, preventing a precise determination of the
branching ratio. By including the relative detector acceptan-
ces for the two product channels at both 1 and 2 m flight
lengths, we estimate the relative yield of the CNϩN channel
for the 160, 107, and 190 transitions to be Ϸ25Ϯ10%.
1
1
—
˜
a ⌬g and d ⌬u—
˜
a 1⌬g transitions. The P(ET) distribu-
˜
tion resulting from the
˜
c 1⌸u—
˜
a 1⌬g ͑000–000͒ transition at
a photon energy of 3.725 eV is shown in Fig. 9. At this
photon energy, the maximum translational energy (EmT ax) is
2.19 and 0.93 eV for C(3P) and C(1D) products, respec-
tively. These values are indicated by the dashed vertical
lines. The onset for dissociation clearly occurs near 2.19 eV
indicating that spin-forbidden triplet products are the primary
photodissociation products. The P(ET) distribution is re-
˜ ˜
markably similar to the B—X transitions, showing resolved
structure at high ET with peak spacing of Ϸ290 meV and a
FWHM of Ϸ150 meV while the distribution is significantly
less structured for ETϽ0.8 eV. It is not obvious from the
distribution if C(1D) products are present. However, the
FIG. 8. P(ET) distributions analyzed with a 14:26 mass ratio for the 160, 107,
1
FIG. 9. P(ET) distribution for N2ϩC products from the
˜c
1⌸u—
˜a
⌬
g
Ϫ
Ϫ
and 190 transitions of the B
⌺
band. The distributions for the 106
3
3
˜
˜
u —X ⌺
000–000 band. The two dashed vertical lines ͑— — —͒ and ͑-- - --͒ denote
EmT ax for C(3P) and C(1D) products, respectively. Data are shown with open
circles, while the results of a fit to the product internal energy distribution
are shown with a solid line.
g
and 170 transitions were collected using a 2 m flight distance between the
dissociation laser and the detector while a 1 m flight length was used for the
190 transition.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Photodissociation dynamics of NCN
4929
The anisotropy parameter, ␤, can range from ϩ2 to Ϫ1,
corresponding to cos2 ␪ and sin2 ␪ distributions, respectively.
The angular distributions for the N2ϩC product channel for
all three electronic transitions were found to be nearly iso-
tropic with ␤Ϸ0, while the NϩCN channel was found to
have an anisotropic angular distribution with ␤ϭ0.9. The
positive ␤ parameter is consistent with the expected parallel
Ϫ
Ϫ
3
3
˜
˜
transition dipole moment for the B ⌺u —X ⌺g band.
IV. ANALYSIS
The P(ET) distributions in Figs. 7–10 demonstrate how
the excess energy above the dissociation threshold is distrib-
uted between the photofragments. The energy balance for
NCN photodissociation to N2ϩC is described by Eq. ͑2͒,
1
FIG. 10. P(ET) distribution for N2ϩC products from a d 1⌬u—
˜a ⌬ tran-
g
˜
sition at 37,476 cmϪ1. Data are shown with open circles, while the results of
a fit to the product internal energy distribution are shown with a solid line.
The dashed vertical line represents EmT ax for C(1D) products.
h␯ϩEint͑NCN͒ϩEelec͑NCN͒
ϭD0 NCN͒ϩE ϩE ͑N ͒ϩE ͑N ͒ϩE
C͒,
͑
elec
͑2͒
͑
T
v
2
R
2
relative intensities of features with ETϾ0.9 eV are similar to
those for ETϽ0.9 eV, indicating that the C(1D) products are
at best a minor channel; this point is revisited in Sec. IV.
where h␯ is the photon energy, Eint͑NCN͒ is the average
rotational energy of the parent radical, Eelec͑NCN͒ is the ini-
tial electronic state of the radical, D0 is the dissociation en-
ergy, ET is the measured translational energy, Ev(N2) and
ER(N2) are the N2 vibrational and rotational energies, re-
spectively, and Eelec͑C͒ is the atomic state of carbon. An
analogous equation can be written for the CNϩN channel.
The parent rotational temperature of 50 K yields Eint͑NCN͒
Ϸ33 cmϪ1. D0(NCN) for the N2 loss channel can be ex-
tracted from these distributions by determining EmT ax , the
translational energy corresponding to photofragments with
zero internal energy.
1
Excitation of the ͑010–010͒ 1⌬g— ⌽u transition yields
a P(ET) distribution ͑not shown͒ which is very similar to
that observed for the ͑000–000͒ transition, in that there is a
structured progression of Ϸ290 meV for ETϾ0.8 eV which
becomes less structured for ETϽ0.8 eV. The FWHM of the
structured features are Ϸ180 meV, somewhat broader than
those for the ͑000–000͒ transition. There does not appear to
be any significant difference in the C(1D):C(3P) product
branching ratios for the two transitions.
1
˜
a 1⌬g transition
Although EmT ax is not always obvious from a P(ET) dis-
tribution, it can be readily determined from the distributions
in Fig. 7 from the steep falloff in intensity toward the high
energy side marked by the vertical dashed lines for each
photon energy. Each P(ET) distribution then provides an
independent measurement of D0 , all of which agree within
our experimental resolution of 30 meV. An average of all of
these values yields D0ϭ2.54Ϯ0.03 eV and ⌬Hf,0 K(NCN)
ϭ4.83Ϯ0.03 eV. The latter agrees with the values of 4.89
Ϯ0.22 eV and 4.69Ϯ0.13 eV from JANAF thermochemical
tables42 and recent ab initio calculations,17 respectively,
but our error bar is smaller. Contamination of the P(ET)
distributions for the CNϩN channel by N2ϩC products
prevents an exact measurement of EmT ax for the CNϩN
channel directly from the P(ET) distribution. From our
experimentally determined ⌬Hf,0 K(NCN) and literature
values for ⌬Hf,0 K(CN)ϭ4.51Ϯ0.02 eV ͑Ref. 43͒ and
˜
The P(ET) distribution for the d ⌬u—
at an excitation energy of 4.646 eV is shown in Fig. 10. The
Ϫ
1
3
˜
singlet–triplet splitting, E(˜a ⌬gϪX ⌺g ), is 1.010
Ϯ0.010 eV.18 Thus, the total energy of the radical above the
ground electronic state after excitation is 5.656 eV with ex-
pected EmT ax values of 3.12 and 1.86 for C(3P) and C(1D)
products, respectively. The onset for dissociation occurs near
1.8 eV indicating that the C(1D)ϩN2 products are the domi-
nant channel. A small fraction ͑Ͻ5%͒ of the total signal
appears at translational energies greater than 1.8 eV. This
˜
feature is attributed to a hot band absorption from the X state
which dissociates to C(3P)ϩN2. The general features of the
˜ ˜
The distribution is highly structured with a peak spacing of
approximately 290 meV and a FWHM of about 150 meV.
P(ET) distribution are similar to those for the B—X band.
˜
However, in contrast to the P(ET) distributions for the B
˜
˜
⌬Hf,0 K(N(4S))ϭ4.8796Ϯ0.0010 eV,42
D0(N–CN)ϭ4.56Ϯ0.04 eV.
we
calculate
—X and
˜c—
˜
a transitions, the P(ET) distribution for the d
—
˜a transition exhibits resolved structure at translational en-
ergies as low as 0.2 eV. It is worth emphasizing here that
The P(ET) distributions for CϩN2 products and the
NϩCN products exhibit resolved structure corresponding to
the vibrational excitation of the molecular fragment. In an
attempt to determine the vibrational distribution of the di-
atomic fragment, we have fit each P(ET) distribution to a
series of distribution functions fv separated by the vibronic
energy levels of the diatomic fragment ͑N2 or CN͒44 with the
total distribution given by Eq. ͑3͒:
4
2
ϩ
˜
although the N( S)ϩCN(X ⌺ ) channel is energetically
accessible by more than 1 eV, only a 12:28 mass ratio was
observed for the photofragments.
The photofragment angular distributions are described
by Eq. ͑1͒:41
I ␪͒ϭ1/ 4␲͒ 1ϩP cos ␪͒ .
͑1͒
͑
͑
͓
͑
͔
2
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

4930
J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Bise, Choi, and Neumark
v
Ј
F E ͒ϭ
␣ f E Ϫ h␯Ϫ ␻ ϪD Ϫ⌬͒,T,␦ . ͑3͒
͓ ͑ v ͔
v v
T 2 0
͑
͚
T
vϭ0
The individual distribution functions, fv , are Gaussians
convoluted with a Boltzmann distribution. The offset ⌬ and
rotational temperature T were adjusted to fit the peak and
width of each vibrational feature. Attempts to model the
P(ET) distributions with a single rotational temperature for
each vibrational product state proved less successful. The
Ϫ
3
˜
P(ET) distributions CϩN2 mass channel for the B ⌺u
Ϫ
—X ⌺g transition was fit with the C(3P) products while
3
˜
the
˜
c 1⌸u—
˜
a 1⌬g transitions were fit with both C(3P) and
C(1D) products. No substantial improvement to the fit was
obtained by including the C(1D) products. We estimate that
the C(1D) products comprise р10% of the total photodisso-
n ˜ ˜
1⌸u—
FIG. 11. The average Npeak values for all the 10 B—X transitions in Fig. 7
1
͑squares͒, as well as the Npeak values for the
˜c
˜a
⌬
origin ͑triangles͒
˜
a 1⌬g transition was fit using
1
g
˜
ciation signal. The d ⌬u—
and transition c of the d 1⌬u—
˜a
⌬
band ͑circles͒ are plotted versus the N2
1
˜
C(1D) products exclusively.
g
vibrational quantum number.
˜ ˜
The vibrational distributions for the B—X transitions,
Table II, do not reveal any clear trends regarding the extent
of vibrational excitation for the N2 fragment as a function of
excitation energy. There also does not appear to be any sig-
nificant difference in the vibrational distributions resulting
from the different NCN excited electronic states. The P(ET)
distributions for the 170 and 190 transitions show population of
volved in the dissociation of this triatomic radical. The
˜
Ϫ
3
B ⌺u state yields N2ϩC(3P) products at all photon ener-
gies explored. At photon energies greater than 4.9 eV, the
Ϫ
3
2
ϩ
4
˜
˜
B ⌺u state also produces CN(X ⌺ )ϩN( S) products as a
minor channel Ϸ25Ϯ10%. C(3P) products are the dominant
both the ϭ0 and ϭ1 states of the CN fragment with an
v
v
dissociation channel ͑у90%͒ for the
˜
c 1⌸u state even though
intensity ratio ϭ0: ϭ1 of Ϸ3:2.
v
v
C(1D) products are energetically available by more than 1.8
We have also extracted the rotational distribution of the
molecular fragment for each product vibrational state. The
offset ⌬ and rotational temperature T in Eq. ͑3͒ are used to
determine Npeak , the rotational quantum number correspond-
ing to the peak of each distribution, fv . A unique Npeak value
for each N2 product vibrational state is then obtained for
every P(ET) distribution. The P(ET) distributions for the 10n
1
1
˜
eV, while the d ⌬u state dissociates to C( D) products ex-
clusively. The B ⌺u and ˜
c 1⌸u states possess very similar
Ϫ
3
˜
P(ET) distributions with well resolved vibrational features
for ETϾ0.8 eV and broad unresolved structure at lower
1
˜
translational energies. The d ⌬u state in comparison exhibits
resolved vibrational structure for translational energies as
low as 0.2 eV. Finally, the CN fragment is produced with
substantially less rotational excitation (Npeakϭ10) than the
N2 photofragment (Npeakϭ20–40), implying that the mecha-
nisms for these two dissociation channels are distinct.
The ground and all excited electronic states accessed in
this study are linear, but the dominance of CϩN2 photoprod-
ucts indicates that the photodissociation mechanism must in-
volve cyclic and/or bent states. The extensive vibrational and
rotational excitation of the N2 product are also consistent
with dissociation through a bent or cyclic state with the N–N
bond forming and C–N bonds breaking at large N–N dis-
tances leading to an extended vibrational progression in the
N2 photoproduct. Rotational excitation of the N2 fragment is
generated from parent molecular rotation and from the
torque applied at the transition state. The trend of increasing
rotational excitation with N2 vibrational quantum number for
˜ ˜
B—X transitions yield very similar Npeak values ͑Ϯ3 quanta͒
for the same N2 product vibrational state ͑i.e., Npeak is inde-
pendent of n͒. However, the Npeak values exhibit a pro-
nounced dependence on the vibrational quantum number of
the N2 fragment.
In Fig. 11, the average Npeak values for the 1n0 B—X
˜ ˜
transitions for each N2 product state are plotted versus the N2
vibrational quantum number; the Npeak values for the ˜
c 1⌸u
1
1
—
˜
a ⌬g (000–000) and d ⌬u—
˜
a 1⌬g transitions are also
˜
shown. For all three transitions, the N2 fragment displays
substantial rotational excitation with distributions peaking
between 20–40 quanta of rotation depending on the vibra-
tional quantum number of the fragment and the initial elec-
Ϫ
tronic transition. The Npeak values for the B ⌺u and ˜
c 1⌸u
3
˜
states increase substantially for higher vibrational states of
1
˜
the N2 fragment, while the Npeak values for the d ⌬u state
Ϫ
3
the B ⌺u and ˜
c 1⌸u states is similar to that seen by Con-
˜
are generally lower than for the other two states and vary
much less with N2 vibrational state. The CN photofragment
was found to have a rotational distribution described by a
Boltzmann temperature of 700 K with NpeakϷ10.
tinetti et al.32 for the photodissociation of N3 and by Zare
and co-workers45 for the photodissociation of ICN.
Formation of product from the two end atoms of a tri-
atomic, particularly a linear one, is relatively uncommon as it
requires a low energy cyclic ͑or at least strongly bent͒ tran-
sition state. Such a transition state has been proposed in the
photodecomposition of OClO which yields the lowest energy
products, O2ϩCl, as a minor channel ͑р4%͒.46–48 The larg-
V. DISCUSSION
The product state energy distributions for the fragmen-
tation of NCN reveal that complicated dynamics are in-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Photodissociation dynamics of NCN
4931
est relative yields of this channel result from excitation of the
bend and symmetric stretch modes, implying a concerted dis-
sociation involving a transition state of approximately C2v
symmetry. Ab initio calculations49–53 for O3 have located
minima for ring structures of D3h symmetry, however, the
poor Franck–Condon overlap with the ground state and a
large barrier to dissociation via C2v symmetry suggests that
this pathway does not contribute. The dissociative ionization
of CS2 yields Sϩ2 as minor channel ͑Ͻ9%͒ from the decay of
the linear excited states of CSϩ2 .54,55 The translational en-
ergy distributions of the Sϩ2 ion56 are structureless single
peaks with only 10%–26% of the available energy projected
into translation indicating extensive vibrational and rota-
tional excitation of the Sϩ2 fragment, consistent with a disso-
ciation pathway involving a cyclic intermediate.
increased vibrational excitation. The ˜
c 1⌸u state has been
rotationally resolved by Kroto7 with an instrument resolution
of 0.1 cmϪ1 and appears to be instrument limited, indicating
that the lifetime of the
time, spin-forbidden products, and P(ET) distribution sug-
gest that the
c 1⌸u state first intersystem crosses to the
˜
c 1⌸u state is Ͼ50 ps. The long life-
˜
Ϫ
3
˜
B ⌺u state prior to dissociation.
1
˜
The d ⌬u state, in contrast to the
˜
c 1⌸u state, leads to
spin-allowed C(1D) dissociation products. Although the
4
2
ϩ
˜
N( S)ϩCN(X ⌺ ) products are energetically accessible by
1.08 eV, these spin-forbidden products are not observed. The
1
˜
P(ET) distribution which results from excitation to the d ⌬u
state, Fig. 8, exhibits an extended vibrational progression of
the N2 photofragment. However, Fig. 11 shows that the de-
1
˜
Martin et al.24 have performed ab initio calculations for
linear, bent, and cyclic structures for NCN and CNN. Their
calculations within the C2v point group have located a local
minimum with 3A2 symmetry and a transition state with 3B1
symmetry at 3.01 and 4.76 eV, respectively, above the linear
gree of N2 rotational excitation is lower for d ⌬u dissocia-
Ϫ
3
tion than for B ⌺u or ˜
c 1⌸u state dissociation. In addition,
˜
the rotational excitation of the N2 fragment is approximately
the same for all N2 vibrational states, in contrast to the
Ϫ
3
B ⌺u and ˜
c 1⌸u states. It therefore appears that the triplet
˜
3
˜
and singlet surfaces have different topologies.
ground state of NCN. As shown in Fig. 1, the A ⌸u state
Ϫ
3
3
In NCN predissociation, the N2 rotational distribution is
determined largely by the energy and geometry of the tran-
sition state, and by the torque exerted on the N2 in the exit
channel. The lower rotational excitation for C(1D)ϩN2
products can therefore be attributed to a number of factors.
The experiments of Milligan and Jacox11 indicate that that
C(1D) atoms react readily in a nitrogen matrix to form NCN
while C(3P) atoms do not, implying that the C(1D) disso-
ciation products do not have a barrier to recombination,
while a substantial barrier exists for the triplet channel. This
is consistent with the extensive work on reactions of atomic
oxygen, in which O(1D) but not O(3P) is known to undergo
a variety of insertion reactions with no barrier.57 However,
one must be careful in applying these considerations to our
experiment, because recombination presumably occurs on
the lowest energy singlet or triplet surface of NCN whereas
the dissociation dynamics in our experiment are launched
from an electronically excited state.
˜
adiabatically correlates to the B1 state while the B ⌺u state
correlates to the lower lying 3A2 state. Both the 3A2 and 3B1
states correlate to ground state products in C2v symmetry. If
we assume that the dissociation proceeds via C2v symmetry,
3
˜
then the A ⌸u state cannot dissociate since it does not cor-
relate to the lower energy 3A2 state and does not have
enough energy to access the 3B1 state. On the other hand, the
Ϫ
3
˜
B ⌺u state has a relatively low energy pathway through a
cyclic transition state to C(3P)ϩN2 products. This simple
Ϫ
3
˜
picture therefore explains why only the B ⌺u state dissoci-
ates, even though both states lie above the CϩN2 asymptote.
Furthermore, the calculations predict a N–N bond distance
3
of 1.737 Å for the A2 state, consistent with the extensive
vibrational excitation of the N2 photofragment (Re
ϭ1.098 Å͒.
3
Martin et al. also found a less symmetric A (C ) tran-
Љ
s
sition state through uphill following of the bending mode of
3
˜
CNN. This state lies slightly lower in energy than the A ⌸u
While the dominant photodissociation products observed
in this study are N2ϩC, the CNϩN channel is observed from
3
˜
state, but the absence of A ⌸u state dissociation suggests
Ϫ
3
that this transition state does not play a role in the dissocia-
tion dynamics.
˜
the B ⌺u state at photon energies Ͼ4.9 eV. Both the
CN( ϭ0) and vibrationally excited CN( ϭ1) products are
v
v
We next consider the dissociation mechanism of the
observed, consistent with an expected large change in CN
˜
c 1⌸u state. Martin et al. report only one singlet minimum
Ϫ
3
˜
bond distance from the B ⌺u state of the NCN radical to the
1
energy structure with symmetry A1 located 1.57 eV above
2
ϩ
˜
CN(X ⌺ ) fragment. The rotational distribution peaks at
Nϭ10, a value less than half that observed for the N2 loss
channels. The limited rotational excitation and positive pho-
tofragment anisotropy, ␤ϭ0.9, are consistent with a linear
dissociation pathway. Freysinger et al.58 have examined the
diabatic and adiabatic electronic correlations for the linear
states of isoelectronic Nϩ3 ion into products NϩϩN2 and
Nϩ2 ϩN,58 which are isoelectronic with the CϩN2 and
the ground state, which correlates adiabatically with the
˜
c 1⌸u state. However, since spin-forbidden C(3P) products
are clearly the dominant dissociation channel from the ˜
c 1⌸u
state, it is unlikely that this structure is involved in the dis-
sociation mechanism. The C(3P):C(1D) branching ratio
does not appear to change significantly for the ͑010–010͒
1
1⌬g— ⌽u transition suggesting that this upper state is also
1
not strongly coupled to the A1 state.
Ϫ
3
˜
The formation of C(3P) products from the c 1⌸u state
˜
NϩCN products observed in this study. The B ⌺u state
4
2
ϩ
˜
clearly indicates that the dissociation mechanism involves
correlates adiabatically to N( S)ϩCN(X ⌺ ) and corre-
intersystem crossing to a triplet surface. The P(ET) distribu-
lates diabatically to the higher energy products N(2D)
Ϫ
3
2
4
2
ϩ
c 1⌸u states are noticeably similar
˜
˜ ˜
ϩCN(A ⌸), while the N( S)ϩCN(X ⌺ ) products corre-
late diabatically to a higher-lying ⌺u state. The crossing of
tions for the B ⌺u and
˜
Ϫ
3
with both displaying an increase in rotational excitation with
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06

4932
J. Chem. Phys., Vol. 111, No. 11, 15 September 1999
Bise, Choi, and Neumark
these two diabatic surface will likely produce a conical in-
tersection leading to a barrier to dissociation along the adia-
batic surface. The CNϩN channel is first observed at photon
energies which exceed D0(N–CN) by more than 0.39 eV
providing an upper limit for the barrier height.
13 S. A. Beaton, Y. Ito, and J. M. Brown, J. Mol. Spectrosc. 178, 99 ͑1996͒.
14 S. A. Beaton and J. M. Brown, J. Mol. Spectrosc. 183, 347 ͑1997͒.
15 K. D. Hensel and J. M. Brown, J. Mol. Spectrosc. 180, 170 ͑1996͒.
16 M. Wienkoop, W. Urban, and J. M. Brown, J. Mol. Spectrosc. 185, 185
͑1997͒.
17 E. P. Clifford, P. G. Wenthold, W. C. Lineberger, G. A. Petersson, and G.
B. Ellison, J. Phys. Chem. A 101, 4338 ͑1997͒.
It is surprising that the CNϩN channel, which appears to
dissociate via a linear pathway, does not become the domi-
nant channel once it is energetically accessible, comprising
only 25Ϯ10% of the total dissociation signal. The domi-
nance of the N2ϩC(3P) channel is further proof that the
18 T. R. Taylor, R. T. Bise, K. R. Asmis, and D. M. Neumark, Chem. Phys.
Lett. 301, 413 ͑1999͒.
19 Z. L. Cai, G. H. Sha, C. H. Zhang, and M. B. Huang, Theochem 85, 303
͑1992͒.
20 G. Berthier and L. Kurdi, Acad. Sci., Paris, Ser. II ͑Mec., Phys., Chim.,
Sci. Univers. Sci. Terre͒ 299, 1171 ͑1984͒.
Ϫ
3
˜
coupling of the B ⌺u state to bent or cyclic states is highly
21 H. U. Suter, M. B. Huang, and B. Engels, J. Chem. Phys. 101, 7686
͑1994͒.
efficient.
22 C. Thomson, J. Chem. Phys. 58, 841.
23 G. R. Williams, Chem. Phys. Lett. 25, 602 ͑1974͒.
24 J. M. L. Martin, P. R. Taylor, J. P. Francois, and R. Gijbels, Chem. Phys.
Lett. 226, 475 ͑1994͒.
VI. CONCLUSIONS
Ϫ
3
The photodissociation dynamics of the B ⌺u ,
˜
˜
c 1⌸u ,
˜
25 V. E. Bondybey and J. H. English, J. Chem. Phys. 67, 664 ͑1977͒.
26 M. E. Jacox, J. Mol. Spectrosc. 72, 26 ͑1978͒.
1
and d ⌬u states of NCN radical have been investigated by
Ϫ
27 T. D. Goldfarb and G. C. Pimentel, J. Am. Chem. Soc. 82, 1865 ͑1959͒.
28 M. C. Curtis, A. P. Levick, and P. J. Sarre, Laser Chem. 9, 359 ͑1988͒.
29 G. W. Robinson and M. McCarty, Jr., J. Am. Chem. Soc. 82, 1859 ͑1960͒.
30 E. Wasserman, L. Barash, and W. A. Yager, J. Am. Chem. Soc. 87, 2075
͑1965͒.
3
˜
fast beam photofragment spectroscopy. Both the B ⌺u and
˜
c 1⌸u states photodissociate to ground state products
3
1
1
˜
N2ϩC( P) while the d ⌬u state dissociates to N2ϩC( D).
Based on the identity of the products, one expects dissocia-
tion to occur through a cyclic or strongly bent transition
state; this is consistent with the extensive vibrational and
rotational excitation of the N2 photoproduct from all three
31 H. Choi, D. H. Mordaunt, R. T. Bise, T. R. Taylor, and D. M. Neumark,
J. Chem. Phys. 108, 4070 ͑1998͒.
32 R. E. Continetti, D. R. Cyr, D. L. Osborn, D. J. Leahy, and D. M. Neu-
mark, J. Chem. Phys. 99, 2616 ͑1993͒.
33 D. J. Leahy, D. L. Osborn, D. R. Cyr, and D. M. Neumark, J. Chem. Phys.
103, 2495 ͑1995͒.
Ϫ
states. For the B ⌺u and ˜
c 1⌸u states, the rotational excita-
3
˜
tion appears to increase with vibrational excitation. The simi-
larity in the P(ET) distributions and the production of the
34 D. L. Osborn, H. Choi, D. H. Mordaunt, R. T. Bise, D. M. Neumark, and
C. M. Rohlfing, J. Chem. Phys. 106, 3049 ͑1997͒.
35 D. L. Osborn, D. J. Leahy, D. R. Cyr, and D. M. Neumark, J. Chem. Phys.
104, 5026 ͑1996͒.
Ϫ
3
˜
same photoproducts for the B ⌺u and
that the two states dissociate along the same surface, requir-
ing that the
c 1⌸u state undergoes intersystem crossing to the
˜
c 1⌸u states suggest
36 J. M. B. Bakker, J. Phys. E 6, 785 ͑1973͒.
˜
37 J. M. B. Bakker, J. Phys. E 7, 364 ͑1974͒.
Ϫ
3
38 D. P. de Bruijn and J. Los, Rev. Sci. Instrum. 53, 1020 ͑1982͒.
39 R. E. Continetti, D. R. Cyr, D. L. Osborn, D. J. Leahy, and D. M. Neu-
mark, J. Chem. Phys. 99, 2616 ͑1993͒.
˜
B ⌺u state prior to dissociation.
Finally, the CNϩN product channel was observed from
Ϫ
3
˜
the B ⌺u state for photon energies greater than 4.9 eV com-
prising Ϸ25Ϯ10% of the total dissociation signal. The rota-
tional distribution and anisotropic angular distribution, ␤
ϭ0.9, suggests that these photoproducts are formed via a
linear dissociation pathway.
40 This gas phase value for
is determined from ϭ1254 cmϪ1 ͑Ref. 3͒
vЉ
vЈ
and
Ϫ
ϭϪ5 cmϪ1 ͑Ref. 6͒.
vЈ vЉ
41 R. N. Zare, Mol. Photochem. 4, 1 ͑1972͒.
42 M. W. Chase, Jr., C. A. Davies, J. R. Downey, Jr., D. J. Frurip, R. A.
McDonald, and A. N. Syverud, J. Phys. Chem. Ref. Data Suppl. 1, 14
͑1985͒.
43 Y. Huang, S. A. Barts, and J. B. Halpern, J. Phys. Chem. 96, 425 ͑1992͒.
44 K. P. Huber and G. Herzberg, Constants of Diatomic Molecules ͑Van
Nostrand Reinhold, New York, 1979͒.
ACKNOWLEDGMENTS
This research is supported by the Director, Office of En-
ergy and Research, Office of Basic Energy Sciences, Chemi-
cal Sciences Division, of the U.S. Department of Energy
under Contract No. DE-AC03-76F00098. We would like to
thank Alexandra Hoops for assistance in data collection.
45 M. A. O’Halloran, H. Joswig, and R. N. Zare, J. Chem. Phys. 87, 303
͑1987͒.
46 H. F. Davis and Y. T. Lee, J. Phys. Chem. 96, 5681 ͑1992͒.
47 E. Bishenden and D. J. Donaldson, J. Chem. Phys. 99, 3129 ͑1993͒.
48 E. Bishenden and D. J. Donaldson, J. Chem. Phys. 101, 9565 ͑1994͒.
49 S. Xantheas, G. J. Atchity, S. T. Elbert, and K. Ruedenberg, J. Chem.
Phys. 94, 8054 ͑1991͒.
50 S. Xantheas, S. T. Elbert, and K. Ruedenberg, J. Chem. Phys. 93, 7519
͑1990͒.
1 L. J. Butler and D. M. Neumark, J. Phys. Chem. 100, 12801 ͑1996͒.
2 K. R. Jennings and J. W. Linnett, Trans. Faraday Soc. 56, 1737 ͑1960͒.
3 G. P. Smith, R. A. Copeland, and D. R. Crosley, J. Chem. Phys. 91, 1987
͑1989͒.
51 T. Muller, S. S. Xantheas, H. Daschel, R. J. Harrison, N. Jaroslaw, R.
Shepard, G. S. Kedziora, and H. Lischa, Chem. Phys. Lett. 293, 72 ͑1998͒.
52 A. Banichevich and S. D. Peyerimhoff, Chem. Phys. 174, 93 ͑1993͒.
53 M. Braunstein, P. J. Hay, R. L. Martin, and R. T Pack, J. Chem. Phys. 95,
8239 ͑1991͒.
4 C. R. O’Dell, C. O. Miller, A. L. Cochran, W. D. Cochran, C. B. Opal,
and E. S. Barker, Astrophys. J. 368, 616 ͑1991͒.
5 A. D. Walsh, J. Chem. Soc. 1953, 2266.
54 B. Brehm, J. H. D. Eland, R. Frey, and A. Kustler, Int. J. Mass Spectrom.
Ion Phys. 12, 212 ͑1973͒.
6 G. Herzberg and D. N. Travis, Can. J. Phys. 42, 1658 ͑1964͒.
7 H. W. Kroto, Can. J. Phys. 45, 1439 ͑1967͒.
55 J. H. D. Eland and J. Berkowitz, J. Chem. Phys. 70, 5151 ͑1979͒.
56 M. Hamdan, F. M. Harris, and J. H. Beynon, Int. J. Mass Spectrom. Ion
Processes 74, 303 ͑1986͒.
8 H. W. Kroto, J. Chem. Phys. 44, 831 ͑1966͒.
9 H. W. Kroto, T. F. Morgan, and H. H. Sheena, Trans. Faraday Soc. 66,
2237 ͑1970͒.
57 M. Alagia, N. Balucani, P. Casavecchia, D. Stranges, and G. G. Volpi, J.
Chem. Soc., Faraday Trans. 91, 575 ͑1995͒.
10 D. E. Milligan, M. E. Jacox, and A. M. Bass, J. Chem. Phys. 43, 3149
͑1965͒.
11 D. E. Milligan and M. E. Jacox, J. Chem. Phys. 45, 1387 ͑1966͒.
12 D. McNaughton, G. F. Metha, and R. Tay, Chem. Phys. 198, 107 ͑1995͒.
58 W. Freysinger, F. A. Khan, P. B. Armentrout, P. Tosi, O. Dmitriev, and D.
Bassi, J. Chem. Phys. 101, 3688 ͑1994͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.252.67.66 On: Sun, 21 Dec 2014 05:57:06