The rate coefficient of the C3H3 + C 3H3 reaction from UV absorption measurements after photolysis of dipropargyl oxalate

Article abstract of DOI:10.1039/b308518j

The kinetics of the C₃H₃ + C₃H ₃ reaction was investigated by using dipropargyl oxalate (DPO) as a new, halogen-free photolytic source for propargyl radicals in the gas phase. After laser-flash photolysis of DPO at 193 nm, the initial absorbance was determined at different wavelengths, and the results were compared with values obtained in analogous experiments using propargyl halides as precursors. A satisfactory agreement of the absorbances was found between 295 and 355 nm but differences were observed near 242 nm. The latter wavelength has also been proposed for C₃H₃ detection. Our results, however, indicate that this absorption is probably due to halogen-containing species. The rate coefficient of the C₃H₃ + C₃H ₃ reaction was then determined from time-resolved absorption measurements at 332.5 nm with DPO as precursor. Values of (2.7 ± 0.6) × 10^-11 cm³ molecule^-1 s^-1 at 373 K, (2.8 ± 0.6) × 10^-11 cm³ molecule ^-1 s^-1 at 425 K, (3.5 ± 0.8) × 10 ^-11 cm³ molecule^-1 s^-1 at 500 K, and (4.1 ± 0.8) × 10^-11 cm³ molecule ^-1 s^-1 at 520 K were obtained with no significant pressure dependence between 1 and ca. 100 bar (140 bar for T = 373 K).

Full text of DOI:10.1039/b308518j

View Article Online / Journal Homepage / Table of Contents for this issue
The rate coefficient of the C H + C H reaction from UV absorption
3
3
3
3
measurements after photolysis of dipropargyl oxalate
B. R. Giri, H. Hippler, M. Olzmann* and A. N. Unterreiner
Institut f u¨ r Physikalische Chemie, Universit a¨ t Karlsruhe, Fritz-Haber-Weg 4, D-76128,
Karlsruhe, Germany. E-mail: olzmann@chemie.uni-karlsruhe.de
Received 23rd July 2003, Accepted 29th August 2003
First published as an Advance Article on the web 17th September 2003
3 3 3 3
The kinetics of the C H + C H reaction was investigated by using dipropargyl oxalate (DPO) as a new,
halogen-free photolytic source for propargyl radicals in the gas phase. After laser-flash photolysis of DPO
at 193 nm, the initial absorbance was determined at different wavelengths, and the results were compared with
values obtained in analogous experiments using propargyl halides as precursors. A satisfactory agreement of
the absorbances was found between 295 and 355 nm but differences were observed near 242 nm. The latter
wavelength has also been proposed for C
is probably due to halogen-containing species. The rate coefficient of the C H + C H reaction was then
3 3
H detection. Our results, however, indicate that this absorption
3
3
3
3
determined from time-resolved absorption measurements at 332.5 nm with DPO as precursor. Values
ꢂ11
3
cm molecule
ꢂ1 ꢂ1
ꢂ11
ꢂ11
3
ꢂ1 ꢂ1
of (2.7 ꢀ 0.6) ꢁ 10
s
at 373 K, (2.8 ꢀ 0.6) ꢁ 10
cm molecule s
at 425 K,
ꢂ1 ꢂ1
cm molecule s at 520 K
ꢂ11
3
cm molecule
ꢂ1 ꢂ1
3
(
3.5 ꢀ 0.8) ꢁ 10
s
at 500 K, and (4.1 ꢀ 0.8) ꢁ 10
were obtained with no significant pressure dependence between 1 and ca. 100 bar (140 bar for T ¼ 373 K).
ꢂ11
3
cm molecule s . Most recently, Knya-
ꢂ1 ꢂ1
(
4.0 ꢀ 0.4) ꢁ 10
Introduction
1
3
zev et al. investigated reaction (1) by laser photolysis/
photoionization mass spectrometry in the temperature range
500–1000 K and at pressures between ꢃ2 and ꢃ10 mbar.
The mechanism of soot formation in the pyrolysis and com-
bustion of hydrocarbons has become an important topic of
1
–3
ꢂ11
research. A critical step in this mechanism is the formation
of the first aromatic ring from small aliphatics. Among the
different reaction routes suggested, the self-combination of
propargyl radicals
The rate coefficient determined decreases from 3.3 ꢁ 10
ꢂ1 ꢂ1
cm molecule s
3
ꢂ1 ꢂ1
ꢂ11
3
cm molecule
s
at 500 K to 1.2 ꢁ 10
at 1000 K. The latter value agrees very well with results from a
1
shock tube study by Scherer et al. These authors report a rate
4
ꢂ11
3 ꢂ1 ꢂ1
cm molecule s for
coefficient of (1.2 ꢀ ꢃ0.3) ꢁ 10
15
C3H3 þ C3H3 ! products
is considered to be of crucial importance.
ð1Þ
Due to its reso-
T ¼ 1000–1250 K and P ꢃ 1 ꢂ 2 bar. Furthermore, DeSain
1
,2,4,5
ꢂ11
3
and Taatjes obtained a value of (3.9 ꢀ 0.6) ꢁ 10
cm
at 296 K and a pressure of 8 mbar, using infra-
ꢂ1 ꢂ1
nance stabilization energy, which is of the order of 50 kJ
mol , the propargyl radical is a comparatively stable species
molecule
s
ꢂ1 6
red laser absorption spectroscopy to detect C H . In a recent
3 3
7
–9
16
and also exhibits a low reactivity towards O
2
.
Therefore, it
theoretical work Miller and Klippenstein calculated the rate
coefficient and the product branching ratios of reaction (1)
over a large parameter range by solving the time-dependent
master equation. For temperatures below 500 K, a virtually
can reach high concentrations in flames, and the bimolecular
self-reaction appears to be the dominant loss channel.
Despite its important role in combustion systems, kinetic
investigations of reaction (1) are scarce and mostly restricted
ꢂ11
3
pressure-independent rate coefficient of ca. 5 ꢁ 10
cm
with a slightly positive temperature dependence
1
0
ꢂ1 ꢂ1
to low pressures. In an early work Alkemade and Homann
ꢂ11
molecule
s
3 ꢂ1 ꢂ1
cm molecule s
obtained a rate coefficient of 5.7 ꢁ 10
is predicted.
In the present work we use UV absorption to monitor
. The first investigation of the electronic spectrum of
the propargyl radical was performed by Ramsay and Thistle-
thwaite. These authors found diffuse bands in the region
290 to 345 nm during flash photolysis of several propargyl
radical precursors. Their results were essentially confirmed
by an argon cryogenic matrix study of Jacox and Milligan.
More recently, Fahr et al. monitored the transient UV
absorption between 230 and 300 nm after laser-flash photolysis
of C H Cl at 193 nm and mixtures of C H Cl with COCl at
3 3 3 3 2
248 nm, respectively. They observed a relatively broad absorp-
tion band with a maximum at 242 nm which was attributed to
(cross section: 1.2 ꢁ 10
same assignment was made by Deyerl et al. to interpret
H-atom yields in C photodissociation experiments, where
the propargyl radicals were produced by supersonic jet flash
at temperatures ranging from 623 to 673 K and pressures
between 3 and 6 mbar, using a low-pressure flow reactor with
3 3
C H
mass-spectrometric detection. The C
ated by the reaction of sodium atoms with propargyl halides.
3 3
H radicals were gener-
1
7
1
In a subsequent study Morter et al. performed laser-flash
1
photolysis experiments of propargyl halides at 193 nm and
1
8
monitored C
a rate coefficient of (1.2 ꢀ 0.2) ꢁ 10
was obtained at a temperature of 295 K and pressures near
0 mbar. In a more recent investigation, Atkinson and Hud-
3
H
3
by time-resolved infrared laser absorption;
ꢂ10
3
ꢂ1 ꢂ1
19
cm molecule
s
2
8
ꢂ11
3
gens determined a rate coefficient of (4.3 ꢀ 0.6) ꢁ 10
cm
using time resolved cavity ring-down spectro-
ꢂ1 ꢂ1
molecule
scopy to monitor C
s
ꢂ17 ꢂ1
2
19
H
3 3
at 332.5 nm after photolysis of pro-
pargyl chloride, propargyl bromide, propyne, and allene,
C
3
H
3
cm molecule , base e). The
2
0
1
2
respectively. Fahr and Nayak performed photolysis experi-
ments using C H Cl/CH N CH mixtures and determined
the relative yields of the stable products by GC/MS. From a
complex modeling calculation they obtained a value of
3 3
H
3
3
3
2
3
2
1
3 3
pyrolysis of C H Br. Atkinson and Hudgens, however,
obtained strong evidence that the absorption around 240 nm
DOI: 10.1039/b308518j
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646
This journal is # The Owner Societies 2003
4641

View Article Online
after flash photolysis of C
3
H
3
X (X ¼ Cl, Br) has to be attrib-
through the cell. The laser was coupled in by a dielectric mirror
with a high reflectivity at 193 nm. A second mirror of the same
type located behind the reactor prevented the arc lamp from
being exposed to the laser light and directed the beam to an
energy meter (Soliton ED-500). The transmission of the mir-
rors for wavelengths between 310 and 360 nm (detection range)
was 80 to 90%, and in order to avoid unwanted photolysis
of the precursor by the arc lamp, an electromagnetic shutter
was opened only a few milliseconds before the laser shot.
The absorption–time profiles were recorded using a combi-
nation of a standard prism monochromator in Czerny–Turner
configuration [Zeiss M4 Q III; slit entrance: 0.25 mm;
Dl ¼ 2.5 nm at 330 nm (full width at half maximum)] and a
UV-sensitive photomultiplier (PMT, Hamamatsu 1P28A).
The data were sampled by a digital storage oscilloscope
(LeCroy 9361, 300 MHz) and further processed in a personal
computer.
uted to adduct radicals formed in a consecutive reaction
of the photolytically generated halogen atoms with excess
precursor:
X þ C3H3X ! C3H3X2
Additionally, a smaller contribution to the absorption may
ð2Þ
2
1
be due to C H . These authors also recorded the propargyl
3
2
8
spectrum between 310 and 360 nm, using several precursors.
They found a maximum absorption at 332.5 nm with an absorp-
ꢂ18
2
cm molecule (base e).
ꢂ1
tion coefficient of (4.1 ꢀ 0.6) ꢁ 10
In the present work we report on the experimental determi-
nation of the rate coefficient for reaction (1) in the temperature
interval 373–520 K and at pressures between 1 and 140 bar.
The extended pressure range allows us to reliably detect or
exclude falloff effects in the self-reaction of C H at these
3
3
temperatures and to verify the theoretical results of ref. 16
Moreover, we introduce dipropargyl oxalate (DPO),
HCCCH O(O)C–C(O)OCH CCH, as a new halogen-free
2 2
photolytic source for propargyl radicals. This idea is based
Propargyl chloride was of commercial grade (98%, Lancas-
ter) and subjected to several freeze–pump–thaw cycles before
use. Mixtures containing 5–10% of C H Cl in N (99.996%,
3
3
2
2
2–24
on earlier work using the thermal decomposition of diallyl
Messer Griesheim) were prepared and stored in a commercially
available gas cylinder at room temperature. From this cylinder
the amounts necessary for the corresponding initial concen-
trations were admitted into the preheated reactor, and N2
was added up to the desired pressure.
2
5
and dipropargyl oxalate to produce the respective unsatu-
rated radicals. With this precursor the above mentioned
complications by simultaneously generated halogen atoms
3 3
are avoided, and the spectroscopic features of C H are
not obscured by halogen adducts. This enables us to study
the UV absorption of C H under halogen-free conditions.
In the case of DPO such storable premixtures could not be
prepared due to its low vapour pressure (<0.02 mbar at room
temperature). Instead, a sample compartment containing the
solid precursor was directly connected to the evacuated reactor
and heated to ꢃ360 K. Because the reactor was already kept
at the final temperature (T ꢄ 373 K), condensation of DPO on
its walls could be avoided. After pressure equilibration, the
3
3
Experimental
Experimental technique
The investigations have been performed in a static reaction cell
using a laser-flash photolysis/UV absorption setup as shown
in Fig. 1. The cylindrical high-pressure cell was made of stain-
less steel and is equipped with Quartz windows (Suprasil I, 12
mm thick). It has an inner diameter of 8 mm and an optical
path length of 10.1 cm. External heating wires allow tempera-
tures up to 800 K, which can be controlled by NiCr–Ni
thermocouples. The cell was connected to a standard gas
handling system provided with appropriate pressure gauges
connection was closed and the reactor filled with N
final pressure.
up to the
2
To ensure complete mixing, the content of the reaction cell
was generally allowed to stand for at least 15 min. before each
experiment. We further note that the windows of the cell had
to be cleaned each time after 2 or 3 experiments due to soot
deposition. The cleaning was accomplished by filling the reac-
2
tor with O (4–5 bar) and exposing it to 3000–4000 laser shots.
Additionally, the windows were manually cleaned by disassem-
bling the reactor at regular intervals and using acetone as a
solvent.
(
VDO, 0–1000 mbar and 1–1000 bar).
Propargyl radicals were generated by the photolysis of
C H Cl and DPO, respectively, at 193 nm using an ArF exci-
3
3
mer laser (Lamda Physik, Compex 102). The laser fluence was
ca. 14 mJ cm , and an aperture with a diameter of 10 mm was
used to reduce the laser spot size and to minimize effects
of scattered light. Laser beam and detection light (Hg–Xe
high-pressure arc lamp, UXM-200H) propagated anti-parallel
Synthesis of DPO
ꢂ2
2
6–28
Only very few synthesis routes for DPO are known.
method we used is based on a procedure suggested by Yin
and will be briefly summarized in the following. A mixture
of 12.6 g oxalic acid, 30 ml propargyl alcohol and 25 ml
benzene was refluxed at 120 C for 12–14 h using a water
separator. After water formation had stopped, the mixture
was cooled down to 0 C. Then, benzene and the excess pro-
pargyl alcohol were decanted from the precipitated solid con-
sisting of DPO and oxalic acid. This crude product was heated
The
2
8
ꢅ
ꢅ
ꢅ
until the residual benzene (at ꢃ80 C) and propargyl alcohol
ꢅ
(
at ꢃ117 C) were evaporated. After the molten mixture, upon
solidification, had cooled down to room temperature, it was
dissolved in 80 ml ethyl acetate. This solution was washed with
2 3
aqueous Na CO solution (1%) and water until a neutral aqu-
eous phase was obtained. The organic phase was dried with
Na
2
SO
4
and CaCl
2
and reduced to half of its volume by eva-
ꢅ
poration. Finally, DPO was allowed to crystallize at ꢂ20 C
in a refrigerator. We obtained white crystals, which were
cleaned by several recrystallization cycles using ethyl acetate.
The overall yield was 46%.
The melting point of the product was 96 C in good agree-
2
ment with values published earlier,
by H-NMR (CDCl , 250 MHz) gave two distinct peaks at
ꢅ
6–27
and characterization
1
3
Fig. 1 Experimental setup; for explanations see text.
2
4.8 ppm (d, 4H, 2CH ) and 2.5 ppm (t, 2H, 2CH). The UV
4
642
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646

View Article Online
ꢂ1
ꢂ1
6
absorption spectra recorded in acetonitrile exhibits a broad
band with a maximum at 185 nm (see Fig. 2).
energy. With a value of ꢃ50 kJ mol for this quantity, a bond
dissociation energy of ꢃ300 kJ mol results.
The photolysis wavelength of 193 nm corresponds to an
excitation energy of 619 kJ mol . If one assumes that DPO
ꢂ1
dissociates after a fast relaxation into the electronic ground
state, C H split-off is the dominating process, because it has
Results and discussion
3
3
the lowest threshold energy. From the excess energy of ꢃ320
DPO as C
3
H
3
precursor
ꢂ1
ꢂ1
kJ mol , the largest part (ꢃ220 kJ mol , statistical energy
partitioning assumed) is deposited into the complementary
Our idea to use the photolysis of dipropargyl oxalate for the
production of propargyl radicals was essentially based on ear-
fragment, O(O)CC(O)OC
ate into a second C radical and 2 CO
tive mechanism can be formulated if the central C–C bond in
DPO is assumed to be broken first. Therefore, 2 C radicals
H
3 3
, which then can further dissoci-
2
2–24
H
3
2
. A similar consecu-
lier work,
cally produce allyl radicals. Later we became aware of one
in which diallyl oxalate was used to thermolyti-
3
2
5
3 3
H
work, in which indeed the thermolysis of the dipropargyl
ester was employed to generate propargyl radicals (for a
matrix-isolation study). No information, however, is available
on the kinetics and mechanism of this reaction. As DPO does
not contain H atoms in the b-position to the carbonyl group,
per DPO molecule are probably formed. We note, however,
that knowledge of the quantum yield is not required for our
kinetic analysis; nonetheless, a detailed investigation of the
photodissociation pathway is underway in this laboratory.
As already mentioned in the introduction section, there is
a discrepancy regarding the UV absorption spectrum of
2
9
the most common ester-decomposition channel, i.e., the ole-
fin elimination via a 6-membered transition state, is not feasi-
ble. Also other molecular elimination pathways proposed for
such esters are not likely to occur due to the stiff structure
of the linear propargyl groups. Hence, simple bond-fission
reactions forming radicals appear to be the dominating uni-
molecular steps.
8
,19–21
C
3
H
3
.
It was argued that this is probably due to inter-
fering effects of halogen atoms in some of the previous experi-
2
9
2
1
ments. Therefore, it seems worthwhile to compare spectra
obtained by using both halogen-containing and halogen-free
precursors.
To this end we measured the initial absorbance after laser-
flash photolysis of C H Cl and DPO, respectively, at several
3 3
To get a rough estimate for the bond dissociation energies,
we have to refer to dimethyl oxalate (DMO) as the only
oxalic acid ester for which at least some of the necessary
thermochemical data are available. With the heats of
wavelengths in the ranges 230–242 nm and 295–355 nm. A sig-
nificant absorption in the former interval could only be
observed in the case of C H Cl, whereas either precursor leads
n
ꢂ1
formation
DH298 K [C(O)OCH
sociation energy of 371 kJ mol
DH298 K [DMO] ¼ ꢂ709
kJ
mol and
a bond dis-
3
3
n
ꢂ1 30
to an absorption between 295 and 355 nm. This is an indica-
tion that indeed the presence of chlorine atoms is responsible
for the absorption around 240 nm; probably adduct species
3
] ¼ ꢂ169 kJ mol
,
for the central C–C
ꢂ1
bond in DMO is obtained. As the heats of formation
for the other possible radicals are not known, we
adopt the thermochemical quantities from methyl
2
1
from Cl and C H Cl are formed according to eqn. (2). The
3
3
results for wavelengths in the range 295–355 nm are compared
in Fig. 3. The good agreement between C H Cl and DPO as
n
ꢂ1
acetate. With DH
[CH C(O)OCH ] ¼ ꢂ413 kJ mol
,
3
3
2
98 K
3
3
ꢂ1
n
n
precursor shows that propargyl radicals are formed in either
case. We note that the relative absorbances from our measure-
ments were scaled by arbitrarily setting the maximum value at
DH298 K [CH
3
C(O)O] ¼ ꢂ208 kJ mol , DH298 K [CH
3
] ¼
ꢂ1
n
ꢂ1
1
46 kJ mol , DH
[CH CO] ¼ ꢂ10 kJ mol , and
2
98 K
3
n
ꢂ1 30
DH298 K [CH O] ¼ 17 kJ mol
kJ mol and 420 kJ mol for the bond dissociation energies
3
,
one obtains values of 351
ꢂ1
ꢂ1
3
son is made with a C spectrum from ref. 8 (propyne as pre-
32.5 nm equal to unity. In Fig. 3 also a qualitative compari-
H
3
3 3 3 3
of the CH C(O)O–CH and the CH C(O)–OCH bond,
3
cursor). Here the relative absorbance was normalized so as to
grossly reproduce the band area from our measurements. The
agreement is reasonable taking into account the different spec-
tral resolution (ꢀ1.5 nm FWHM in this work as compared to
the ꢀ0.5 nm uncertainty in ref. 8) as well as the higher
respectively. In transferring these results to DPO, it is surely
reasonable to assume that a bond dissociation energy of
ꢂ1
ꢃ370 kJ mol for the central C–C bond remains a good esti-
mate, whereas the value for the oxy radical split-off is likely to
be somewhat lowered due to resonance stabilization in the
complementary carbonyl radical. The largest decrease of the
bond dissociation energy, however, can be expected in going
3 3 3
from CH to C H split-off. In this case the difference should
lie at least in the order of the propargyl resonance stabilization
Fig. 3 Initial absorbance between 295 and 355 nm after laser-flash
photolysis of dipropargyl oxalate (S) and propargyl chloride (N),
respectively, relative to the initial absorbance at 332.5 nm (P ¼ 1
Fig. 2 UV absorption spectrum of DPO; solvent: acetonitrile,
[
bar, bath gas N
of C
2
; T ¼ 373 K); grey solid line: absorption spectrum
3 4
ꢂ4
ꢂ1
DPO] ¼ 10 mol l , optical path length: 1 cm.
3
H
3
after photolysis of C
H
from ref. 8 (for explanations see text).
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646
4643

View Article Online
temperature and pressure in our experiments which addition-
ally cause line broadening.
To obtain temperature-dependent absolute values for the
our kinetic experiments were performed under identical spec-
tral conditions, such an effective absorption cross-section can
be used for the kinetic data analysis. Additionally, we checked
the validity of the Beer–Lambert law by performing calibra-
tion measurements for different initial concentrations of
absorption cross-section of C
required for the kinetic analysis, we performed photolysis
experiments with C Cl as radical source. The absorption
3 3
H at 332.5 nm, which are
H
3 3
3 3
C H . No deviations from linearity could be found within
cross-sections were determined from the initial absorption
after the photolysis flash by employing the Beer–Lambert
the concentration range of our experiments. As the tempera-
ture dependence of the absorption cross-section is not signifi-
law. Initial concentrations [C
3
H
3
]
0
were calculated from the
cant (cf. Fig. 4), we used an average value of hs
e
i ¼
3
1
ꢂ18
2
cm molecule in all our kinetic data analyses.
ꢂ1
known values for [C
3
3
H Cl]
0
via
1.9 ꢁ 10
jIL
d
½
C3H3ꢆ ¼
f1 ꢂ exp½ꢂseðC3H3Cl; 193 nmÞ
0
3 3 3 3
The C H + C H reaction
É
ꢁ
½C3H3 Clꢆ ꢁ dꢆ
ð3Þ
0
Reaction (1) was investigated at temperatures of 373 K, 425 K,
500 K, and 520 K and at pressures between 1 and 140 bar with
N as bath gas. The detailed conditions are given in Table 1.
1
1
Here j ¼ 0.93 is the quantum yield for C
3 3
H formation,
ꢂ18
2
ꢂ1
s
e
(C
3
H
3
Cl, 193 nm) ¼ 3.8 ꢁ 10
cm molecule the absorp-
d ¼ 10.1 cm the optical path
the laser fluence. In our evaluation of eqn. (3),
we neglected the temperature dependences of and
(C Cl, 193 nm), because both quantities are not expected
2
8
,19
tion cross-section of C H Cl,
We note that experiments below 373 K were not possible
due to the low vapour pressure of DPO.
The kinetic analysis is based on a second-order rate law
3
3
length, and I
L
j
s
e
3 3
H
d½C3H3ꢆ
2
to substantially vary in the limited temperature range of our
experiments. Whereas the quantum yield is generally not acces-
sible with our experimental setup, the absorption cross-section
ꢂ
¼ 2k1½C3H3ꢆ
ð4Þ
dt
3 3
where [C H ] is the concentration of the propargyl radicals
and k the rate coefficient of reaction (1). With the definition of
of C
3
3
H Cl at 193 nm can be estimated by measuring
1
the attenuation of the photolysis laser-beam. We obtained
ꢂ18
the absorbance A and the Beer–Lambert law
2
cm molecule at 370 K and
ꢂ1
values of (3.4 ꢀ 0.6) ꢁ 10
Iz
ꢂ18
2
cm molecule at 450 K. A comparison
ꢂ1
(
4.5 ꢀ 0.9) ꢁ 10
A ꢇ ln I ¼ hsei½C3H3ꢆd
ð5Þ
ð6Þ
with the room-temperature value given above shows that the
temperature dependence indeed is weak and not significant
eqn. (4) upon integration becomes
within our error margins. The results for s
e
(C
3
H
3
, 332.5 nm)
ꢀ
ꢁꢂ1
1
2k1t
at 5 different temperatures are shown in Fig. 4. The deviation
from the room-temperature value reported by Atkinson and
Hudgens may be due to the higher pressure in our study. A
A ¼
þ
A0 hseid
8
Here I
the photolysis laser-flash, i.e., for [C
corresponding to a given concentration [C
ln(I /I represents the initial absorbance for [C
[C . A typical absorbance–time profile is shown in Fig.
. These decay profiles were analyzed in terms of eqn. (6) by
adjusting the parameters A and k in a non-linear fitting pro-
z
denotes the measured intensity at 332.5 nm before
] ¼ 0, I is the intensity
], and A
further reason might be that the light intensity within our spec-
tral bandwidth (332.5 nm ꢀ 1.25 nm, FWHM) is not uniform
but increases towards higher wavelengths. This is caused by
the strong emission band of the mercury arc-lamp that is cen-
3
H
3
3
H
3
0
¼
z
0
)
3
H
3
] ¼
3
1
3 3 0
H ]
tered at 334.1 nm. Consequently, the measured absorption
cross-section, which is an average over the bandwidth, does
not correspond to the absorption maximum at 332.5 nm but
to a slightly higher wavelength at the slope of the band. Hence,
a somewhat lower value for the absorption cross-section is
obtained, which is operational because it contains information
on the non-uniform intensity distribution of the lamp. As all of
5
0
1
cedure. The rate coefficients obtained in this way are collected
in Table 1; the grey line in Fig. 5 represents the result of such a
fit. Additionally, the adequacy of a second-order rate law was
verified by checking the linearity of the plots 1/A vs. t; for an
example see inset of Fig. 5.
We obtained rate coefficients, which exhibit a slight positive
temperature dependence but no significant pressure depen-
dence. Hence, for a compact data representation it is useful
to average the values obtained for different pressures at
the same temperature. To judge the quality of our results, we
Table 1 Rate coefficients k
from the non-linear fits (averaged for experimental runs performed
under identical conditions, see text)
1
(ꢀsingle standard deviation) obtained
No of
Experiments
ꢂ11
3 ꢂ1 ꢂ1
cm molecule s
T/K
P/bar
k
1
/10
3
73
1
4
3
3
2
4
1
2
2
2
1
2
1
2.6 ꢀ 0.1
2.8 ꢀ 0.1
2.7 ꢀ 0.1
2.7 ꢀ 0.1
2.7 ꢀ 0.2
2.9
1
2
5
0
0
0
1
1
00
40
1
4
25
2.9 ꢀ 0.1
2.7 ꢀ 0.1
3.6 ꢀ 0.1
3.4
1
00
1
Fig. 4 Absorption cross-section of C
from ref. 8; L: this work (P ¼ 1 bar, bath gas N
standard deviation; solid line: averaged temperature-independent value
3
H
3
at 332.5 nm (base e); K:
500
20
2
); error bars: two-fold
108
1
5
4.1 ꢀ 0.1
4.2
ꢂ18 ꢂ1
2
hs
e
i ¼ 1.9 ꢁ 10
cm molecule used for the data analysis in this
1
08
work (see text).
4
644
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646

View Article Online
slightly decreasing low-pressure rate coefficient with tem-
perature for T < 500 K can surely not be ruled out, if one takes
into account the uncertainties in the thermochemical para-
meters of the model. Here additional work is clearly needed.
To improve the situation at higher temperatures, a shock tube
investigation of reaction (1) is currently underway in our
laboratory.
Summary
3 3 3 3
The kinetics of the C H + C H reaction was investigated at
temperatures between 375 K and 520 K and pressures ranging
from 1 bar to above 100 bar. Pressure-independent rate coeffi-
ꢂ11
ꢂ11
3
cm molecule s
ꢂ1 ꢂ1
cients between 2.7 ꢁ 10
and 4.1 ꢁ 10
with a weak positive temperature dependence were obtained.
For the first time propargyl radicals were generated by
photolysis of dipropargyl oxalate. The absence of interfering
halogen atoms, often present in earlier studies using halogen
containing precursors, enabled us to study the absorption spec-
trum of C H under halogen-free conditions. We could con-
Fig. 5 Absorbance–time profile at 332.5 nm (P ¼ 1 bar, T ¼ 520 K)
ꢂ11
and non-linear fit according to eqn. (6) (A
0
¼ 0.026, k ¼ 4.1 ꢁ 10
1
3
cm molecule
ꢂ1 ꢂ1
s ); inset: linear second-order plot.
3
3
firm an absorption between 295 nm and 355 nm found in an
earlier study but rule out an absorption near 240 nm found
in another work.
estimated a maximum error of ꢀ20 %, which is mainly caused
by uncertainties in the fitting procedure due to the unfavour-
able signal/noise ratio especially at longer reaction times (cf.
Fig. 5). This also limits the applicability of the linear second-
order plot for data analysis.
Acknowledgements
The results of our experiments averaged in this way are com-
pared with previously measured values from other authors in
Financial support from the Deutsche Forschungsgemeinschaft
(
8
,12,15
Table 2. In three very recent studies,
rate coefficients near
ꢂ1 ꢂ1
cm molecule s are reported in very good agree-
SFB 551, Kohlenstoff aus der Gasphase: Elementarreaktio-
ꢂ11
3
4
ꢁ 10
nen, Strukturen, Werkstoffe) is gratefully acknowledged. We
thank Prof. Y. Q. Yin for information on the synthesis of
dipropargyl oxalate.
ment with each other for a temperature of 295 K and pressures
between 3 and 133 mbar. These new results seem to supersede
the threefold higher value obtained by Morter et al. in 1994.
1
1
Our results, though determined at somewhat higher tempera-
tures, also seem to favour a room temperature value near
ꢂ11
3
cm molecule s . They are in good agreement
ꢂ1 ꢂ1
References
4
ꢁ 10
1
3
with the value obtained by Shafir et al. for a temperature
of 500 K. Because the pressure range of our study considerably
differs from the pressures applied in the experiments of Shafir
1
2
M. Frenklach, Phys. Chem. Chem. Phys., 2002, 4, 2028.
H. Richter and J. B. Howard, Prog. Energy Combust. Sci., 2000,
2
6, 565.
1
3
et al., the pressure dependence of k
appears to be weak. This is also predicted in a recent theoreti-
1
at these temperatures
3
K. H. Homann, Angew. Chem., 1998, 110, 2572; K. H. Homann,
Angew. Chem. Int. Ed. Engl., 1998, 37, 2434.
1
6
4
5
6
C. H. Wu and R. D. Kern, J. Phys. Chem., 1987, 91, 6291.
J. A. Miller and C. F. Melius, Combust. Flame, 1992, 91, 21.
J. P. Senosiain, J. H. Han, C. B. Musgrave and D. M. Golden,
Faraday Discuss., 2001, 119, 173.
cal investigation by Miller and Klippenstein. Whether the
slightly positive temperature dependence of k at higher pres-
sures obtained in our work and the slightly negative tempera-
ture dependence obtained at lower pressures in the work of
Shafir et al. are due to the different pressure ranges remains
to be clarified. The master-equation modeling by Miller and
Klippenstein predicts a weak positive temperature depen-
dence below 500 K for virtually all pressures. However, a
1
7
8
I. R. Slagle and D. Gutman, Proc. Combust. Inst., 1988, 21, 875.
D. B. Atkinson and J. W. Hudgens, J. Phys. Chem. A, 1999, 103,
4242.
1
3
1
6
9
0
1
D. K. Hahn, S. J. Klippenstein and J. A. Miller, Faraday Discuss.,
2
001, 119, 79.
U. Alkemade and K. H. Homann, Z. Phys. Chem. (Neue Folge),
989, 161, 19.
1
1
1
C. L. Morter, S. K. Farhat, J. D. Adamson, G. P. Glass and R. F.
Curl, J. Phys. Chem., 1994, 98, 7029.
Table 2 Comparison of our results (ꢀestimated maximum error)
with previous experimental measurements of k
1
1
1
2
3
A. Fahr and A. Nayak, Int. J. Chem. Kinet., 2000, 32, 118.
E. V. Shafir, I. R. Slagle and V. D. Knyazev, J. Phys. Chem. A,
submitted for publication.
ꢂ11
cm
3
k
1
/10
molecule
ꢂ1 ꢂ1
s
T/K
P
Bath gas
Ref.
14 S. Scherer, Th. Just and P. Frank, Proc. Combust. Inst., 2000, 28,
511.
1
5
.7
623–673
295
295
(3–6) mbar
ca. 20 mbar
(3–133) mbar Ar, He, N
He, N
He + CO
2
Ref. 10
Ref. 11
15 J. D. DeSain and C. A. Taatjes, J. Phys. Chem. A, 2003, 107, 4843.
16 J. A. Miller and S. J. Klippenstein, J. Phys. Chem. A, 2001, 105,
1
2 ꢀ 2
2
7
254.
4
4
3
.3 ꢀ 0.6
.0 ꢀ 0.4
.9 ꢀ 0.6
2
Ref. 8
Ref. 12
17
18
19
D. A. Ramsay and P. Thistlethwaite, Can. J. Phys., 1966, 44, 1381.
M. E. Jacox and D. E. Milligan, Chem. Phys., 1974, 4, 45.
A. Fahr, P. Hassanzadeh, B. Laszlo and R. E. Huie, Chem. Phys.,
295
296
67 mbar
8 mbar
He
He
Ref. 15
Ref. 13
3
2
1
.30 ꢀ 0.35
.74 ꢀ 0.43
.20 ꢀ 0.14
500
700
ꢃ(2–4) mbar He
ꢃ(3–6) mbar He
ꢃ(4–8) mbar He
1
997, 215, 59.
H.-J. Deyerl, I. Fischer and P. Chen, J. Chem. Phys., 1999, 111,
441.
Ref. 13
Ref. 13
Ref. 14
2
0
1000
3
1
2
2
3
4
.2 ꢀ ꢃ 0.3
.7 ꢀ 0.6
.8 ꢀ 0.6
.5 ꢀ 0.8
.1 ꢀ 0.8
1000–1250 ꢃ(1–2) bar
Ar
21 D. B. Atkinson and J. W. Hudgens, J. Phys. Chem. A, 1999, 103,
7978.
22 D. G. L. James and S. M. Kambanis, Trans. Faraday Soc., 1969,
373
425
500
520
(1–140) bar
(1–100) bar
(1–108) bar
(1–108) bar
N
2
N
2
N
2
N
2
This work
This work
This work
This work
6
3 D. M. Golden, N. A. Gac and S. W. Benson, J. Am. Chem. Soc.,
5, 1350.
2
1
969, 91, 2136.
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646
4645

View Article Online
2
4
T. Sakai, D. Nohara and T. Kunugi, in Industrial and Laboratory
Pyrolyses, ed. L. F. Albright and B. L. Crynes, ACS Symposium
Series 32, American Chemical Society, Washington DC, 1976,
p. 152.
V. A. Korolev, A. K. Mal’tsev and O. M. Nefedov, Bull.
Acad. Sci. USSR Div. Chem. Sci. (Engl. Transl.), 1989, 38,
957.
S. N. Danilov, L. N. Yastrebov, A. L. Galka and A. S. Zanina,
J. Org. Chem. USSR (Engl. Transl.), 1979, 15, 1025.
27 M. Neveux, C. Bruneau, S. L e´ colier and P. H. Dixneuf, Tetra-
hedron, 1993, 49, 2629.
28 Y. Q. Yin, private communication.
29 K. A. Holbrook, in The Chemistry of Acid Derivatives, Suppl. B2,
ed. S. Patai, Wiley, Chichester, 1992, vol. 2, p. 703.
30 CRC Handbook of Chemistry and Physics, CRC Press, Boca
Raton, FL, 80th edn, 1999.
2
5
6
2
31 J. G. Calvert and J. N. Pitts, Jr., Photochemistry, Wiley, New
York, 1966.
4
646
Phys. Chem. Chem. Phys., 2003, 5, 4641–4646