The dynamics of hydrogen abstraction reactions: Crossed-beam reaction Cl+n-C5H12→C5H11+HCl

Article abstract of DOI:10.1063/1.476523

We present a crossed molecular beam study of the dynamics of the metathesis reaction Cl+n-C₅H₁₂→HCl+C₅H₁₁ at a collision energy of 16.8 kcal/mol. The experiments were conducted on the Chemical Dynamics Beamline at the Advanced Light Source, using tunable undulator radiation to effect soft ionization of the pentyl radical product. Laboratory angular distributions and time-of-flight spectra at many laboratory angles were used to obtain the center-of-mass translational energy and angular distributions. These distributions were found to be strongly coupled, with the forward scattered pentyl radical formed extremely cold, while the backscattered radicals were formed leaving nearly 15 kcal/mol in internal energy in the products. These results are contrasted with recent studies on the analogous reaction with propane, suggesting in this case direct involvement of the carbon skeleton in the collision process.

Full text of DOI:10.1063/1.476523

The dynamics of hydrogen abstraction reactions: Crossed-beam reaction Cl +n- C 5 H
2 → C 5 H 11 +HCl
1
Naoki Hemmi and Arthur G. Suits
Citation: The Journal of Chemical Physics 109, 5338 (1998); doi: 10.1063/1.476523
View online: http://dx.doi.org/10.1063/1.476523
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/109/13?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Rotationally resolved reactive scattering: Imaging detailed Cl + C 2 H 6 reaction dynamics
J. Chem. Phys. 125, 133107 (2006); 10.1063/1.2202827
Imaging O ( 3 P)+ alkane reactions in crossed molecular beams: Vertical versus adiabatic H abstraction
dynamics
J. Chem. Phys. 117, 7947 (2002); 10.1063/1.1510444
Dynamics of the C + C 2 H 2 reaction from differential and integral cross-section measurements in crossed-beam
experiments
J. Chem. Phys. 116, 5603 (2002); 10.1063/1.1456508
Crossed beams study of C-H bond activation: Mo ( 5 S 2 )+ CH 4 →MoCH 2 +H 2
J. Chem. Phys. 108, 2665 (1998); 10.1063/1.475658
Crossed-beam reaction of carbon atoms with hydrocarbon molecules. IV. Chemical dynamics of methylpropargyl
radical formation, C4H5, from reaction of C(3Pj) with propylene, C3H6 (X1A)
J. Chem. Phys. 106, 4945 (1997); 10.1063/1.473543
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:
132.174.255.116 On: Sun, 30 Nov 2014 17:30:03

JOURNAL OF CHEMICAL PHYSICS
VOLUME 109, NUMBER 13
1 OCTOBER 1998
The dynamics of hydrogen abstraction reactions: Crossed-beam reaction
Cl؉n-C H ˜C H ؉HCl
5
12
5
11
Naoki Hemmi and Arthur G. Suits
Chemical Sciences Division, Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley,
California 94720
͑
Received 17 June 1998; accepted 29 June 1998͒
We present a crossed molecular beam study of the dynamics of the metathesis reaction Cl
ϩn-C H →HClϩC H at a collision energy of 16.8 kcal/mol. The experiments were conducted
5
12
5
11
on the Chemical Dynamics Beamline at the Advanced Light Source, using tunable undulator
radiation to effect soft ionization of the pentyl radical product. Laboratory angular distributions and
time-of-flight spectra at many laboratory angles were used to obtain the center-of-mass translational
energy and angular distributions. These distributions were found to be strongly coupled, with the
forward scattered pentyl radical formed extremely cold, while the backscattered radicals were
formed leaving nearly 15 kcal/mol in internal energy in the products. These results are contrasted
with recent studies on the analogous reaction with propane, suggesting in this case direct
involvement of the carbon skeleton in the collision process. ͓S0021-9606͑98͒02437-4͔
I. INTRODUCTION
this arises from a shift in the nature of the surfaces from
repulsive to attractive across the series, with the more pro-
ductlike transition state for primary H abstraction favoring
vibrationally cold products while the more reactantlike tran-
sition state for the tertiary H abstraction yields greater vibra-
tional excitation.
Metathesis reactions, in which a single atom is trans-
ferred from a stable molecule to an atom or radical, are the
only truly bimolecular ͑i.e., bimolecular in both directions͒
1
gas-phase reactions. Those involving hydrogen atom trans-
fer arguably represent the most important subclass of this
vast body of reactions, particularly for combustion dynamics.
Free radical abstraction of hydrogen atoms in saturated hy-
drocarbons, for example, are reactions of great importance in
combustion, and the differing propensities for reaction of
primary, secondary, or tertiary H atoms, as well as the dif-
ferent dynamics underlying these pathways, are important to
a detailed understanding of combustion chemistry. These
processes have come under increasing scrutiny in recent
years as experimental innovations have allowed more direct
probing of the dynamics.
Studies of the dynamics of chlorine atom reactions with
hydrocarbons were stimulated in the early 90’s when it was
realized that these reactions represent a possible termination
step for the chlorine-atom catalyzed ozone destruction cycle
in the stratosphere. Theoretical studies3 mapped features of
the transition state for the methane reaction, while an experi-
mental dynamics study in 1990 considered reaction of Cl
,4
5
with deuterated cyclohexane. In the latter case, using time-
resolved infrared absorption probe of the DCl product, Flynn
and co-workers observed cold rotational distributions of the
DCl, and efficient coupling of the initial collision energy into
product translational energy. They framed the dynamics in a
simple ‘‘spectator’’ picture in which the hydrocarbon radical
does not participate in the reaction and little momentum is
transferred in the collision. They also invoked a collinear
transition state to account for the cold rotational distributions
An important pair of papers by Andresen and Luntz in
2
1
980 reported experimental and theoretical studies of the
3
reaction dynamics of O( P) with a variety of saturated hy-
drocarbons chosen to underscore distinct trends in the reac-
tion dynamics for primary, secondary or tertiary H atom ab-
straction. They measured OH product state distributions
from these reactions using laser-induced fluorescence ͑LIF͒
and compared the results to classical trajectory calculations
employing triatomic model surfaces with empirical poten-
tials. The trends they observed for the OH rotational state
distributions have been confirmed in all subsequent studies
of H abstraction dynamics, viz., all the newly formed hy-
dride products, regardless of the identity of the hydrocarbon
or the attacking atom, are formed rotationally cold. This was
ascribed to the importance of a collinear O–H–C transition
state in these reactions. In addition, they found that the more
exothermic the process ͑exothermicity increases in the se-
quence primaryϽsecondaryϽtertiary͒, the greater the yield
of vibrationally excited OH product, with tertiary H atom
abstraction preferentially populating vϭ1. They argued that
6
in the DCl product. Koplitz and co-workers used selective
deuteration to investigate the site propensities for reaction of
Cl with propane, using state-resolved resonant multiphoton
ionization ͑REMPI͒ to detect the HCl or DCl products. They
concluded that the middle carbon site ͑with secondary H
atoms͒ was significantly more reactive than the terminal car-
bons.
A
pioneering series of papers from the Zare
laboratory7
–10
reported studies of Cl reaction with both vi-
brationally excited- and ground-state methane molecules, us-
ing their ‘‘photoloc’’ technique to provide state-resolved dif-
ferential cross sections for these reactions. Reaction of Cl
5
338
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:
32.174.255.116 On: Sun, 30 Nov 2014 17:30:03
1

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
N. Hemmi and A. G. Suits
5339
atoms with vibrationally excited methane was found to be
quite facile, showing the characteristic cold rotational distri-
bution but predominantly forward scattered HCl, especially
for the vibrationally excited HCl product. In addition, al-
though all rotational distributions were relatively cold, for
the vibrationally excited HCl product, the higher J values
were associated with greater backscattering. A significant
steric effect was also reported, with reaction yielding HCl
facile abstraction of the secondary H atoms, while the latter
were believed to result from more strongly coupled colli-
sions involving the terminal H atoms. Here we extend these
studies to reaction of Cl with n-pentane, in an effort to ex-
amine the underlying trends in the reaction dynamics, in par-
ticular allowing us to probe the role of the extended carbon
skeleton in these reactions.
(
vϭ1) favored by perpendicular approach of the Cl atoms to
II. EXPERIMENT
the excited C–H stretch. Their results for reaction with
ground vibrational state methane showed strongly backscat-
tered, rotationally cold distributions. They rationalized this in
terms of a hard-sphere model in which reaction is only pos-
sible for small impact parameter collisions. They also argued
that the association of low rotational excitation with a collin-
ear transition state, as adduced by Andresen and Luntz in the
These experiments were performed using a rotatable
source molecular beam apparatus on the chemical dynamics
beamline at the advanced light source at Lawrence Berke-
15
ley National Laboratory. The apparatus has been described in
16
detail elsewhere. Briefly, a molecular beam of the hydro-
carbon reagent, n-pentane, crosses a chlorine atom beam and
the reactive scattering products are detected by a quadrupole
mass spectrometer. The two molecular beam sources are
fixed at 90° and the whole source chamber is rotatable to
allow data collection at angles of Ϫ20° to 110° between the
pentane beam and detector. The significant feature in the
experimental setup is the use of the vacuum ultraviolet
3
O( P) reactions and Flynn and co-workers for the Cl–C D
6
6
reaction, is strictly appropriate only when there is impulsive
energy release in the reaction. For the Cl-methane case, this
is only applicable for the backscattered products. For the
forward-scattered HCl, which is the dominant product from
reaction of vibrationally excited methane, the cold rotational
distributions were attributed simply to strippinglike mecha-
nism involving large impact parameter collisions.
͑VUV͒ undulator radiation for photoionization probe of the
neutral scattering products, in place of the conventional elec-
tron impact ionizer. The spectrum of the undulator radiation
consists of a fundamental peak of roughly Gaussian energy
profile with a full width at half maximum ͑FWHM͒ of 2.5%,
and a series of harmonics of diminishing intensity extending
beyond nϭ15. The energy of the fundamental may be tuned
continuously over a range from 5 to 30 eV. The harmonics
are suppressed as described below. Under typical operating
conditions, the time structure of the undulator radiation is 35
ps pulses at 500 MHz; for our purposes it may be considered
a continuous source. The photon flux at full storage ring
current is greater than 1016 photons/s. It is focused to a di-
ameter less than 0.5 mm in the detector region.
Varley and Dagdigian have reported an extensive inves-
tigation of the reaction dynamics of Cl with simple
1
1–13
hydrocarbons
using a technique similar to the photoloc
method employed in Zare’s experiments. In this work, state-
resolved differential cross sections were obtained for the HCl
product from reaction with a number of hydrocarbons in-
cluding methane, propane, isobutane, and selectively deuter-
ated propane. They observed backscattered HCl from the
isobutane reaction, and vibrational population ratios that
were similar for propane and isobutane, in contrast to the
3
O( P) results of Andresen and Luntz. Their results for the
partially deuterated propane showed abstraction of the D at-
oms ͑as the primary substituents͒ led to sideways scattered
DCl product, while abstraction of the H atoms ͑the second-
ary substituents͒ led to a more isotropic angular distribution,
yet still favoring the backward direction. It must be noted
that their collision energies were very broad, and extended to
very high energies, owing to the crossed-flow conditions. It
is also important to recognize that the distributions obtained
in these experiments as well as the photoloc experiments, are
largely blind to internal energy in the undetected product.
Often the assumption is made that the undetected product is
internally cold. We will see below that this may be far from
the truth in some cases.
For the reactive scattering studies on pentane with
atomic chlorine, the pentane molecular beam was prepared
by bubbling helium through n-pentane ͑Aldrich, HPLC
grade͒ held at Ϫ40 °C, where it has a vapor pressure of about
19 Torr, with a total stagnation pressure of 150 Torr, yield-
ing a 13% beam. The continuous pentane beam was ex-
panded through a 0.25 mm stainless steel nozzle held at
200 °C to inhibit cluster formation, into a source chamber
Ϫ4
maintained at 1.0ϫ10 Torr. The mean velocity of the pen-
tane beam was 1215Ϯ61 m/s. After collimation by two
skimmers separated by a region of differential pumping, the
pentane beam was crossed by the continuous beam of atomic
chlorine, generated by thermal dissociation of Cl diluted in
2
We have used reaction of Cl atoms with saturated hydro-
carbons as a step toward a detailed investigation of these
metathesis reactions. Initial studies14 from our laboratory fo-
cused on the slightly exoergic reactions of Cl with propane
helium. A mixture of 5% Cl ͑Matheson, HP grade͒ in He
2
͑Matheson, HP grade͒ was used. The pyrolytic molecular
14
beam source consists of a high density graphite nozzle with
a 0.15 mm opening and a resistively heated silicon carbide
front plate. The nozzle tip is heated to about 1500 °C pro-
(
C H ), with probe of the C H product using synchrotron
3 8 3 7
radiation. The use of tunable undulator radiation, offering a
unique combination of universality and selectivity in product
detection, enables us to probe the doubly differential cross
sections for the hydrocarbon radical fragment. The propane
studies suggested different dynamics for the forward and
backward scattered products. The former were ascribed to
viding dissociation of a large fraction of the Cl . The mean
2
velocity of the chlorine beam was 2120Ϯ282 m/s.
The cross-correlation method17 was used for velocity
measurements for both reactant beams as well as the reactive
scattering products. A 17.8 cm diam cross-correlation wheel
with two identical 255-bit pseudorandom sequences of open
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:
132.174.255.116 On: Sun, 30 Nov 2014 17:30:03

5340
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
N. Hemmi and A. G. Suits
FIG. 1. Schematic energy diagram for the reaction Clϩn-C H
5
12
→
HClϩC H .
5 11
and closed slots was used in front of the detector entrance to
modulate the incident flux. The wheel was spun at 326.8 Hz
resulting in a nominal resolution of 6 ␮s in the time of flight
FIG. 2. Laboratory angular distribution for C H product ͑top͒ and Newton
5
11
͑TOF͒. The neutral flight length from the wheel to the photo-
diagram ͑bottom͒ for the title reaction. The circle on the Newton diagram
represents maximum recoil limit for production of 2- or 3-pentyl radicals.
Circles on the angular distribution are experimental data, open triangles are
the best-fit simulations. The connecting line is included to guide the eye.
ionization volume is 11.9 cm. For velocity analysis of either
molecular beam, the beam being analyzed was placed on line
with the detector axis, the aperture at the entrance to the
detector was set to 0.075 mm, and the photoionization en-
ergy was set to 14.5 eV with argon in the gas filter. The
argon gas filter filled at 30 Torr eliminates the undulator
harmonics above 15.0 eV.
1
0.71 kcal/mol, is taken from Ref. 19, while the energy dif-
ferences between this and the 1- and 3-pentyl radicals were
taken from Ref. 20. We have chosen the slightly lower value
for the heat of formation of the 2-pentyl radical based upon
Ref. 19 owing to a significantly more favorable fit to our
experimental results. Barriers for these two pathways are un-
known, but rate constant measurements and analogy to
smaller hydrocarbons suggest no barrier for the secondary H
atom abstraction pathways, and a very small barrier for the
primary abstraction channel.
Detection of the scattered C H reaction products was
5
11
accomplished using 9.5 eV photoionization radiation with a
MgF filter that has transmission cutoff at about 11 eV, thus
2
removing any residual radiation above that energy. The ion-
ization potential ͑IP͒ of n-pentane is 10.35 eV, while for the
pentyl radical it is 7.7 eV. By using a photoionization energy
of 9.5 eV there is almost no background from dissociative
ionization of the C H reactant. The quadrupole was set at a
5
12
The Newton diagram for the reaction is shown in Fig. 2,
along with the measured laboratory angular distribution. The
maximum available energy for the 2-pentyl radical product
from this slightly exothermic reaction gives the recoil limit
circle indicated. Also shown in conjunction with the mea-
sured angular distribution is the simulated distribution result-
ing from the best fit to all of the data. The inherently coupled
nature of the angular and energy distributions can be inferred
directly from inspection of the angular distribution: The for-
ward scattering extends beyond the beam in the forward di-
rection, nearly to the recoil limit. The backscattered distribu-
tion, in contrast, drops sharply at a lab angle of 60°, well
before the translational energy limit. This is confirmed by the
time-of-flight distributions shown in Fig. 3. These distribu-
tions serve to constrain the simulations from which the trans-
lational energy and angular distributions are obtained.
Although it was impossible to obtain a satisfactory fit to
the data with a single uncoupled translational energy ͓P(E)͔
and angular distribution ͓T(⌰)͔, excellent fits were obtained
simply by decomposing the center-of-mass distributions into
two components shown in Fig. 4. The first, shown as ‘‘chan-
nel 1,’’ is exclusively forward scattered, and shows a trans-
lational energy distribution peaking at the limit of the avail-
able energy. In fact, improved fits could be obtained by
mass resolution of about 1.0 amu to maximize transmission
ϩ
for products at m/e 71 (C H ). A multichannel scaler trig-
5
11
gered by the chopping wheel was used to record TOF at
various laboratory scattering angles, defined with respect to
the pentane beam. Signals were averaged for 1–3 h at each
angle and laboratory angular distributions were obtained by
integration of the TOF spectrum for each angle. The center-
of-mass translational energy and angular distributions were
obtained by the forward convolution technique.18
III. RESULTS
The reaction of chlorine atoms with n-pentane proceeds
via either primary or secondary H atom abstraction
Clϩn-C H →HClϩH CH CH CH CH CH ,
͑1͒
͑2͒
͑3͒
5
12
2
2
2
2
2
3
Clϩn-C H →HClϩH CHCH CH CH CH ,
5
12
3
2
2
2
3
Clϩn-C H →HClϩH CH CHCH CH CH .
5
12
3
2
2
2
3
Throughout the remainder of this paper, reactions ͑2͒ and ͑3͒
will be considered together since they both represent second-
ary H atom abstraction and are energetically indistinguish-
able. Figure 1 shows the energetics of these reactions. The
0
heat of formation ͓⌬Hf ͑298.15͔͒ of the 2-pentyl radical,
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:
132.174.255.116 On: Sun, 30 Nov 2014 17:30:03

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
N. Hemmi and A. G. Suits
5341
FIG. 5. Velocity-flux contour map for the title reaction. Inner circle shows
recoil velocity limit for the 1-pentyl radical; outer circle indicates corre-
sponding limit for 2- or 3-pentyl radicals. The contour line spacing is twice
as dense for the lowest 25% of the distribution to reveal the channel 2
contribution more clearly.
backscattered distribution, with a vastly different transla-
tional energy release. The average energy release obtained
for channel 1 is
available energy (Eavail), while for channel 2 it is
͗
E
͘1ϭ20.4 kcal/mol, which is 92% of the
͘2
͗
E
ϭ7.7 kcal/mol, which is 35% of Eavail . The branching frac-
tions are 58% for channel 1 and 32% for channel 2.
As in any forward convolution simulation of crossed-
molecular beam results, the fit obtained is not unique. Nev-
ertheless, the velocity-flux contour map we obtain, shown in
Fig. 5, does faithfully reproduce the measured distributions.
However, for the forward-scattered products in particular the
decomposition into two components is somewhat arbitrary,
and driven by the fit to the backscattered distribution.
ϩ
11
FIG. 3. Time-of-flight spectra for m/e 71 (C H ) from the title reaction at
5
the indicated laboratory scattering angles ͑defined with respect to the pen-
tane beam͒. The filled circles are experimental points, the dashed line rep-
resents the contribution to the fit from ‘‘channel 1,’’ the dotted line that
from ‘‘channel 2’’ and the solid line represents the total simulation using the
distributions in Fig. 4.
IV. DISCUSSION
allowing the exothermicity of the reaction to increase an ad-
ditional 0.5 kcal/mol beyond that shown in Fig. 1, implying a
heat of formation for the 2-pentyl radical of 10.2 kcal/mol.
However, since the fitting approach is not unique and this
difference is small, we do not feel justified in arguing for a
lower heat of formation for the 2-pentyl radical. In contrast
to channel 1, channel 2 shows a broad but predominantly
The dynamics for the forward-scattered channel obtained
from the fit are largely consistent with previous studies of
Cl-hydrocarbon reaction dynamics. The P(E) peaks at the
maximum energy available for formation of the isoenergetic
2
-pentyl ͑butyl-1-methyl͒ or 3-pentyl ͑propyl-1-ethyl͒ radi-
cals, with an average of only 1.8 kcal/mol remaining in ro-
tation and vibration of the products. The bulk of the forward
scattered distribution appears at translational energies ex-
ceeding the limit for formation of the 1-pentyl radical ͓see
the limits in Figs. 4͑B͒ and 5͔. Abstraction of primary H
atoms is thus ruled out for channel 1; however, there is little
here to suggest any different propensities for abstraction of
the hydrogen atoms from the 2- or 3-carbons.
The results for this channel are reminiscent of those re-
5
ported by Flynn et al. for the Cl-cyclohexane reaction, and
1
1–13
identified by others in propane
and even for vibra-
tionally excited methane.7 That is, the products are formed
with very little internal energy. As mentioned in the Intro-
duction, many have attributed this to a collinear transition
state; yet the correlation between forward scattering and a
collinear transition state is problematic. Zare and co-workers
considered several possible mechanisms that could reconcile
these, but finally suggested an alternative interpretation of
–10
7
the dynamics. They argued that if the reaction is not in the
impulsive limit, then a cold product rotational distribution
need not imply a collinear transition state. They considered
the relative speed of the H and Cl at the time of the collision
to estimate the anticipated rotational excitation. A related
view of the collision dynamics can be obtained by asking
what the appropriate impact parameter is for the collision. In
FIG. 4. Center-of-mass angular ͑A͒ and translational energy ͑B͒ distribu-
tions obtained from the fit to the data in Figs. 2 and 3. Arrows mark trans-
lational energy limits for production of indicated product isomers.
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:
32.174.255.116 On: Sun, 30 Nov 2014 17:30:03
1

5342
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
N. Hemmi and A. G. Suits
fact, one may consider a hierarchy of impact parameters: that
between the Cl and the whole molecule, which determines
the overall angular momentum; that between the Cl atom and
the carbon to which the abstracted H is attached, or even that
between the Cl and H atoms. The most appropriate choice
depends on the magnitude of the barrier, the tightness of the
transition state and the particular feature of the dynamics
under consideration. For example, collisions that have a
large impact parameter with respect to the pentane center-of-
mass ͑most collisions, obviously͒ may nevertheless exhibit
small impact parameters with respect to the abstracted H
atom. If there is no barrier for reaction even for a perpen-
dicular C–H–Cl geometry, then the reaction may proceed
with no momentum transfer whatsoever, other than that as-
sociated with change in the H atom motion. The orbital an-
gular momentum will be largely unchanged, yielding for-
ward scattering in a ‘‘spectator-stripping’’ mechanism.
Calculations have indicated a collinear transition state for
the closely related Cl-propane reaction, but the question re-
ally is how tight the transition state is; we suggest the coin-
cident observation of a cold rotational distribution with for-
ward scattering argues for a fairly loose transition state.
The dynamics for channel 2, the backscattered distribu-
tion, diverge dramatically from previous work on H abstrac-
tion dynamics in hydrocarbons, even differing significantly
from our own studies of the analogous Cl-propane reaction.
The P(E) for channel 2 peaks at 6.8 kcal/mol, with an av-
erage release of 7.7 kcal/mol, a fractional energy release of
only 35%. This may be compared to results for Cl-propane,
for which the fractional energy release for the backscattered
component ranged from 52% at 11.5 kcal/mol collision en-
ergy to 48% at the highest 31 kcal/mol collision energy. This
means that an average of 15 kcal/mol remains in internal
degrees of freedom of the products. State-resolved studies
On the other hand, plausible reasons for greater vibra-
tional excitation in the case of pentane vs propane are readily
apparent. There are many more vibrational modes for n-
pentane ͑45 vs 27͒, and perhaps more significantly, many
more involving the low-frequency C–C–C bends. For these
modes, with vibrational frequencies on the order of 800
Ϫ1
cm , the half-period is roughly 20 fs. We can compare this
to the collision time under the conditions of the experiment:
For an effective interaction region of 0.5 Å at 2400 m/s we
estimate a collision time of 21 fs, a remarkable match to the
vibrational period. The combination of the higher density of
states and the excellent match between the collision time and
the bending mode vibrational period greatly favors coupling
between the collision energy and the internal modes of the
hydrocarbon. This suggests that in pentane we are seeing the
direct participation of the extended carbon backbone in the
reaction.
A final question concerns the nature of the two different
channels. In our previous work on the propane reaction, we
argued for the association of the forward-scattered distribu-
tion with the abstraction of secondary H atoms, while the
backscattered product was associated with primary H atoms.
It is even clearer in the pentane case, since the bulk of the
forward scattered products are formed at translational ener-
gies well exceeding that possible for production of the 1-
pentyl radicals, that the forward scattered products are in-
deed the result of secondary H atom abstraction. The
question remains for the backscattered products: Are they
preferentially associated with primary H abstraction? Al-
though attempts were made to discriminate between these
two radical isomers based upon their ionization potentials,
we found no significant differences in the angular distribu-
tions as a function of probe photon energy. However, the
vastly different internal energy distributions for the forward
and backward scattered products may account for the ab-
sence of any significant differences, and would make inter-
pretation of such an effect problematic in any case.
2
1
for propane have shown only a small yield of vibrationally
excited HCl product,1
2,13
so it is unlikely that this accounts
for much of the energy. Furthermore, we anticipate little dif-
ference between the HCl energy content from propane or
pentane, since the energetics are virtually the same for both
reactants. The likely repository of the bulk of this energy is
thus in the hydrocarbon fragment; the question is whether it
is in rotation or vibration. One significant difference between
n-pentane and propane is the average impact parameter: The
long n-pentane molecule naturally will yield abstraction of
primary H atoms a considerable distance from the center of
mass of the molecule. One might expect this to lead to
greater rotational excitation for the pentane case. However,
these large impact parameter ͑b͒ collisions will also yield
fairly large exit impact parameters (bЈ) so that little of the
necessary conversion of orbital to rotational angular momen-
tum may occur. Large rotational excitation in general re-
quires small impact parameters for effective coupling of the
initial linear momentum into rotational excitation. It is only
for these close collisions that we can achieve the required
large value of (bϪbЈ). Certainly some measure of enhanced
rotational excitation is plausible in the case of n-pentane, but
for this to account for an increase to 65% of the available
energy seems unlikely.
Although there is no direct evidence indicating a connec-
tion between the predominantly backscattered ‘‘channel 2’’
and the primary abstraction channel, several lines of evi-
dence point this direction. Firstly, the maximum energy re-
lease shown in Fig. 4͑B͒, and apparent in the contour map in
Fig. 5, extends well below the limit of the available energy
for production of the higher energy isomer. This distribution
came directly from the best fit to the laboratory data and was
in no way enforced. Secondly, there is the total branching
between forward and backward scattered products. For the
propane reaction, the fraction scattered forward of 90° was
very close to 50% of the total. Assuming simply a limiting
case association of forward scattering with secondary H ab-
straction we find, effectively, three times higher reactivity of
the secondary vs primary H atoms. For pentane, then, we
1
3
expect the fraction forward scattered to increase from 2 to ,
4
since there are three times the number of secondary H atoms,
if this limiting assumption is a good one. Indeed, if we look
only at the fraction of the total scattered in the forward di-
rection, in the same way we considered the propane case, we
find a value of 74%, a remarkable agreement. If we compare
the fraction of the total scattered into ‘‘channel 1,’’ then the
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:
132.174.255.116 On: Sun, 30 Nov 2014 17:30:03

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
N. Hemmi and A. G. Suits
5343
value is 64%. If we make the association between channel 1
and secondary H abstraction, and channel 2 and primary H
abstraction, this implies roughly a 1.8-fold higher reactivity
of the secondary H atoms. However, we cannot directly com-
pare this to the propane results since the distributions were
decomposed differently. These values, on the order of two to
threefold higher reactivity of secondary vs primary H atoms,
kcal/mol using the crossed molecular beam technique. The
experiments were conducted on the Chemical Dynamics
Beamline at the Advanced Light Source, using tunable un-
dulator radiation to effect soft ionization of the pentyl radical
product. The center-of-mass translational energy and angular
distributions obtained from the experiments were found to be
strongly coupled, with the forward scattered pentyl radical
formed extremely cold, while the backscattered radicals were
formed leaving nearly 65% of the available energy in inter-
nal energy in the products. These results are contrasted with
recent studies on the analogous reaction with propane, sug-
gesting in this case direct involvement of the carbon skeleton
in the collision process.
6
are consistent with those reported by Koplitz et al. as well
as Varley and Dagdigian13 for selectively deuterated pro-
pane.
We may also consider the role of the transition state and
the collision dynamics in an attempt to understand the nature
of the backscattered component. Given a small enough bar-
rier for primary H abstraction, virtually any direct encounter
between Cl and a primary H atom could be reactive if the
transition state is sufficiently loose. That is, even considering
only the relative energy between Cl and H, we find under our
conditions an effective collision energy of 0.7 kcal/mol or
ACKNOWLEDGMENTS
We thank Dr. D. Blank for valuable experimental ad-
vice, Dr. Y. T. Lee and Dr. C. Taatjes for helpful discus-
sions, and the staff of the Advanced Light Source. This work
was supported by the Director, Office of Energy Research,
Office of Basic Energy Sciences, Chemical Sciences Divi-
sion, of the U.S. Department of Energy under contract No.
DE-AC03-76SF00098.
Ϫ1
about 250 cm . Collisions that possess a zero Cl–H impact
parameter but achieve a perpendicular geometry at the tran-
Ϫ1
sition state still have a relative energy of at least 250 cm . If
the barrier is lower than this, then the scattering distribution
directly reflects the geometry of the transition state rather
than the need for translational energy along the line-of-
centers. The broad, predominantly backscattered distribution
we see in ‘‘channel 2,’’ thus implies a tighter transition state
for this channel than for channel 1. Taken together these
points argue convincingly for the association of the slower,
preferentially backscattered product with abstraction of pri-
mary H atoms, while the very fast, forward scattered product
is unambiguously associated with secondary H abstraction.
These translational energy distributions argue for a bit of
caution on the part of those employing the photoloc tech-
nique and other methods relying on state-resolved measure-
ments of one of the fragments to extract the speed and an-
gular distributions for the reaction. These studies are
generally blind to internal energy in the undetected fragment.
Moreover, the reconstruction of the product distributions re-
quire some assumptions about the unmeasured fragment. If
these assumptions are in error, then the derived distributions
are inaccurate. Usually the assumption is made that the un-
detected fragment is internally cold, although in at least one
1
W. Bartok and A. F. Sarofim, Fossil Fuel Combustion ͑Wiley-
Interscience, New York, 1991͒, p. 88.
P. Andresen and A. C. Luntz, J. Chem. Phys. 72, 5842 ͑1980͒, A. C. Luntz
2
and P. Andresen, ibid. 72, 5851 ͑1980͒.
T. N. Truong, D. G. Truhlar, K. K. Baldridge, M. S. Gordon, and R.
3
Steckler, J. Chem. Phys. 90, 7137 ͑1989͒.
K. D. Dobbs and D. A. Dixon, J. Phys. Chem. 98, 12584 ͑1994͒.
J. Park, Y. Lee, J. F. Hershberger, J. M. Hossenlopp, and G. W. Flynn, J.
4
5
Am. Chem. Soc. 114, 58 ͑1990͒.
Y. Fen, Z. Wang, B. Xue, and B. Koplitz, J. Phys. Chem. 98, 4 ͑1994͒.
W. R. Simpson, A. J. Orr-Ewing, and R. N. Zare, Chem. Phys. Lett. 212,
6
7
1
63 ͑1993͒.
8
9
0
W. R. Simpson, T. P. Rakitzis, S. A. Kandel, A. J. Orr-Ewing, and R. N.
Zare, J. Chem. Phys. 103, 7313 ͑1995͒.
W. R. Simpson, T. P. Rakitzis, S. A. Kandel, T. Lev-On, and R. N. Zare,
J. Phys. Chem. 100, 7938 ͑1996͒.
1
S. A. Kandel, T. P. Rakitzis, T. Lev-On, and R. N. Zare, J. Chem. Phys.
105, 7550 ͑1996͒.
11
D. F. Varley and P. J. Dagdigian, J. Phys. Chem. 99, 9843 ͑1995͒.
D. F. Varley and P. J. Dagdigian, J. Phys. Chem. 100, 4365 ͑1996͒.
D. F. Varley and P. J. Dagdigian, Chem. Phys. Lett. 255, 393 ͑1996͒.
D. A. Blank, N. Hemmi, A. G. Suits, and Y. T. Lee, Chem. Phys. 231, 261
1
1
1
2
3
4
͑
1998͒; D. A. Blank, PhD Thesis, University of California, Berkeley,
8
case this issue was considered explicitly. As we have seen,
1997.
15
for the forward-scattered products in the pentane and pro-
pane case, this is a fairly accurate assumption. For the back-
scattered distributions, however, this is grossly incorrect.
The implications of this for the inferred distributions are not
clear, but these issues should be considered for any quanti-
tative analysis of the photoloc-type experiments.
P. A. Heimann, M. Koike, C. W. Hsu, M. Evans, C. Y. Ng, D. Blank, X.
M. Yang, C. Flaim, A. G. Suits, and Y. T. Lee, Proc. SPIE 2856, 90
͑
1996͒.
16
X. Yang, D. A. Blank, J. Lin, A. G. Suits, Y. T. Lee, and A. M. Wodtke,
Rev. Sci. Instrum. 68, 3317 ͑1997͒.
G. Comsa, R. David, and B. J. Schumacher, Rev. Sci. Instrum. 52, 789
17
͑
1981͒.
1
1
8
9
R. Buss, Ph.D. Thesis, University of California, Berkeley, 1972.
S. G. Lias, J. F. Liebman, R. D. Levin, S. A. Kafafi, and S. E. Stein,
Structures and Properties, NIST Database 25, 1994.
V. CONCLUSION
2
2
0
1
J. A. Seetula and I. R. Slagle, J. Chem. Soc., Faraday Trans. 93, 1709
We have studied the dynamics of the metathesis reaction
͑
1997͒.
ClϩnϪC H →HClϩC H at a collision energy of 16.8
A. Bottoni and G. Poggi, J. Mol. Struct. 337, 161 ͑1995͒.
5
12
5
11
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:
32.174.255.116 On: Sun, 30 Nov 2014 17:30:03
1