Probing E/Z isomerization on the C10H8 potential energy surface with ultraviolet population transfer spectroscopy

Article abstract of DOI:10.1021/ja908103u

The excited-state dynamics of phenylvinylacetylene (1-phenyl-1-buten-3-yne, PVA) have been studied using laser-induced fluorescence spectroscopy, ultraviolet depletion spectroscopy, and the newly developed method of ultraviolet population transfer spectroscopy. Both isomers of PVA (E and Z) show a substantial loss in fluorescence intensity as a function of excitation energy. This loss in fluorescence was shown to be due to the turn-on of a nonradiative process by comparison of the laser-induced fluorescence spectrum to the ultraviolet depletion spectrum of each isomer, with a threshold 600 cm ^-1 above the electronic origin in Z-PVA and 1000 cm^-1 above the electronic origin in E-PVA. Ab initio and density functional theory calculations have been used to show that the most likely source of the nonradiative process is from the interaction of the ππ* state with a close lying πσ* state whose minimum energy structure is bent along the terminal CCH group. Ultraviolet population transfer spectroscopy has been used to probe the extent to which excited-state isomerization is facilitated by the interaction with the πσ* state. In ultraviolet population transfer spectroscopy, each isomer was selectively excited to vibronic levels in the S1 state with energies above and below the threshold for fluorescence quenching. The ultraviolet-excited populations are then recooled to the zero point levels using a reaction tube designed to constrain the supersonic expansion and increase the collision cooling capacity of the expansion. The new isomeric distribution was detected in a downstream position using resonant-2-photon ionization spectroscopy. From these spectra, relative isomerization quantum yields were calculated as a function of excitation energy. While the fluorescence quantum yield drops by a factor of 50-100, the isomerization quantum yields remain essentially constant, implying that the nonradiative process does not directly involve isomerization. On this basis, we postulate that isomerization occurs on the ground-state potential energy surface after internal conversion. In these experiments, the isomerization to naphthalene was not observed, implying a competition between isomerization and cooling on the ground-state potential energy surface.

Full text of DOI:10.1021/ja908103u

Published on Web 01/12/2010
Probing E/Z Isomerization on the C H Potential Energy
10
8
Surface with Ultraviolet Population Transfer Spectroscopy
†
Josh J. Newby, Christian W. M u¨ ller, Ching-Ping Liu, and Timothy S. Zwier*
Department of Chemistry, Purdue UniVersity, West Lafayette, Indiana 47907-2084
Received September 23, 2009; E-mail: zwier@purdue.edu
Abstract: The excited-state dynamics of phenylvinylacetylene (1-phenyl-1-buten-3-yne, PVA) have been
studied using laser-induced fluorescence spectroscopy, ultraviolet depletion spectroscopy, and the newly
developed method of ultraviolet population transfer spectroscopy. Both isomers of PVA (E and Z) show a
substantial loss in fluorescence intensity as a function of excitation energy. This loss in fluorescence was
shown to be due to the turn-on of a nonradiative process by comparison of the laser-induced fluorescence
-1
spectrum to the ultraviolet depletion spectrum of each isomer, with a threshold 600 cm above the electronic
-
1
origin in Z-PVA and 1000 cm above the electronic origin in E-PVA. Ab initio and density functional theory
calculations have been used to show that the most likely source of the nonradiative process is from the
interaction of the ππ* state with a close lying πσ* state whose minimum energy structure is bent along the
terminal CCH group. Ultraviolet population transfer spectroscopy has been used to probe the extent to
which excited-state isomerization is facilitated by the interaction with the πσ* state. In ultraviolet population
1
transfer spectroscopy, each isomer was selectively excited to vibronic levels in the S state with energies
above and below the threshold for fluorescence quenching. The ultraviolet-excited populations are then
recooled to the zero point levels using a reaction tube designed to constrain the supersonic expansion and
increase the collision cooling capacity of the expansion. The new isomeric distribution was detected in a
downstream position using resonant-2-photon ionization spectroscopy. From these spectra, relative
isomerization quantum yields were calculated as a function of excitation energy. While the fluorescence
quantum yield drops by a factor of 50-100, the isomerization quantum yields remain essentially constant,
implying that the nonradiative process does not directly involve isomerization. On this basis, we postulate
that isomerization occurs on the ground-state potential energy surface after internal conversion. In these
experiments, the isomerization to naphthalene was not observed, implying a competition between
isomerization and cooling on the ground-state potential energy surface.
I. Introduction
incorporate radical reactions shared by analogous models of
combustion processes, where soot formation could follow similar
pathways.
One of the areas in which both combustion and atmospheric
models are still under active development is in the pathways
from benzene and its simple derivatives to naphthalene, an-
thracene, and larger PAHs. Among the postulated pathways are
The mechanism of formation of aromatic molecules is
important in a variety of contexts ranging from combustion
processes to planetary atmsopheres.
7
,9,10
1-8
As such, there are many
previous studies focusing on reactions that lead from small
hydrocarbons to polycyclic aromatic hydrocarbons (PAHs),
including studies spanning a wide range of temperatures and
pressures, and encompassing hundreds of molecules of different
sizes and functional complexity. Titan, one of the moons of
Saturn, has a photochemically driven atmosphere rich in
hydrocarbons, including benzene, with a wide array of larger
PAHs also anticipated. The models of Titan’s atmosphere often
those that utilize a series of hydrogen-abstraction-C
2
H
2
-addition
7
,10,11
(
HACA) steps,
a recently postulated ethynyl addition
1
2
mechanism (EAM), and cyclopentadienyl recombination to
1
3
form napthalene.
In the context of such models, it is important to understand
how the aromatic-forming reactions turn on or off depending
on the reaction conditions, which can differ substantially
between combustion (high T, thermally driven), planetary
atmospheres (low T, intermediate pressures), and interstellar
†
Present address: Division of Medical Engineering Research, National
Health Research Institutes, 35 Keyan Road, Zhunan 350, Taiwan.
(
(
(
(
1) Bauschlicher, C. W.; Ricca, A. Chem. Phys. Lett. 2000, 326, 283.
2) Filley, J.; McKinnon, J. T. Combust. Flame 2001, 124, 721.
3) Frenklach, M. Phys. Chem. Chem. Phys. 2002, 4, 2028.
4) Goldaniga, A.; Faravelli, T.; Ranzi, E. Combust. Flame 2000, 122,
(9) Krasnopolsky, V. A. Icarus 2009, 201, 226.
3
50.
(10) Wilson, E. H.; Atreya, S. K. J. Geophys. Res. [Planets] 2004, 109.
(11) Atreya, S. Science 2007, 316, 843.
(
5) Roesler, J. F.; de Tessan, M. A.; Montagne, X. Chemosphere 2001,
2, 823.
6) Schulz, K.; Hofmann, J.; Zimmermann, G. Liebigs Ann.-Recueil 1997,
4
(12) Mebel, A. M.; Kislov, V. V.; Kaiser, R. I. J. Am. Chem. Soc. 2008,
130, 13618.
(
2
535.
(13) Marinov, N. M.; Pitz, W. J.; Westbrook, C. K.; Castaldi, M. J.; Senkan,
S. M. Modeling of aromatic and polycyclic aromatic hydrocarbon
formation in premixed methane and ethane flames; 4th International
Congress on Toxic Combustion Byproducts, Berkeley, CA, 1995.
(
7) Wilson, E. H.; Atreya, S. K. Planet. Space Sci. 2003, 51, 1017.
(8) Wilson, E. H.; Atreya, S. K.; Coustenis, A. J. Geophys. Res. [Planets]
2
003, 108.
10.1021/ja908103u  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 1611–1620 9 1611

A R T I C L E S
Newby et al.
space (very low T, few collisions). For example, the recombina-
can serve as the basis for state-specific studies carried out in
both directions, E-PVAfZ-PVA and Z-PVAfE-PVA. In both
PVA isomers, previous studies show a turn on of a nonradiative
process in regions similar to that observed in stilbene. In these
studies, it was shown that the most plausible and dominant
nonradiative process involved the crossing of the optically
accessed ππ* state with a πσ* state involving bending of the
CtCH away from planarity; however, little could be inferred
tion of propargyl radicals (2C
3
H
3
fC
6 6
H ) leads efficiently to
1
4,15
benzene in flames
isomer wells under the conditions of Titan’s stratosphere.
Similar issues are likely to arise in naphthalene formation, where
even less is known about the C10 potential energy surface
and all the various entry ways onto the surface or isomerization
pathways once C10 is formed. In the context of the photo-
chemical models, one intriguing possibility worth exploring is
whether pathways that lead initially to other C10 isomers might
produce photochemically active C10 intermediates that could
but are trapped in intermediate C H
6 6
1
6
H
8
H
8
23,24
about the role of isomerization in the probed energy regime.
H
8
In recent years, we have introduced a series of experimental
methods that combine isomer-specific laser excitation early in
a supersonic expansion with a recooling step prior to interroga-
tion, thereby enabling the isomeric products to be selectively
detected downstream in the collision-free region of the
H
8
subsequently undergo photoisomerization to naphthalene.
Such issues form one motivation for the studies reported here.
Phenylvinylacetylene (1-phenyl-1-buten-3-yne, PVA) is a struc-
2
5-31
tural isomer of naphthalene (C10
H
8
), which was recently
expansion.
These prior studies employed one of two
2
5,26
identified as a photoproduct of the reaction between UV excited
excitation methods, infrared (IR)
pumping (SEP).
tional levels within the first 4000 cm of the ground electronic
state, necessitating removal of internal energies in this range
prior to interrogation. Most of the work to date has thus focused
or stimulated emission
Both methods involve excitation of vibra-
1
7
25,28
diacetylene (C
4
H
2
*) and styrene. PVA is also a logical
-1
intermediate formed in a variety of ways following formation
of benzene or styrene. In fact, PVA has been shown to isomerize
6
,18
to naphthalene in shock tube studies.
2
5-29
The molecule is itself composed of two structural isomers,
E-PVA and Z-PVA, in which the phenyl and ethynyl groups
are trans or cis, respectively. One question we seek to address
is whether either of the two isomers of PVA (E- or Z-)
photoisomerize to form naphthalene under conditions relevant
to Titan’s atmosphere.
on conformational isomerization
or cluster photodissocia-
30,31
tion,
where the barriers to isomerization or dissociation were
less than this amount.
Here, we extend population transfer methods to structural
isomerization initiated by ultraviolet (UV) excitation, where
absorption of a single photon provides 10 times the energy
involved in infrared or stimulated emission pumping excitation.
We present a modified version of the population transfer scheme
in which a reaction tube appended to the end of the pulsed valve
is used to constrain the initial expansion, increasing the number
of cooling collisions available to the UV-excited molecule. In
what follows, we describe UV-population transfer methods for
the first time, and their application to the ETZ isomerization
of PVA. We use UV-population transfer to determine relative
product quantum yields for isomerization as a function of
internal energy in the two directions, EfZ and ZfE. When
combined with laser-induced fluorescence (LIF) excitation and
UV-depletion spectra, which determine the relative fluorescence
quantum yields of the same bands, a detailed picture of the state-
specific isomerization dynamics for PVA is obtained.
A second motivation for this work is the opportunity it affords
to study the state-specific dynamics of photoisomerization. PVA
is a close analogue to cis/trans-stilbene, a molecule that has
played a seminal role in the search for vibrationally mode-
specific dynamics, and more generally as a large-molecule test
case for RRKM descriptions of unimolecular reaction due to
the low energy threshold for isomerization in its S
1
state (∼1200
-1
19-22
cm in trans-stilbene).
In this regard, Z-PVA and E-PVA
form a cis/trans pair involving even simpler functional groups,
with the second phenyl ring of stilbene replaced by an ethynyl
group. Our group has recently undertaken a detailed study of
the vibronic spectroscopy of E-PVA and Z-PVA, which serves
II. Experimental and Computational Methods
Ultraviolet population transfer spectroscopy is a double reso-
nance, pump/probe style of experiment (Figure 1) that has been
developed as a method for probing isomer-selective excited-state
isomerization dynamics. The methodology of UV-population
transfer is closely analogous to that of infrared population transfer,
which has been used to investigate the ground-state isomerization
2
3,24
as the foundation for the present work.
Unlike stilbene,
where one of the isomers (cis) has an unbound excited state,
both E-PVA and Z-PVA have sharp vibronic transitions that
2
6,27
dynamics of molecules with many conformational isomers.
However, because the photon energy absorbed by the molecule is
nearly an order of magnitude larger than in past work, UV-
population transfer spectroscopy must remove much larger amounts
(
(
(
14) Miller, J. A.; Klippenstein, S. J. J. Phys. Chem. A 2001, 105, 7254.
15) Miller, J. A.; Klippenstein, S. J. J. Phys. Chem. A 2003, 107, 7783.
16) Newby, J. J.; Stearns, J. A.; Liu, C. P.; Zwier, T. S. J. Phys. Chem.
A 2007, 111, 10914.
(
(
17) Robinson, A. G.; Winter, P. R.; Zwier, T. S. J. Phys. Chem. A 2002,
1
06, 5789.
(25) Dian, B. C.; Clarkson, J. R.; Zwier, T. S. Science 2004, 303, 1169.
(26) Dian, B. C.; Florio, G. M.; Clarkson, J. R.; Longarte, A.; Zwier, T. S.
J. Chem. Phys. 2004, 120, 9033.
18) Hofmann, J.; Schulz, K.; Altmann, A.; Findeisen, M.; Zimmermann,
G. Liebigs Ann.-Recueil 1997, 2541.
(
(
19) Felker, P. M.; Zewail, A. H. J. Phys. Chem. 1985, 89, 5402.
(27) Dian, B. C.; Longarte, A.; Winter, P. R.; Zwier, T. S. J. Chem. Phys.
2004, 120, 133.
20) Zwier, T. S.; Carrasquillo, E.; Levy, D. H. J. Chem. Phys. 1983, 78,
5
493.
(28) Selby, T. M.; Clarkson, J. R.; Mitchell, D.; Fitzpatrick, J. A. J.; Lee,
H. D.; Pratt, D. W.; Zwier, T. S. J. Phys. Chem. A 2005, 109, 4484.
(29) Zwier, T. S. J. Phys. Chem. A 2006, 110, 4133.
(30) Clarkson, J. R.; Herbert, J. M.; Zwier, T. S. J. Chem. Phys. 2007,
126, 15.
(
(
(
21) Amirav, A.; Jortner, J. Chem. Phys. Lett. 1983, 95, 295.
22) Tahara, T.; Hamaguchi, H. Bull. Chem. Soc. Jpn. 1996, 69, 925.
23) Liu, C. P.; Newby, J. J.; Muller, C. W.; Lee, H. D.; Zwier, T. S. J.
Phys. Chem. A 2008, 112, 9454.
(
24) Newby, J. J.; Liu, C. P.; Muller, C. W.; James, W. H.; Buchanan,
E. G.; Lee, H. D.; Zwier, T. S. J. Phys. Chem. A, DOI: 10.1021/
jp909243y.
(31) Zwier, T. S. Laser-initiated isomerization in H-bonded solute-solvent
complexes; 230th National Meeting of the American Chemical Society,
Washington, DC, 2005.
1
612 J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010

Probing E/Z Isomerization on the C10
H
8
Potential Energy Surface
A R T I C L E S
Figure 1. Schematic energy level diagram of the UV-PT experiment
illustrating a case where both excited-state and ground-state isomerization
can occur, using the Z to E isomerization of PVA as an example.
of internal energy. To that end, the aforementioned reaction channel
has been incorporated so that one can efficiently remove the excess
internal energy by constraining the expansion and therefore greatly
increasing the number of cooling collisions experienced by a UV-
excited molecule.
Figure 2. Cross section of the (a) experimental source chamber and (b)
reaction channel used in population transfer experiments. Molecules are
initially cooled in region 1 as they are expanded into the reaction channel.
Selective excitation of an isomer is accomplished by timing the pump laser
pulse to intersect the gas pulse while it is still within the reaction channel
in region 2. Upon exiting the channel, the isomeric distribution is cooled
to the zero point level and enters the collision free region in region 3. The
change in isomeric population is monitored downstream in the ionization
region via R2PI.
The experimental setup for UV-population transfer spectroscopy
makes use of a Wiley-McLaren type time-of-flight mass spec-
32-34
trometer that has been described in detail previously.
Samples
of E/Z-PVA are introduced to the ion source region via a pulsed
valve (R. M. Jordan) operating at 20 Hz, entrained in helium buffer
at a stagnation pressure of 2.7 bar. A 2 mm ID × 2 cm reaction
tube was affixed to the end of the pulsed valve (Figure 2), equipped
with a ceramic end-plate that minimizes the production of photo-
electrons due to the UV laser interaction with the valve face. The
effective temperature of gas samples in this reaction tube has been
previously estimated to be ∼75 K with pressures of ∼50 mbar,
where it is subsequently cooled to the product zero point level,
producing a gain in population in that product isomer population.
Under hole-filling conditions, the pump laser excites the molecules
while they are in the reaction tube, where hundreds of collisions
occur. Therefore, the population “hole” created by the pump laser
has the opportunity to “fill” via collisional cooling if they do not
escape the reactant well. When the probe laser is fixed on a
transition out of the zero point level of reactant A, a UV-population
transfer scan measures how much of the reactant does not return
to the reactant zero point level as a function of pump laser
wavelength. When the probe laser is fixed on a transition of the
product B, the transfer of population from A to B produces a gain
signal that measures how much population is moved into the probed
product well.
35
assuming plug flow. Upon exiting the reaction channel, the sample
is further cooled by expansion into the vacuum chamber. The free-
jet expansion is then crossed by the doubled output of a Nd:YAG
pumped dye laser operating in the 300-280 nm region, where the
vibrationally resolved, electronic spectra can be recorded using
resonant-2-photon ionization (R2PI).
In UV-population transfer experiments, a 20 Hz probe laser is
fixed on a intense transition observed in the R2PI spectrum (usually
the S
through the S
0
-S
1
origin band transition), while a 10 Hz pump is tuned
rS region of interest (Figure 3). The difference
1
0
signal (without vs with the pump laser) is monitored using active
baseline subtraction through a gated integrator. However, unlike
most double-resonant techniques, the two UV beams are not
counter-propagated. Instead, the pump laser is aligned to counter-
propagate the expansion axis (orthogonal to the probe beam). In
UV-population transfer experiments, the pump laser is timed so
that the UV pulse intersects the gas pulse while it is in the reaction
tube (pump preceding the probe by ∼65 µs). This allows for cooling
back to the zero point level (ZPL) after any isomerization has
occurred.
In UV-population transfer experiments (Figure 3a), there are two
possible outcomes: (1) population could remain in the reactant well
and be recooled to the reactant zero point level (hole-filling) or (2)
isomerization could occur, transferring population to a product well
Alternatively, one can measure where the population is going at
fixed pump energy using UV hole-filling spectroscopy. In UV hole-
filling spectroscopy (Figure 3b), the pump laser is fixed on a
transition of the reactant while the probe laser is scanned through
the region of interest. The beam alignment and laser timing
conditions are otherwise identical to that used in UV-population
transfer. As the probe laser is scanned, a gain (depletion) is recorded
whenever the intensity in a transition is increased (decreased) by
the pump laser. Ultraviolet hole-filling spectra are used to confirm
the selective excitation of the reactant, successful isomerization and
recooling of the product, and provide a measure of the relative sizes
of depletion and gain signals at a fixed pump excitation wavelength.
Finally, UV depletion spectroscopy (Figure 3b) is another
variation of UV-UV double-resonance technique, often used to
36,37
record conformer/isomer specific UV spectra.
Previous studies
(
(
(
(
32) Arrington, C. A.; Ramos, C.; Robinson, A. D.; Zwier, T. S. J. Phys.
Chem. A 1998, 102, 3315.
of E-PVA and Z-PVA used UV hole-burning spectroscopy (Figure
3d) to accomplish this task (in which the pump laser wavelength
is fixed while the probe laser wavelength is tuned). Ultraviolet
33) Bandy, R. E.; Lakshminarayan, C.; Frost, R. K.; Zwier, T. S. J. Chem.
Phys. 1993, 98, 5362.
34) Frost, R. K.; Zavarin, G. S.; Zwier, T. S. J. Phys. Chem. 1995, 99,
9
408.
35) Stearns, J. A.; Zwier, T. S.; Kraka, E.; Cremer, D. Phys. Chem. Chem.
Phys. 2006, 8, 5317.
(36) Michel, P.; Gennet, D.; Rassat, A. Tetrahedron Lett. 1999, 40, 8575.
(37) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010 1613

A R T I C L E S
Newby et al.
3
7
38
39
functional at the 6-311+G(d,p) basis set using Gaussian 03.
These geometries served as the reference state for configuration
4
0
interaction singles (CIS) calculations using this same basis set to
perform excited-state geometry optimizations, calculate excited-
state harmonic vibrational frequencies, and vertical excited-state
excitation energies. For comparison, CASSCF calculations were
also carried out at the 6-31G(d) basis set using 10 electrons in
4
1
3
8
1
0 orbitals to fully represent the ππ* transition.
Transition state geometries and vibrational frequencies were
calculated on the C10H ground-state potential energy surface to
8
aid in understanding the barriers to isomerization between E-PVA,
Z-PVA, and naphthalene. Transition states were located using the
4
2,43
synchronous transit-guided quasi-Newton method (STQN)
using B3LYP/6-311+G(d,p) for the ground state.
III. Results and Analysis
A. Review of the Spectroscopy of PVA. Because the popula-
tion transfer measurements described here build on a foundation
2
3
24
of vibronic-level spectroscopy on E-PVA and Z-PVA, we
briefly summarize here those key aspects of that spectroscopy
relevant to the present work. Figure 4 compares the laser-
1
induced excitation spectra and S origin single vibronic level
fluorescence (SVLF) emission spectra of E-PVA (Figure 4a)
and Z-PVA (Figure 4b). The electronic spectra of both molecules
are similar to what is generally seen in other substituted
benzenes, with most of the Franck-Condon activity lying in
ring and substituent-sensitive modes including ν
1
, ν
6
, ν12, and
ν
13 (in the Wilson Scheme). The excitation spectrum of E-PVA
-
1
shows vibronic activity up to 1000 cm , with a drop-off in
intensity above 1000 cm . In the absence of wavelength-
dependent nonradiative processes, the LIF and S origin SVLF
-
1
Figure 3. Energy level diagrams for the double resonance techniques used
in this work: (a) UV-population transfer, (b) UV hole-filling, (c) UV
depletion, and (d) UV hole-burning. In (a) and (b), there are sufficient
cooling collisions with the buffer gas for UV excited molecules to recool
to the zero point level, while in (c) and (d) no recooling occurs.
1
spectra will show similar Franck-Condon activity that reflects
the geometry change between these two states (Figures 4a).
While this reflection symmetry is apparent in the low-energy
region, the LIF excitation scan shows a cutoff in intensity above
-1
1000 cm that is not apparent in the SVLF spectrum, indicating
depletion spectroscopy does the opposite, fixing the probe laser
wavelength on an isomer-specific transition while the hole-burn
laser is tuned (Figure 3c). The principal advantage of UV depletion
spectroscopy is that one can observe and quantify the absorption
intensities of transitions not easily detected in the R2PI spectrum.
In the present work, because the pump laser is normally aligned to
counter propagate the supersonic expansion, we retain this config-
uration to obtain mass-selective UV depletion spectra, choosing
the delay between pump and probe laser (40 µs) so that the pump
laser intersects the gas pulse just outside the reaction channel after
the molecules have been fully cooled to the zero point level, in the
collision free regime. In so doing, the pump laser intersects all
molecules reaching the ion source region, making alignment of
pump and probe lasers straightforward.
a turn-on in nonradiative processes in these higher S vibronic
1
2
4
levels. In Z-PVA, the LIF excitation spectrum shows an
analogous cutoff in fluorescence intensity already 600-700
-
1
1
cm above the S origin.
B. Ultraviolet Depletion Spectra. As a first step in probing
the excited-state dynamics, Figure 5 presents a comparison of
LIF and UV depletion spectra of E- and Z-PVA acquired using
fluorescence detection. The data on Z-PVA have been presented
2
4
previously and are included here for ready comparison with
the analogous results on E-PVA. Relative fluorescence quantum
yields are extracted from the data simply by taking the ratio of
the intensities of transitions in LIF (ILIF R Nexc ·Φ ) to those in
f
UV depletion recorded under otherwise identical conditions
While described in the case of an R2PI experiment, UV depletion
spectroscopy can also be readily carried out utilizing fluorescence
(IUVD R Nexc). Because the laser power used in UV depletion is
higher than that in a typical LIF spectrum, many transitions in
Figure 5 show some degree of saturation compared to the LIF
spectrum in Figure 4. Figure 6 overlays on the LIF excitation
spectra of the two isomers the relative single vibronic level
2
4
detection, as was done recently for Z-PVA. Here, we also report
fluorescence-based UV depletion spectra for E-PVA, acquired using
2
7
a fluorescence chamber described previously. These spectra will
be used to determine relative fluorescence quantum yields for
E-PVA to compare with those in Z-PVA.
(38) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80,
3265.
(39) Frisch, M. J.; Gaussian 03, revision E.01; Gaussian, Inc.: Wallingford,
CT, 2004. The full reference is given in the Supporting Informa-
tion.
E-PVA and Z-PVA were synthesized according to the procedure
24,36
of Michel, Gennet, and Rassat.
Each isomer of PVA was stored
as a dilute solution (<10% in hexane) at 0 °C to minimize
polymerization and decomposition. Each isomer was determined
be nearly isomerically pure (>90%) by gas chromatography.
To aid in the analysis of the isomerization of PVA, ground-state
geometries and harmonic vibrational frequencies were calculated
for both isomers using DFT calculations that employed the B3LYP
(
40) Foresman, J. B.; Head-Gordon, M.; Pople, J. A.; Frisch, M. J. J. Phys.
Chem. 1992, 96, 135.
(
(
41) Eade, R. H. A.; Robb, M. A. Chem. Phys. Lett. 1981, 83, 362.
42) Peng, C. Y.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. J. Comput.
Chem. 1996, 17, 49.
(43) Peng, C. Y.; Schlegel, H. B. Isr. J. Chem. 1993, 33, 449.
1
614 J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010

Probing E/Z Isomerization on the C10
H
8
Potential Energy Surface
A R T I C L E S
Figure 4. Comparison of the laser-induced fluorescence excitation spectrum (black) and the S
1
origin dispersed fluorescence spectra (red) of (a) E-PVA and
(
b) Z-PVA. The loss of reflection symmetry between excitation (black) and emission (red) suggests the turn-on of a nonradiative process in the excited state.
above 1000 cm- in (a) E-PVA and above 600
1
Figure 5. Comparison of the LIF (black) and UV depletion spectra (red) shows a significant decrease in Φ
f
-1
cm in (b) Z-PVA.
rel
fluorescence quantum yields, Φ
f
, normalized to 1.00 at the
7a) appear as depletions in the UV hole-filling spectrum
starting from Z-PVA as reactant, as should occur if population
is lost from that isomer by UV excitation. At the same time,
4
6
rel
electronic origin. As expected, in both isomers, Φ
f
shows a
precipitous drop by over a factor of 50 as their respective
thresholds are overcome. Furthermore, below the threshold,
individual transitions show some evidence for mode-specific
effects, a point to which we will return in the Discussion.
C. UV Hole-Filling Spectra. A first step in demonstrating
the ability to carry out state-specific studies of structural
isomerization is to determine whether the experimental
scheme shown in Figure 1 can successfully recool the UV
excited products to their zero-point levels. Ultraviolet hole-
filling spectroscopy (section II) answers this question by
fixing the UV wavelength of the pump laser on a transition
of the reactant of interest (e.g., E-PVA) while the probe laser
is scanned through vibronic transitions due to both reactant
0
transitions out of the E-PVA S zero point level appear as
gain signals, indicating that isomerization from ZfE has
occurred, and the excess energy has been completely removed
by collisions in the reaction channel and postchannel expan-
sion. No other gain or depletion signals appear in this
wavelength region, demonstrating that ZfE isomerization
is being carried out cleanly under UV hole-filling conditions,
with recooling complete prior to interrogation. In the same
way, the middle trace of Figure 7b demonstrates the reverse
isomerization process, from EfZ following excitation of the
1
-1
25
0
transition of E-PVA, 931 cm above its electronic
origin. Once again, selective excitation of E-PVA and
complete removal of excess energy from the UV-excited
molecules is apparent.
D. UV-Population Transfer Spectra. Figure 8 shows a series
of UV-population transfer spectra carried out on near-pure
samples of E-PVA (Figure 8a) and Z-PVA (Figure 8b) while
and product. The middle trace of Figure 7a shows a UV hole-
1
filling scan carried out with the pump laser fixed on the 24
0
vibronic band of Z-PVA, 933 cm- above its S
1
origin. LIF
1
scans of E-PVA and Z-PVA are shown above and below the
UV hole-filling scan for direct comparison.
As anticipated, under hole-filling conditions, vibronic
transitions arising out of the Z-PVA zero point level (Figure
monitoring the S
(top trace) or reactant (middle trace). The bottom trace of
0
1
-S origin transitions of the isomer product
J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010 1615

A R T I C L E S
Newby et al.
prominently evident in the product gain spectra. In the EfZ
UV-population transfer spectrum, the broad feature begins to
-1
rise about 715 cm above the electronic origin, the same energy
region where the fluorescence quantum yield drops. A similar
background is evident in the ZfE UV-population transfer
spectrum (Figure 8b, top trace), with a somewhat less clear
threshold, in part because of the extensive hot band structure
present in the spectrum. Nevertheless, the rising baseline can
be traced back to a similar energy above its S
1
origin (∼700
-
1
-1
cm ) as in the EfZ case, slightly below the 1000 cm
nonradiative threshold observed in E-PVA. It seems plausible,
then, that this rising background is associated with overcoming
a barrier to another excited state, which is also responsible for
the nonradiative process.
Third, the UV-population transfer spectra monitoring the
return to the reactant well (EfE in Figure 8a and ZfZ in
Figure 8b) show small depletions reflecting a small fractional
loss in population from the reactant due to isomerization.
These spectra follow the intensity patterns in the UV
depletion spectra below them, but with reduced intensity due
to the partial refilling that occurs under hole-filling conditions.
In the following section, we will consider the extent to which
these data can be used to obtain isomerization quantum yields.
Finally, UV-population transfer spectra monitoring the
product show large fractional population changes due to the
fact that the initial samples have only a minor impurity of
the other isomer, leading to large fractional population
changes.
E. Relative Isomerization Quantum Yields from UV-Po-
pulation Transfer Data. To this point, we have measured
relative single vibronic level fluorescence quantum yields for
both ZfE and EfZ isomerization that show a sharp drop at
-
1
thresholds ∼600 and 1000 cm above the S
1
zero point level
of Z- and E-PVA, respectively. The UV-population transfer
scans of Figure 8 provide qualitative evidence that isomer-
ization occurs both below and above these thresholds.
However, we have not yet extracted quantitative values for
isomerization quantum yields to determine whether the drop
in fluorescence quantum yield is reflected in an analogous
turn-on in isomerization quantum yield, that is, whether the
nonradiative process funnels population selectively into the
product isomer well.
Figure 6. Experimentally observed relative fluorescence quantum yields,
with error bars, superimposed on the LIF spectrum for (a) E-PVA and (b)
Z-PVA.
each series is a UV depletion spectrum showing the fraction
of the reactant population removed by photoexcitation when
recooling of the UV excited molecules does not occur (carried
out with R2PI detection). In E-PVA, we were able to record
the UV-population transfer spectrum from the electronic
origin to almost 1500 cm- above it. In Z-PVA, the limited
size of our synthesized sample required us to carry out UV-
In the previously mentioned studies using IR population
transfer spectroscopy, an experimental protocol was developed
for measuring absolute isomerization product quantum yields.
1
26,27
population transfer spectra with pump laser tuned only over
These equations have analogues in UV-population transfer
spectroscopy, suitably modified to account for differences in
the experimental protocol associated with counter-propagation
of the laser and the presence of the reaction channel, with its
increased collisional cooling capacity. As detailed in the
Supporting Information, these differences make it difficult to
extract absolute product quantum yields from the present
measurements. As a result, we have chosen instead to report
the region from 500-1500 cm- above the Z-PVA S
1
-S
0
1
origin.
Several deductions regarding the ETZ isomerization dynam-
ics can be immediately drawn from these spectra. First, product
signal gains are observed in both directions (EfZ and ZfE)
already at the S
0
1
-S electronic origin, well below the nonra-
diative threshold identified by the fluorescence quantum yield
44
1
measurements. Thus, already at the S origin of both isomers,
here product quantum yields on a relative scale, normalized to
a fraction of the population undergoes internal conversion to
the ground state, leading to isomerization.
45
1
the S origin according to eq 1.
Second, in all of the UV-population transfer spectra, a broad
background grows in at higher excitation energies, most
(
45) To make quantitative comparisons to the UVPT spectra, the UVD
spectra were retaken under ionization condition and have been found
to be nearly identical to those acquired under fluorescence conditions.
(
44) Population transfer at the electronic origin of Z-PVA was confirmed
by specifically probing that transition in lieu of a full scan near the
electronic origin transition.
(46) Taylor, J. T. An Introduction to Error Analysis: The Study of
Uncertainties in Physical Measurements; University Science Books:
Sausalito, 1996.
1
616 J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010

Probing E/Z Isomerization on the C10
H
8
Potential Energy Surface
A R T I C L E S
UVPT
AB
UVD
AB
0
0
the zero point level of each reactant, where we have not included
intensity due to the broad background when determining UV-
population transfer transition intensities. Error bars were
estimated by comparing the area of the transitions from several
scans of the UV-population transfer spectrum and propagated
using standard methods. The table also includes the measured
relative fluorescence quantum yields for the same bands (section
III.B and Figure 6) for comparison. It is clear from this
comparison that the isomerization quantum yields for these sharp
vibronic bands do not change significantly as the fluorescence
quantum yield drops by almost a factor of 50. We surmise on
this basis that the nonradiative threshold responsible for the drop
ΦAB(i)
I
(i)
0
I
(0 )
)
·
(1)
0
UVPT
AB
UVD
(i)
AB
ΦAB(0 )
I
(0 )
I
0
0
These relative isomerization yield measurements are in keeping
with our primary interest in determining whether the wavelength-
dependent UV-population transfer scans (Figure 8) reflect a turn-
on in isomerization quantum yield as the threshold for the
nonradiative process is reached.
Table 1 presents relative product isomerization yields for
EfZ and ZfE isomerization for a select set of vibronic
transitions of E-PVA (left-hand columns) and Z-PVA (right-
hand columns). These transitions are strong transitions out of
46
Figure 7. Comparison of LIF spectra (red and blue traces) with hole-filling spectra (red trace) for the (a) ZfE and (b) EfZ isomerizations. In both panels,
the middle spectrum is the UV hole-filling spectrum in which population gain produces an increase in signal and population loss a depletion in signal. The
LIF spectrum of the depleted species (blue trace) has been inverted for easier comparison.
Figure 8. Comparison of the UV population transfer spectra for forming product (red) and reforming reactant (blue) starting from (a) E-PVA and (b)
Z-PVA. The UV depletion spectrum, in which no refilling occurs, is shown in the black trace. Transitions marked with asterisks are assigned to hot-bands
of Z-PVA present at the temperature in the reaction channel.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010 1617

A R T I C L E S
Newby et al.
Table 1. Comparison of the Fluorescence Quantum Yields with
Product Quantum Yields
ν/cm-
1
Φ
r
f
a
(E)
Φ
(EZ)
b
∆ν/cm-1
Φ
r
f
c
(Z)
d
∆
r
r
Φ (ZE)
0
3
1.00
0.67
0.90
1.26
0.19
0.18
0.13
0.11
1.00
0
535
702
937
967
1060
1152
1223
1.00
0.30
0.02
0.02
0.03
0.01
0.02
0.02
26
15
31
214
279
292
434
0.77 ( 0.23
0.90 ( 0.27
0.86 ( 0.26
0.96 ( 0.29
0.77 ( 0.23
0.81 ( 0.24
0.61 ( 0.18
1.00
8
9
1
1
1
1
1.46 ( 0.44
0.95 ( 0.29
1.59 ( 0.48
1.54 ( 0.46
1.05 ( 0.32
1.19 ( 0.36
a
b
Fluorescence quantum yields relative to the origin band of E-PVA. Flu-
orescence quantum yields relative to the origin band of Z-PVA.
c
Isomerization quantum yields relative to the origin band of E-PVA.
d
1
Isomerization quantum yields relative to the 29
0
band of Z-PVA.
Figure 9. Excited-state potential energy curve for hinder rotation about
the double bond in PVA at the CIS and CASSCF levels of theory.
in fluorescence quantum yield is not one that is directly
associated with ETZ isomerization. At the same time, additional
absorption leading to isomerization produces a broad background
that does show a threshold in the same energy region as the
drop in fluorescence quantum yield.
can be used to state-selectively excite and probe both isomers.
As a result, the ETZ photophysics and photoisomerization
dynamics can be probed as a function of internal energy from
either starting geometry.
F. Ultraviolet Population Transfer Spectra Searching for
Naphthalene Products. Because PVA is a structural isomer of
naphthalene, it is natural to ask if PVA can photoisomerize to
naphthalene. In fact, Zimmer and co-workers reported formation
In the present work, we have used a combination of LIF,
UV depletion, and UV-population transfer spectra to provide
relative single vibronic level fluorescence quantum yields and
isomerization product yields as a function of internal energy
for both E and Z isomers. We are now in a position to pull
together our data into a single coherent picture of the excited-
state dynamics of PVA.
6
of naphthalene from the flash vacuum pyrolysis of PVA. On
the basis of the geometry of the isomers, it would seem most
plausible for Z-PVA to isomerize to naphthalene rather than
E-PVA, because the acetylenic moiety in Z-PVA is in close
proximity to the phenyl ring, as needed for ring closure to occur.
To this end, we have tried to observe the photoinduced ring
closure from Z-PVA to naphthalene.
To search for naphthalene production, we employed UV hole-
filling spectroscopy, where the pump laser was fixed on a
vibronic transition of Z-PVA while the probe laser was scanned
through the region where a portion of the 1-color R2PI spectrum
The relative fluorescence quantum yields show sharp drops
-1
1
beginning 600 and 1000 cm above the S origins of Z- and
E-PVA, respectively. These thresholds are analogous to the
corresponding thresholds in trans-stilbene, which have been the
subject of ongoing discussions regarding whether the nonra-
diative rates are governed by statistical or nonstatistical energy
19-22
flow.
In PVA, the spectra of both isomers are sharp at the
vibronic level below these thresholds, opening the possibility
for a quantitative analysis starting from either reactant well. In
the present studies, we have not obtained a quantitative
measurement of the excited-state lifetime at the S origin, which
1
would be necessary for a complete RRKM analysis. The
isomerization quantum yield measurements prove that nonra-
of naphthalene associated with the S
0
2
-S transition is known
47
to occur (282.5-275 nm). During this search for naphthalene
products, many different pump laser frequencies were used,
ranging from as far red as the electronic origin of Z-PVA
-1
-1
(
33 838 cm ) to as far blue as 4000 cm above its S
0
-S
1
electronic origin. The population transfer gain signal into the E
isomer served as a check that we were still able to transfer
population efficiently; however, in no case was naphthalene
detected unambiguously. For completeness, the search for
naphthalene was also attempted from the E isomer, but again
was not detected.
diative pathways compete with fluorescence already at the S
1
origin. Furthermore, on the basis of the fluorescence decay
profile observed in these LIF measurements with nanosecond
lasers, we can only place an upper bound on the S origin
1
lifetimes of a few nanoseconds in both isomers. It would be
very interesting to have picosecond fluorescence decay profiles
or high resolution UV spectra that resolve or partially resolve
the rotational structure, which could both be used to deduce
IV. Discussion
4
8
the excited-state lifetime. In the absence of that data, one can
only estimate the excited-state radiative lifetime on the basis
of the calculated oscillator strength for the ππ* transition. On
the basis of the B3LYP 6-311+G(d,p)//TD-DFT/6-311+G(d,p)
calculations, we estimate radiative lifetimes of 1.9 and 2.5 ns
for the E- and Z-PVA isomers.
A. The Photophysics and Photoisomerization of PVA: Com-
parison with Stilbene. In the introduction, Z-PVA and E-PVA
were posed as close analogues of cis- and trans-stilbene, a
molecule that has served as a prototype for studies aimed at
understanding the time scales, reaction pathways, and state-
19-21
specific energy flow associated with cis/trans isomerization.
Figures 9 and 10 present two different one-dimensional cuts
along the excited-state surfaces of E- and Z-PVA. Figure 9
shows a series of single-point calculations projecting up from
the ground-state surface, stepping along the CdC twisting
coordinate that is the direct analogue of the isomerization
coordinate in stilbene. The 0° and 180° end-points are single-
In stilbene, the excited-state surface accessed from the cis ground
state has no bound levels, while the trans isomer has a sharp
spectrum with a nonradiative pathway turning on about 1200
cm above the trans S origin. An intriguing aspect of the
1
current studies of PVA is that the ππ* excited-state surface of
both isomers (E and Z) is bound, with sharp transitions that
-1
(
47) Beck, S. M.; Powers, D. E.; Hopkins, J. B.; Smalley, R. E. J. Chem.
Phys. 1980, 73, 2019.
(48) Alvarez-Valtierra, L.; Yi, J. T.; Pratt, D. W. J. Phys. Chem. A 2009,
113, 2261.
1
618 J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010

Probing E/Z Isomerization on the C10
H
8
Potential Energy Surface
A R T I C L E S
Figure 10. Potential energy surfaces along the acetylenic bending coordinate for the first two excited states of A′ (blue) and A′′ (red) symmetry of (a)
E-PVA and (b) Z-PVA.
point projections from the ground-state minima for E- and
Z-PVA, respectively. The CIS calculations become unstable near
We conclude that excited-state isomerization is not driving the
reduction in fluorescence quantum yield, so that the nonradiative
process that turns on in the excited states of E- and Z-PVA is
not the critical bifurcation point for forming one or the other
isomer. Instead, the sharp turn-on in nonradiative rate is
associated with the surface crossing from ππ* to πσ* states,
but the πσ* state must itself be an intermediate that facilitates
internal conversion to the ground state, with isomerization
occurring on the ground-state surface. Furthermore, the close
proximity of the πσ* state to the ππ* state in the Franck-Condon
region may also mean that this state plays a role in the
9
0° due to instabilities in the reference state wave function about
this geometry; however, the CASSCF calculations show a
smooth increase in energy to a transition state about 8500 cm
above the S origins of E- and Z-PVA. While this coordinate
-1
1
most closely mimics that in stilbene, it seems unlikely to be
responsible for the nonradiative process detected in the first 1500
-
1
cm above the S
1
origins of E- and Z-PVA, where isomerization
origin, with a fast nonradiative
can occur already at the S
1
threshold encountered at energies 10 times smaller than the
CASSCF barrier.
photoisomerization observed at the S
although direct internal conversion to S
1
origins of the two isomers,
cannot be ruled out.
Instead, as previously identified in our discussion of the
vibronic spectroscopy and fluorescence quantum yields of
0
Before closing this section of the discussion, two final
comments should be made. First, UV-population transfer spectra
provide evidence that overcoming the nonradiative threshold
accesses a dense manifold of dark states that lead to isomer-
ization, producing the broad backgrounds observed (Figure 8).
These background states must gain oscillator strength from the
ππ* state, but are hard to distinguish from simple spectral
congestion in the UV depletion scans. However, the UV-
population transfer spectra show evidence for the onset of this
background absorption near the thresholds for nonradiative
decay, suggesting that they arise from accessing the πσ* state.
Finally, while the isomerization quantum yields show no
mode-specific effects, the relative fluorescence quantum yields
do show variations somewhat outside the error bars (Figure 6).
While not large, they suggest that, even below the barrier, the
coupling to the nonradiative πσ* intermediate state may have
some mode specificity, an issue worth exploring in greater detail,
perhaps by high-resolution methods.
2
4
Z-PVA, the acetylenic group is not a passive substituent, but
actively participates in the nonradiative pathways in the excited
state. Figure 10 presents potential energy curves stepping along
the CtCH bending coordinate, tracking the four lowest energy
excited states, two ππ* states of A′ symmetry and two πσ*
states of A′′ symmetry. The striking result of these calculations
is that the lowest energy excited singlet state of PVA is a πσ*
state of A′′ symmetry, with a nonlinear CCH geometry bent
∼
60° away from linear. In the Franck-Condon accessible
region, the lowest ππ* state is calculated to be slightly lower
in energy than the πσ* state, but the two are in close proximity
over a large range of angles, with crossing points near (25° in
both E- and Z-isomers. While the absolute magnitudes of the
crossing points are somewhat higher (2300/1900 cm- for E/Z)
1
-1
than the experimental nonradiative thresholds (1000/600 cm ),
their relative energy ordering is in keeping with experiment.
As a result, we postulate that the first πσ* singlet state is
responsible for the sharp threshold in nonradiative decay in both
isomers, a coordinate that is not directly associated with ETZ
isomerization at all.
8
B. The C10H Potential Energy Surface and Isomerization
to Naphthalene. The one-dimensional cuts on the excited-state
surfaces shown in Figures 8 and 9 still leave open the question
of how isomerization proceeds following internal conversion
on the ground-state surface. The recent ab initio study of the
This is borne out also by the relative isomerization quantum
yield measurements of Table 1. At least for the strong vibronic
transitions for which we have quantitatively reliable data, the
relative isomerization quantum yields are relatively constant with
increasing internal energy, maintaining a similar value above
the nonradiative threshold as below it, despite the fact that the
fluorescence quantum yields dropped by almost a factor of 50.
8
ground-state C10H surface by Dyakov et al. had a somewhat
different focus (the photodissociation pathways of azulene);
nevertheless, it included consideration of many minima and
transition states on the ground-state surface, including E- and
J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010 1619

A R T I C L E S
Newby et al.
channel, the C
6
H
6
products formed were shown not to be
6 6
benzene, indicating that collisional trapping of the C H product
1
6
in other isomer wells must be occurring.
Along the isomerization pathway to naphthalene, we have
identified at least one stable intermediate: a biradical species
that forms after the initial ring closure, but before the sigmat-
ropic H-atom transfers required to make naphthalene. It is
possible that, while the naphthalene product channel is initially
energetically accessible, collisional cooling in the reaction
8
channel traps the C10H product in an isomer that is optically
dark in the region probed (300-275 nm).
V. Conclusions
The isomerization dynamics of E-PVA and Z-PVA have been
studied using the double-resonance techniques of UV depletion
and the newly developed techniques of UV-population transfer
and UV hole-filling spectroscopy. While Z-PVA and E-PVA
structural isomers are in principle direct analogues of cis- and
trans-stilbene, the presence of the acetylenic group opens up
new nonradiative pathways involving a low-lying πσ* state that
changes the photophysics and photoisomerization mechanisms
in fundamental ways. Nevertheless, the ability to state-selectively
photoexcite both isomers and follow the isomerization via
pump-probemethodssuchasreportedheremakestheE-PVATZ-
PVA isomerization process a prime candidate for further
theoretical studies seeking quantitative descriptions of the
intramolecular processes involved. Furthermore, the presence
of other important isomerization channels with similar ground-
state reaction thresholds, including naphthalene and azulene,
makes the E-PVA/Z-PVA isomers attractive entry points for
further studies seeking to understand photoisomerization on this
Figure 11. Ground-state potential energy surface for the isomerization from
E-PVA and Z-PVA to naphthalene (kJ/mol) calculated employing the
B3LYP function using the 6-311+G(d,p) basis set.
49
Z-PVA. We have reproduced key aspects of these calculations
relevant to the present work at a slightly different level of theory
(
DFT B3LYP/6-31+G(d)), which are included in Figure 11.
As one can see, the calculations predict that the E-isomer is
more stable than Z-PVA by 3.6 kJ/mol. The two isomers are
connected by a transition state on the ground-state surface for
ETZ isomerization, which is 201 kJ/mol above the E minimum,
well below the energy imparted by UV photoexcitation (660
kJ/mol). This makes clear that, if isomerization is occurring on
the ground-state surface, it does so well away from the
thresholds probed in UV excitation.
Not surprisingly, the ground-state energy of naphthalene is
calculated to be much lower than that of either E- or Z-PVA
8
fundamentally important C10H potential energy surface.
(∼253 kJ/mol). As Dyakov et al. have pointed out, the pathway
from PVA to naphthalene occurs out of the Z-PVA well, which
is poised for ring closure. According to the calculations,
formation of naphthalene occurs via a diradical intermediate
that is 98 kJ/mol higher in energy than Z-PVA (Figure 11), with
barriers into and out of the intermediate well 235 and 121 kJ/
mol above E-PVA. These energies are only 34 kJ/mol above
the barrier for ETZ isomerization (201 kJ/mol), also well below
the UV photoexcitation energy. Furthermore, the stability
of the naphthalene product will greatly favor its formation in
the absence of collisional cooling, when the products are under
thermodynamic control.
The unsuccessful search for naphthalene products under UV-
population transfer conditions indicates that collisional cooling
must be fast compared to the time scale for isomerization on
the ground-state surface. Under the fast flow conditions in the
reaction channel, collisional cooling is complete on the tens of
microsecond time scale. In previous experiments in our labora-
tory on propargyl recombination that employed the reaction
Acknowledgment. We gratefully acknowledge support from the
NASA Planetary Atmospheres program (NNG06GC57G) for this
research. We express thanks to Information Technology at Purdue-
the Rosen Center for Advanced Computing, West Lafayette,
Indiana, for computational resources, H. Daniel Lee for synthetic
assistance, and Marek Zgierski for helpful discussion identifying
the πσ* states as possible causes of the nonradiative thresholds
involved. C.W.M. would like to thank the “Deutsche Akademie
der Naturforscher Leopoldina” for a postdoctoral scholarship (grant
number BMBF-LPD 9901/8-159 of the “Bundesministerium f u¨ r
Bildung und Forschung”). C.-P.L. thanks Taiwan NSC for the
financial support through Taiwan Merit Scholarships (TMS-094-
2
-B-020) during her postdoctoral research at Purdue University.
Supporting Information Available: Extracting quantum yields
from UV-population transfer spectra and full ref 39. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(
49) Dyakov, Y. A.; Ni, C. K.; Lin, S. H.; Lee, Y. T.; Mebel, A. M. J.
Phys. Chem. A 2005, 109, 8774.
JA908103U
1
620 J. AM. CHEM. SOC. 9 VOL. 132, NO. 5, 2010