The thermal decomposition of 5-membered rings: A laser pyrolysis study

Article abstract of DOI:10.1039/b315093c

The mechanisms of pyrolysis of cyclopentadiene, furan, pyrrole and thiophene have been investigated using a combination of IR laser powered homogeneous pyrolysis, chemical and physical trapping of radical intermediates, and use of precursors specifically designed to generate selected radical intermediates. The results confirm the central role played by free radicals in the cases of cyclopentadiene and thiophene, and the dominant step of 1,2-H shifts in the cases of furan and pyrrole. The experimental results may be interpreted according to the high level ab initio calculations recently reported in the literature.

Full text of DOI:10.1039/b315093c

View Article Online / Journal Homepage / Table of Contents for this issue
P A P E R
The thermal decomposition of 5-membered rings: a laser
pyrolysis study
Nathan R. Hore and Douglas K. Russell*
Department of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand.
E-mail: d.russell@auckland.ac.nz; Fax: þ64 9 373 7422
Received (in Montpellier, France) 20th November 2003, Accepted 19th January 2004
First published as an Advance Article on the web 1st April 2004
The mechanisms of pyrolysis of cyclopentadiene, furan, pyrrole and thiophene have been investigated using
a combination of IR laser powered homogeneous pyrolysis, chemical and physical trapping of radical
intermediates, and use of precursors specifically designed to generate selected radical intermediates. The results
confirm the central role played by free radicals in the cases of cyclopentadiene and thiophene, and the dominant
step of 1,2-H shifts in the cases of furan and pyrrole. The experimental results may be interpreted according to
the high level ab initio calculations recently reported in the literature.
5
-bromofuroic acid according to the method of Clennan and
Introduction
3
4
Mehrsheikh-Mohammabi, and the purity of the final product
checked using GC/MS. Since cyclopentadiene dimerises at
room temperature, the vapour of the monomer was generated
by heating the pyrolysis cell (see below) at a temperature insuf-
ficient to induce further pyrolysis; the exclusive presence of
the monomer was verified by FTIR spectroscopy. All samples
were handled on a rigorously pre-conditioned vacuum line
fitted with greaseless Young’s taps.
The mechanism of pyrolysis of 5-membered cyclic compounds
is of considerable interest in our understanding of combustion.
In particular, systems such as furan, pyrrole, cyclopentadiene
and especially thiophene have all been implicated as significant
1
species in the pyrolysis of fossil fuels. In addition, furan has
also been identified as a significant component of tobacco
2
smoke, while the pyrolysis of thiophene has applications in
3
the synthesis of conducting polymers. While there have been
several experimental studies of the kinetics of decomposition
with some
1
,4–6
1,7–11
1,12–17
of furan,
pyrrole
and cyclopentadiene
IR laser powered homogeneous pyrolysis
mechanistic interpretation, there has been rather little experi-
mental work specifically directed towards full elucidation of
mechanistic detail, and very little at all has been published on
thiophene apart from a very recent product-oriented investiga-
All experimental pyrolysis studies were carried out using the
method of infrared laser powered homogeneous pyrolysis, IR
LPHP. This technique has been reviewed in detail else-
3
5–38
1
8
where,
and therefore only a brief description is provided
tion. On the other hand, a number of recent high-level
theoretical investigations have thrown new light on possible
reaction routes.
here. Pyrolysis is carried out in a cylindrical Pyrex cell (length
10 cm, diameter 3.8 cm) fitted with ZnSe windows. The cell is
filled with a few torr (1 torr ¼ 133.3 Pa) of the vapour under
1
9–24
We have shown that the technique of infrared laser powered
homogeneous pyrolysis (IR LPHP), allied with standard
mechanistic tools such as matrix isolation spectroscopies, reac-
tion in the presence of trapping agents, and ab initio or semi-
empirical calculation of activation energies, can provide
unique insights into the mechanisms of pyrolysis in systems
6
study and approximately 7 torr of SF . The contents of the cell
are then exposed to the output of a free running continuous
CO laser at a wavelength of 10.6 mm and a power level
depending on the temperature required. As shown else-
2
3
5–38
where,
is then converted to heat via rapid inter- and intramolecular
relaxation. The low thermal conductivity of SF ensures
6
the SF strongly absorbs the laser radiation, which
2
5
as varied as azines, organometallic chemical vapour deposi-
tion (CVD) precursors,
2
6–31
32
glycine and chlorofluorocar-
In particular, the use of matrix isolation EPR
6
3
3
that this process generates a strongly inhomogeneous tempera-
ture profile, in which the centre of the cell may reach tempera-
tures as high as 1500 K while the cell walls remain at room
temperature.
bons.
spectroscopy and of radical traps such as H2 or D2 allows
the disentanglement of molecular and radical pathways. In
2
5
the case of azines, we have also shown that the use of precur-
sors specifically designed to proceed through relevant radical
pathways, principally bromine-substituted analogues, allows
even more detailed mechanistic information. In the present
work, we apply these techniques to a study of the 5-membered
ring systems cyclopentadiene, pyrrole, furan and thiophene.
The IR LPHP technique has a number of well-documented
advantages. The first of these is that pyrolysis is initiated
directly in the gas phase, eliminating the complications of com-
peting surface reactions frequently encountered in more con-
ventional techniques. The second is that less volatile primary
products are rapidly ejected into cool regions of the cell, where
they are not subject to further reaction. In favourable cases,
these products may be accumulated for further reaction or
analysis. Furthermore, the very small quantities of material
used render the technique ideal for mechanistic studies invol-
ving isotope labelling, and the judicious use of additional
reagents in co-pyrolysis studies is straightforward. Finally,
we have shown that the technique may be incorporated into
a flow system coupled with cryogenic techniques, permitting
Experimental methods and calculations
Compounds
All compounds used in this study, with the exception of 2-
bromofuran, were ‘AR’ grade (Aldrich or Lancaster); these
and SF (BOC) were purified prior to use by repeated
6
freeze-pump-thaw cycles. 2-Bromofuran was synthesised from
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
6
06
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3

View Article Online
the identification of short-lived species through matrix
isolation IR or EPR spectroscopy.
steady state concentration and hence on its rates of both
formation and subsequent decomposition.
On the other hand, the temperature of the pyrolysis is
neither well-defined nor easily determined, so that comparison
with more conventional methods is not straightforward.
Furthermore, the effective temperature is in part determined
by the conductivity of the cell contents, and hence may vary
over the course of reaction. Despite these factors, an initial
effective temperature may be estimated from the rate of disap-
pearance of a compound with known (and comparable) kinetic
parameters. In the present experiments, the effective first-order
rate constants for thermal decomposition of the target com-
Calculations
When available, relevant results of high level ab initio calcula-
tions were obtained from the literature. When not available,
indicative calculations of structures, energies and transition
states were carried out at the ab initio (Hartree–Fock 3-21G
or 6-31G*) or B3LYP density functional level using the PC
0
39
Spartan 02 molecular modelling package. All structures
were located by minimising the norm of the analytical energy
6
,11
ꢀ9
pounds themselves reported by Mackie et al.
and Burcat
gradient to less than 5 ꢁ 10 N. Transition states were located
1
7
and Dvinyaninov were used to provide estimates of the pyr-
olysis temperatures for furan, pyrrole and cyclopentadiene,
respectively. It must be emphasised that IR LPHP is a techni-
que with relatively long contact times (several minutes) and
product conversions at the same temperature are therefore
much higher than those achieved with methods such as shock
tube or flash vacuum techniques. Consequently, it is possible
to probe considerably lower effective temperatures, with the
result that the technique discriminates more strongly in favour
of lower activation energy routes when multiple parallel
pathways exist.
by ensuring that the Hessian matrix (matrix of energy second
derivatives) possessed only one negative frequency, corre-
sponding to the reaction coordinate. Reaction coordinates
were verified using the visualisation facilities of the package.
Although the absolute accuracy of these methods is less than
more sophisticated methods reported in the literature, bench-
marking against high-level calculations on similar systems
has shown that they do provide sufficiently reliable discrimina-
tion among possible reaction pathways.
Results and discussion
Furan and bromofurans
Analysis
Of the published studies of the kinetics of the thermal decom-
position of furan, two were carried out under the single pulse
shock tube conditions that are of most relevance to the present
work. Lifshitz et al. studied the decomposition of furan at tem-
peratures between 1050 and 1460 K, and concluded that two
parallel major routes operate, one leading to the formation
of CO and propyne and/or allene, and the second to acetylene
and ketene, although ketene was not directly detected in that
All stable products were identified by a combination of FTIR
and GC/MS. All FTIR spectra were recorded directly (either
as vapour or as liquid or solid deposits on the cell windows)
using the 10 cm long pyrolysis cell described above. Although
the ZnSe windows have a relatively high cut-off at the low
frequency end (approximately 500 cm ), for our LPHP work
it possesses many advantages over materials such as NaCl
or KCl.
ꢀ
1
5
6
work. The work of Organ and Mackie covered the tempera-
ture range 1100 to 1700 K; in addition to propyne, allene was
also identified as a major product. The latter workers con-
cluded that both routes originate in a common first step, that
of homolytic C–O bond cleavage resulting in a biradical
species. Very recently, Sendt et al. have carried out high level
ab initio calculations on the furan system and concluded that
neither C–O not C–H fission is likely to be a major initiation
process; instead, 1,2-H transfers resulting in cyclic carbenes
provide the first step in the lowest energy pathways to the
All GC/MS work was carried out using a Hewlett Packard
890 Series II gas chromatograph coupled with a Hewlett
6
Packard 5873 mass selective detector. The GC employed a
HPI phenyl methyl siloxane gum capillary column of 0.33
mm film thickness and 25 m by 0.2 mm internal diameter. Pyr-
olysis samples for GC/MS analysis were taken directly from
the pyrolysis cell by means of a rubber septum and a sealable
gas-tight syringe. This method possesses several practical
advantages over the solvent-based methods used earlier and
also is more reliably quantitative.
1
9
observed products.
EPR spectroscopy was carried out using a Varian E-4 EPR
spectrometer. Samples for EPR spectroscopy were collected
In the LPHP work reported here, laser powers were set to
produce lower temperatures between 950 and 1000 K, as esti-
mated from the rate of disappearance of furan and the effective
2
8–30
using the methods described in detail elsewhere.
Briefly,
6
the laser pyrolysis cell is connected to a vacuum line and liquid
nitrogen cooled cold finger assembly. A bulb containing the
kinetic parameters of Organ and Mackie. The major products
were those reported by earlier workers, with the CO þ C H
3
4
required mixture of SF
to the cell and the pyrolysis gas is slowly flowed through the
cell while being exposed to the CO laser radiation. The result-
6
and starting material is also attached
route dominating over ketene and acetylene; there was a signi-
ficant production of benzene in addition (approximately 15%).
2
However, in LPHP the major C
FTIR and GC/MS is clearly allene, with very little detectable
propyne. Co-pyrolysis with H (furan:H
3
product identified by both
ing gas containing pyrolysis products, including any free radi-
cal intermediates, is then trapped onto the cold finger. This
results in a matrix containing SF , unreacted starting material
2
2
¼ 1:1) resulted in a
modest decrease in allene, propyne and benzene relative to
CO, acetylene and ketene; no deuterium incorporation into
the initial furan was detected on co-pyrolysis with D . No
6
and intermediates and final products. The entire cold finger
assembly can be demounted and installed into the EPR
spectrometer for examination.
For the present study, the technique was modified to incor-
porate a pressure reduction glass diaphragm between the
pyrolysis cell and the cold finger assembly. This permitted
pyrolysis at higher pressures than previously, in turn allowing
the attainment of significantly higher temperatures and
exploration of a wider range of intermediates. Loss of radicals
through the diaphragm did not appear to be marked and, in
2
EPR signals were observed under any conditions and no furan
oligomers were detected.
As shown in our studies of azines, the pyrolysis of bromi-
nated analogues can provide additional insight into the signi-
ficance of radical routes that may operate for the parent
compound. Accordingly, the pyrolysis of both 2-bromofuran
and 3-bromofuran was studied at temperatures near 850 K
in order to identify the part played by furyl radicals. These
two compounds yielded almost identical results, but dramati-
cally different from the parent furan. CO and benzene
( > 70%) now appear as substantially the major products, with
2 3
any event, was reduced by coating with B O to minimise sur-
face-catalysed recombination. The observation (or otherwise)
of signals arising from a given free radical depends on its
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3
607

View Article Online
appreciable quantities of bromobenzene ( > 20%) and much
smaller amounts of the C H isomers hexa-1,3-dien-5-yne
cyclic carbene is formyl allene, which subsequently undergoes
a 1,4-hydrogen shift and simultaneous C–C s-bond cleavage
to yield CO and propyne. However, it is well-known that for-
myl allene undergoes ready conversion to the more stable vinyl
6
6
and hexa-1,5-dien-3-yne, acetylene, 1,3-butadiyne, 1-buten-3-
yne, and allene or propyne. Again, GC/MS revealed no trace
of furan oligomers. Co-pyrolysis with H reduced the yield of
4
1,42
2
ketene;
ketenes very readily decarbonylate and the resul-
C
6
H
6
isomers at the expense of allene and propyne. Relevant
tant vinyl methylene is likely to suffer the usual 1,2-hydrogen
shift to yield the observed allene. The calculated activation
energies of potential reaction pathways are summarised in
Table 1 and the dominant reaction pathways for both furan
itself and the bromofurans are illustrated in Scheme 1.
regions of the FTIR spectrum of the pyrolysis products of
furan and 3-bromofuran are illustrated in Fig. 1.
The very marked change in product distribution on using
bromofurans as starting materials and the absence of EPR sig-
nals from furan, together with the lack of sensitivity towards
2
co-pyrolysis with H , combine to rule out a significant radical
route initiated by C–H bond homolysis in furan. However, the
substantial production of benzene from bromofurans ( > 70%)
confirms the furyl radical as the principal source of benzene, as
assumed in earlier work. This conclusion is consistent with the
Pyrrole
Two single pulse shock tube studies of pyrrole pyrolysis, by
1
0
11
Lifshitz et al. and Mackie et al., have covered the tempera-
ture ranges 1050–1450 K and 1200–1700 K, respectively. The
latter workers identified a reversible 1,2-hydrogen shift to form
the unstable pyrrolenine as the first step, followed by C2–N
homolysis. The resultant biradical then breaks down in a man-
ner analogous to that for furan, yielding HCN and C H
(allene or propyne) or acetylene and acetonitrile. Alternatively,
it may rearrange to form the CN-terminal pyrrole isomers allyl
cyanide or crotonitrile in a route not available for furan. As in
the case of furan, there have been several recent high-level cal-
culations in an attempt to discriminate among possible initia-
ꢀ
1
high energy of the C–H bond (over 500 kJ mol according to
the CASPT2 calculations of Sendt et al. ) compared with less
1
9
ꢀ
1
19
than 300 kJ mol for 1,2-H shifts. Similar remarks apply to
the biradical route arising from C=O homolysis (calculated at
slightly below or above 400 kJ mol , depending on spin
ꢀ1
3
4
1
9
state ). However, it is clear from the bromofuran results that
both 2-furyl and 3-furyl radicals suffer ready loss of CO to
yield C H , which is a well-known precursor to benzene.
3 3
4
0
In view of the low energy barrier to H migration in the parent
compound, it is likely that the two furyl radicals also intercon-
vert very readily, resulting in identical end products from
2
0–23
tion routes.
the relatively high N–H bond energies in pyrrole (397 kJ mol
at the CASPT2 level ) and our own calculations show that
direct C–H fission at either the 2 or 3 positions is even less
All these theoretical studies have confirmed
ꢀ1
either precursor. In the presence of H
are evidently trapped as the isomers of C H rather than
2 3 3
, the C H radicals
2
0
3
4
combining to yield benzene.
Not predicted from earlier work, or by the calculations of
ꢀ
1
favourable at 473 and 444 kJ mol , respectively. Direct ring
C–N or C–C scissions are also unfavourable, at 393 and 557
1
Sendt et al., was the almost exclusive production of the allene
9
ꢀ
1
23
isomer of C H . These calculations suggest that the first stable
3
kJ mol , respectively. As in the case of furan, the most likely
routes involve 1,2-hydrogen shifts, yielding initially cyclic
4
intermediate of the ring-opening of the 3,2-hydrogen shifted
Fig. 1 Partial FTIR spectra of the products of IR LPHP of furan (above) and 3-bromofuran (below). Dotted line ¼ before pyrolysis, full line ¼
after pyrolysis. Peaks are labelled as: 1 ¼ ketene, 2 ¼ CO, 3 ¼ allene, 4 ¼ acetylene, 5 ¼ benzene; unassigned peaks are due to starting material
6
or SF .
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
6
08
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3

View Article Online
a
Table 1 Structures and calculated energies of possible initial intermediates in the thermal decomposition of 5-membered rings
Number
I
II
III
VI
V
VI
VII
Structure
b
b
b
g
A ¼ CH
A ¼ NH
A ¼ O
2
0
0
0
0
–
–
346/346/379
Polyaromatics
310/309/428
306/275/393
167/0
–
190/58/317
C
258/203
2
H
2
þ allene
C
302/265/332
2
H
2
þ propyne
f
f
f
f
c
d
d
d
494/494
495/495
492/492
495/495
397/397
HCN þ propyne
Cyanopropenes
–
e
e
–
–
270/229/346
þ H CCO
305/261/590
290/245/275
CO þ C
C
2
H
2
2
3 4
H
f
f
g
g
A ¼ S
488/488/488
Bisthiophene
476/476/476
Bisthiophene
326/279/611
–
a
ꢀ1
0
All energies in kJ mol . A ¼ CH
2
, NH, O or S; A ¼ CH or N. For each numerical entry, the first number is the energy of the transition state
for formation, the second the energy of the species, and the third (where given) the energy of the transition state for subsequent decomposition,
b
all relative to the parent compound; the major products of the specified decomposition route are also listed for each species. Ref 24. Ref 20.
c
d
e
Ref 23. Ref 19. Ref 46. This work (B3LYP).
f
g
carbenes (or, in the case of pyrrole, pyrrolenine), followed by
ring opening and rearrangement.
2
Co-pyrolysis with H suppressed the formation of benzene
and on co-pyrolysis with D , some deuteration of the open-
2
IR LPHP of pyrrole at temperatures between 1000 and 1050 K
confirmed cis- and trans-crotonitrile (E- and Z-1-cyanopropene,
chain pyrrole isomers was detected, but none of the residual
pyrrole. Unfortunately, C-substituted bromopyrroles are
well-known to be unstable towards auto-polymerisation,
precluding any study of these compounds; other potential
sources of pyrrolyl radicals do not lend themselves readily to
comparison with the IR LPHP observations.
4
3–45
2
1
0% and 9%, respectively) and allyl cyanide (3-cyanopropene,
1%) as the major low temperature pyrolysis products. In addi-
tion, significant amounts of acrylonitrile (cyanoethene, 16%),
acetylene, HCN and allene or propyne (38%) were also pro-
duced, together with benzene and other C
6
H
6
isomers in smaller
The IR LPHP results are entirely consistent with earlier
investigations and the results of high-level calculations and
so little additional insight into the decomposition mechanism
for this compound is afforded. The calculated activation ener-
gies of potential reaction pathways are summarised in Table 1
and the dominant reaction pathways for pyrrole are illustrated
in Scheme 2.
quantities (5% total). A typical GC-MS trace is shown in Fig. 2.
Only very weak EPR signals were detectable, indicating a very
minor radical route; no pyrrole oligomers were detected.
Cyclopentadiene
There have been several experimental studies of the thermal
decomposition of cyclopentadiene.
1
2–17
The major products
at high temperatures are polycyclic aromatics such as naphtha-
lene, etc.; at lower temperatures, products include relatively
minor quantities of benzene, acetylene and allene and/or pro-
pene. It has generally been accepted that the principal initia-
tion step in the pyrolysis of cyclopentadiene is H loss to
form the stable c-C H radical; however, to our knowledge,
5
5
there has been no prior report of the direct confirmation of
this process. Following their similar studies of the isoelectronic
furan and pyrrole, Bacskay and Mackie have carried out
ab initio calculations at the CASPT2 level and have shown
that 1,2-H shifts to produce cyclic carbenes may also be
very important in the production of C H and C H in the
2
2
3
4
2
4
cyclopentadiene case.
IR LPHP of cyclopentadiene led to an extensive array of
products; the major products detectable by GC/MS included
benzene (51%) and the other aromatic hydrocarbons identified
by previous workers, predominantly toluene (20%), naphtha-
lene (12%), styrene (7%) and indene (4%). Higher molecular
weight products may also be formed, but are unlikely to have
sufficient vapour pressure to be recovered for GC/MS analy-
sis. Low molecular weight products in relatively low yield
included methane, acetylene, propyne and/or allene, 1,3-buta-
diene and but-1-ene-3-yne. Co-pyrolysis with excess H
a reduction in the aromatic products at the expense of
smaller products, while co-pyrolysis with D led to extensive
deuteration of residual cyclopentadiene. Significantly, the
benzene product was not deuterated at all, while the allene was
2
led to
2
Scheme 1 Major pathways in the thermal decomposition of furan
(above) and the bromofurans (below); activation energies are given
in Table 1.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3
609

View Article Online
Fig. 2 Gas chromatogram of products of IR LPHP of pyrrole.
very heavily deuterated (the ratio of mass 41 to mass 40 was
almost 3:1 in the presence of D , while this ratio was 1:20 in
the absence of D ). Very strong EPR signals, unambiguously
attributable to cyclopentadienyl radicals, were observed, as
shown in Fig. 3. Brominated compounds were considered unli-
kely to reveal additional information in this case and no study
of these was undertaken.
back reaction, almost certainly involving H atoms, since the
reaction of C H with H is endothermic to the extent of some
2
5
1
5
2
ꢀ
2
100 kJ mol . It is therefore reasonable to conclude that these
radicals do not break down, but recombine in a number of
ways to yield the observed polyaromatic hydrocarbons (and
eventually soot, at higher temperatures still). This conclusion
2
4
is in agreement with the calculations of Bacskay and Mackie,
as well as those of other theoretical investigations; thus the
4
6
Clearly, C
tities, but are relatively stable to the further breakdown to
5
H
5
radicals are formed in very substantial quan-
ꢀ1
C–H bond strength is calculated to be 343 kJ mol at the
CASPT2 level, while the subsequent steps of 1,2-H shift and
1
2–17
C
2
H
2
and C
they recombine with loss of H
as shown in our work on cyclopentadienyl containing transi-
3
H
3
assumed in some earlier work.
Instead,
ꢀ
1
2
to form largely naphthalene,
C–C bond fission require a further 403 kJ mol . Bacskay
and Mackie have further asserted that the small molecule route
2
8
tion metal compounds and elsewhere; while well-attested,
the detailed mechanism of this step is unknown and merits
further theoretical investigation. The extensive D-incorpora-
5 5
is therefore unlikely to involve C H radicals and instead fol-
lows 1,2-H shifts to yield cyclic carbenes, followed by ring
opening and C–C fission; the activation energies of the two
possible routes were found to be 393 and 427 kJ mol . There
ꢀ1
tion observed on co-pyrolysis with D
2
implies considerable
Scheme 2 Major pathways in the thermal decomposition of pyrrole; activation energies are given in Table 1.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
610
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3

View Article Online
organic systems and is a major sulfur-containing component of
fossil fuels.
IR LPHP of thiophene at laser powers estimated to produce
temperatures of 1150 to 1200 K resulted in solid deposits of
polythiophene, with acetylene as the major gas phase product;
trace quantities of benzene and methane were observed after
prolonged laser exposure. In addition, considerable quantities
of the three isomeric bisthiophenes were observed, constituting
over 60% of the observable products, as well as smaller
quantities of benzothiophene and phenylthiophene. At higher
laser powers, CS
were also observed. Co-pyrolysis of thiophene with D
thiophene:D
This shows very clear evidence of extensive deuteration of
the parent at both the 2 and 3 positions, together with peaks
attributable to more highly deuterated products; this is con-
firmed by GC-MS. Very weak EPR signals were observed,
but could not be analysed satisfactorily. LPHP of 2-Br and
2
and increased amounts of polyaromatics
2
(
2
¼ 1:1) produced the FTIR spectrum of Fig. 4.
Fig. 3 EPR spectrum of free radicals produced in the IR LPHP of
cyclopentadiene.
3
-Br thiophene (at much lower laser powers) resulted in the
is some support for this assertion in the analogous saturated
cyclopentylidene carbene. When formed through low energy
routes [such as abstraction of the O atom from cyclopentanone
production of methane and thiophene itself, as well as deposits
identified as polythiophene. All three isomeric bisthiophenes
were again observed, with product ratios dependent to some
extent on the nature of the starting material. Co-pyrolysis of
the bromo compounds with D also resulted in extensively
2
deuterated thiophene. The relative yields of the isomeric
bisthiophenes are illustrated in Fig. 5.
The higher temperature required to decompose thiophene is
consistent with the absence of a low activation energy initia-
tion step. Calculations place C–H bond strengths at 490 kJ
mol
with its lighter analogues, 1,2-H transfer is the most probable
step (activation energy 300 kJ mol ), but it is clear from the
results above that both thienyl radicals and any ring carbenes
must also be sufficiently stable to internal reaction to allow for
extensive exchange reactions before bond fission. It is entirely
reasonable to assume that recombination of thienyl radicals
leads to the major observed products, the isomeric bisthio-
phenes, although radical reaction with the parent cannot
be ruled out. The calculated activation energies of potential
reaction pathways are summarised in Table 1 and the domi-
nant reaction pathways for both thiophene itself and the
bromothiophenes are illustrated in Scheme 4.
4
7
by W(CO) in IR LPHP ] only the 1,2-H migration product
n
cyclopentene is observed, whereas the high energy carbene
(
formed by O-abstraction by C atoms) also leads to appreci-
4
8
able quantities of allene and ethylene. The calculated activa-
tion energies of potential reaction pathways are summarised in
Table 1 and the dominant reaction pathways for cyclopenta-
diene are illustrated in Scheme 3.
ꢀ
1 49
ꢀ1
and the ring C–S bond at 560 kJ mol . By analogy
ꢀ
1
Thiophene and bromothiophenes
There have been very few detailed studies of the thermal
decomposition of thiophene, apart from a brief report on the
production of bisthiophenes in a study of the relative thermal
1
stability of a series of heterocycles by Bruinsma et al. A recent
analysis of breakdown products has confirmed bisthiophenes
as the major product at temperatures below 1100 K, with sig-
2
nificant production of CS and sulfur-containing polyaro-
1
8
matics as the temperature is raised. The reason for this is
the thermal stability of the compound; it is frequently encoun-
tered as an end product of the thermolysis of sulfur-containing
Conclusions
Decomposition of the compounds described here may concei-
vably be initiated by one or more of a number of steps, which
are discussed in turn below.
(
i) Direct scission of a ring C–C or heteroatom–C bond will
result in biradicals or linear carbenes. However, all calcula-
tions have shown that this route demands very high energy;
direct C–O bond scission to produce the lowest triplet state
ꢀ1
of a linear carbene from furan requires 322 kJ mol (at the
CASPT2 level ) and other molecules require even higher ener-
1
9
ꢀ1
gies (for thiophene, for example, 560 kJ mol at the Hartree–
Fock level). It should be noted that further energy barriers to
decomposition of these initial intermediates are also substan-
tial. We therefore conclude that direct ring-bond scission is
unlikely to be the lowest energy initiation step in any system,
and it will not be considered further.
(
ii) Scission of a ring C–H or heteroatom–H bond will pro-
duce cyclic free radicals. C–H bond strengths for a range of
heterocyclic compounds and the N–H strength in pyrrole have
recently been calculated at the B3LYP/6-31G level. C–H
4
9
ꢀ
1
scission generally requires on the order of 500 kJ mol , the
ꢀ1
notable exception being cyclopentadiene at 346 kJ mol . This
step is likely to be important only when lower energy routes
are unavailable, as in cyclopentadiene and thiophene. Since
C–Br scission in the brominated compounds is undoubtedly
Scheme 3 Major pathways in the thermal decomposition of cyclo-
pentadiene; activation energies are given in Table 1.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3
611

View Article Online
Fig. 4 Partial FTIR spectrum of the products of IR LPHP of a 1:1 mixture of thiophene and D
pyrolysis. Labelled peaks are assigned to partially deuterated thiophene as indicated; unassigned peaks are due to starting material or SF
2
. Dotted line ¼ before pyrolysis, full line ¼ after
6
.
the lowest energy step, the products of radical pathways in
these compounds provide a clear indication of the ultimate fate
of such radicals.
observation of identical products using bromothiophenes as
radical sources. On the other hand, the lower ring bond disso-
ciation energy in cyclopentadienyl also leads to the observation
of small molecule products, in contrast with thiophene. The
relative unimportance of radical routes in furan and pyrrole
is also confirmed by our observations; this is confirmed by
the observation of very different products using bromofurans
as a source of furyl radicals. In these latter systems, the initial
steps are 1,2-hydrogen shifts, followed by internal scission
and/or rearrangement of the cyclic carbenes. These assertions
are confirmed by the calculations reported in Table 1, which in
each case indicate that the lowest energy predicted pathway
leads to the observed products.
(
iii) Indirect ring-bond scission, preceded by 1,2-H shifts
from C to C or C to heteroatom (or vice versa), results in
carbenes, followed by ring opening and rearrangement. This
appears to be the lowest energy route in furan and pyrrole,
and may also contribute to small molecule production in
cyclopentadiene—it appears to be insignificant in the study of
thiophene in the present work, however.
The calculated relative energies of the relevant steps for
pathways (ii) and (iii) are given in Table 1, which will serve
as a basis for drawing general conclusions.
The pivotal role of cyclopentadienyl radicals is graphically
confirmed by their direct observation by EPR and supported
by isotope exchange on co-pyrolysis of cyclopentadiene in
The role played by the cyclic carbenes formed by 1,2-H
shifts in cyclopentadiene, furan and pyrrole, strongly indicated
by relative activation energy calculations, will be explored
in future work by direct generation using the method of
the presence of D . Similarly, the observation of bisthiophenes
2
as the major product of thiophene supports the role of
C-centred radicals in this system, a view supported by the
O-abstraction from suitable carbonyl precursors by W(CO)
n
species recently discovered in our laboratory.
4
8
Fig. 5 Partial gas chromatogram of the products of IR LPHP of thiophene (top trace), 2-bromothiophene (middle trace) and 3-bromothiophene
bottom trace), indicating relative yields of 2,2-bisthiophene, 2,3-bisthiophene and 3,3-bisthiophene. The peak marked X arises from one isomer of
phenylthiophene, the second isomer being obscured.
(
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
612
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3

View Article Online
Scheme 4 Major pathways in the thermal decomposition of thiophene and the bromothiophenes; activation energies are given in Table 1.
2
1
G. B. Bacskay, M. Martoprawiro and J. C. Mackie, Chem Phys.
Lett., 1999, 300, 321.
Acknowledgements
2
2
2
3
L. Zhai, X. Zhou and R. Liu, J. Phys. Chem. A, 1999, 103, 3917.
M. Martoprawiro, G. B. Bacskay and J. C. Mackie, J. Phys.
Chem. A, 1999, 103, 3923.
G. B. Bacskay and J. C. Mackie, Phys. Chem. Chem. Phys., 2001,
3
The authors gratefully acknowledge financial support from the
Department of Chemistry and the Research Committee of
the University of Auckland, and useful correspondence with
Professor J. C. Mackie of the University of Sydney.
2
2
4
5
, 2467.
N. R. Hore and D. K. Russell, J. Chem. Soc., Perkin Trans. 2,
998, 269.
26 D. K. Russell, Coord. Chem. Rev., 1992, 112, 131.
1
References
2
7
A. S. Grady, R. D. Markwell and D. K. Russell, J. Chem. Soc.,
Chem. Commun., 1991, 929.
1
2
3
4
5
6
7
D. S. C. Bruinsma, P. J. J. Tramp, H. J. J. de Sauvage Nolting and
J. A. Moulijn, Fuel, 1988, 67, 334.
G. Holzner, J. Oro and W. Bertsch, J. Chromatogr., 1976,
28 D. K. Russell, I. M. T. Davidson, A. M. Ellis, G. P. Mills, M.
Pennington, I. M. Povey, J. B. Raynor, S. Saydam and A. D.
Workman, Organometallics, 1995, 14, 3717.
29 D. K. Russell, G. P. Mills, J. B. Raynor and A. D. Workman,
Chem. Vap. Deposition, 1998, 4, 61.
30 D. K. Russell, I. M. T. Davidson, A. M. Ellis, G. P. Mills,
M. Pennington, I. M. Povey, J. B. Raynor and S. Saydam, Chem.
Vap. Deposition, 1998, 4, 103.
31 R. A. Berrigan, J. B. Metson and D. K. Russell, Chem. Vap.
Deposition, 1998, 4, 23.
32 J. E. Everett, N. D. Renner and D. K. Russell, Chem. Commun.,
1998, 341.
33 G. R. Allen, N. D. Renner and D. K. Russell, Chem. Commun.,
1998, 703.
34 E. L. Clennan and M. E. Mehrsheikh-Mohammabi, J. Am. Chem.
Soc., 1984, 106, 7112.
35 W. M. Shaub and S. H. Bauer, Int. J. Chem. Kinet., 1975, 7, 509.
36 J. Pola, Collect. Czech. Chem. Commun., 1981, 46, 2856.
37 D. K. Russell, Chem. Soc. Rev., 1990, 19, 407.
38 D. K. Russell, Chem. Vap. Deposition, 1996, 2, 223 and references
therein.
39 Spartan 02, Wavefunction Inc., Irvine, CA, USA, 2002.
40 J. A. Miller, R. J. Kee and C. K. Westbrook, Ann. Rev. Phys.
Chem., 1990, 41, 345.
126, 771.
G. Schopf and G. Kobmehl, Polythiophenes-Electrically Conduc-
tive Polymers, Springer, Berlin, 1997 and references therein.
M. A. Grela, V. T. Amorebieta and A. J. Colussi, J. Phys. Chem.,
1
985, 89, 38.
A. Lifshitz, M. Bidani and S. Bidani, J. Phys. Chem., 1986, 90,
373.
P. P. Organ and J. C. Mackie, J. Chem. Soc., Faraday Trans.,
991, 87, 815.
5
1
J. M. Patterson, A. Tsamasfyros and W. T. Smith, J. Heterocycl.
Chem., 1968, 5, 727.
8
9
W. R. Johnson and J. C. Kang, J. Org. Chem., 1971, 36, 189.
A. E. Axworthy, V. H. Dayan and G. B. Martin, Fuel, 1978,
5
7, 29.
A. Lifshitz, C. Tamburu and A. Suslensky, J. Phys. Chem, 1989,
3, 5802.
J. C. Mackie, M. B. Colkett, P. F. Nelson and M. Esler, Int.
J. Chem. Kinet., 1991, 23, 733.
M. Szwarc, Chem. Rev., 1950, 47, 135.
1
0
1
9
1
12
13
14
0
R. Spielman and C. A. Cramers, Chromatographia, 1972, 5, 295.
E. Gey, B. Ondrushka and G. Zimmerman, J. Prakt. Chem., 1987,
329, 511.
1
5
6
R. G. Butler, MSc Thesis, Princeton University, Princeton, NJ,
USA, 1992.
M. B. Colket, The Pyrolysis of Cyclopentadiene, Eastern States
section of the Combustion Institute Annual Meeting in Orlando,
FL, 1990.
A. Burcat and M. Dvinyaninov, Int. J. Chem. Kinet., 1997,
29, 505.
J. K. Winkler, W. Karow and P. Rademacher, J. Anal. Appl. Pyr-
olysis, 2002, 62, 123.
K. Sendt, G. B. Bacskay and J. C. Mackie, J. Phys. Chem. A,
41 H. Bibas, M. W. Wong and C. Wentrup, J. Am. Chem Soc., 1995,
117, 9382.
42 H. Bibas, M. W. Wong and C. Wentrup, Chem.-Eur. J., 1997,
3, 237.
43 G. A. Cordell, J. Org. Chem., 1975, 40, 3161.
44 H. M. Gilow and D. E. Burton, J. Org. Chem., 1981, 46, 2221.
45 P. Audebert and G. Bidan, Synth. Met., 1986, 15, 9.
46 L. V. Moskaleva and M. C. Lin, J. Comput. Chem., 2000, 21, 415.
47 G. R. Allen and D. K. Russell, Chem. Commun, 2002, 1960.
48 G. Xu, T.-M. Chang, J. Zhou, M. L. McKee and P. B. Shevlin,
J. Am Chem. Soc., 1999, 121, 7150.
1
17
18
19
20
2000, 104, 1861.
G. B. Bacskay, M. Martoprawiro and J. C. Mackie, Chem Phys.
Lett., 1998, 290, 391.
49 C. Barckholtz, T. A. Barckholtz and C. M. Hadad, J. Am. Chem.
Soc., 1999, 121, 491.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 6 0 6 – 6 1 3
613