Thermal rearrangements of perchlorohexatrienes-structures and experimental and theoretical evaluation of pathways to isomerization and cyclization

Article abstract of DOI:10.1002/ejoc.200801076

We have prepared trans- (1) and cis-octachloro-1,3,5-hexatriene (2) by known routes and studied their thermal behavior experimentally and theoretically by ab initio calculations. The three double bonds in 1 and 2 are completely decoupled due to steric hindrance by the eight Cls, as indicated by calculations as well as the single-crystal X-ray structure of 1. The cis isomer 2 can be isomerized to the trans isomer 1 by heating it to 220-250 °C either neat or dissolved in highboiling solvents, leading to a roughly 2:1 mixture of trans and cis isomers. Calculations at several different levels of theory predict 1 and 2 to be isoenergetic within 2 kJmol^-1. Unimolecular cis/trans isomerization is predicted to occur through an unusual vinylcyclobutene intermediate 7, whose formation faces a barrier of more than 150 kJmol ^-1, but whose stability is comparable to that of 1 and 2. The isomerization rate is strongly enhanced by the addition of small amounts of Br₂ or Cl₂ or by 3 and can be explained by a radical-induced isomerization mechanism. The heating of trienes 1 and 2 to 250 °C leads to cyclization, yielding 71% of the cyclopentene isomer 3. Compound 3 can be dechlorinated by treatment with copper powder to give fulvene derivative 4. Using flash vacuum pyrolysis, the thermal conversion of trienes 1 and 2 to hexachlorobenzene (5) occurs at higher temperatures between 600-1000 °C, likely via perchlorinated 1,3-cyclohexadiene (6) as an intermediate. The elimination of molecular Cl₂ from 3 and 6 requires very high activitation energies in agreement with calculations. Wiley-VCH Verlag GmbH & Co. KGaA, 2009.

Full text of DOI:10.1002/ejoc.200801076

FULL PAPER
DOI: 10.1002/ejoc.200801076
Thermal Rearrangements of Perchlorohexatrienes–Structures and Experimental
and Theoretical Evaluation of Pathways to Isomerization and Cyclization
Heiner Detert,*[a] Dieter Lenoir,[b] and Hendrik Zipse*[c]
Dedicated to Barry Dellinger
Keywords: Perchlorinated compounds / Isomerization / Cyclization / Pyrolysis / Aromaticity / Thermochemistry / Ab initio
calculations
We have prepared trans- (1) and cis-octachloro-1,3,5-hexatri- is strongly enhanced by the addition of small amounts of Br2
ene (2) by known routes and studied their thermal behavior
experimentally and theoretically by ab initio calculations.
The three double bonds in 1 and 2 are completely decoupled
due to steric hindrance by the eight Cls, as indicated by
calculations as well as the single-crystal X-ray structure
of 1. The cis isomer 2 can be isomerized to the trans isomer
1 by heating it to 220–250 °C either neat or dissolved in high-
boiling solvents, leading to a roughly 2:1 mixture of trans and
cis isomers. Calculations at several different levels of theory
predict 1 and 2 to be isoenergetic within 2 kJmol–1. Unimo-
lecular cis/trans isomerization is predicted to occur through
an unusual vinylcyclobutene intermediate 7, whose forma-
tion faces a barrier of more than 150 kJmol–1, but whose sta-
bility is comparable to that of 1 and 2. The isomerization rate
or Cl2 or by 3 and can be explained by a radical-induced
isomerization mechanism. The heating of trienes 1 and 2 to
250 °C leads to cyclization, yielding 71% of the cyclopentene
isomer 3. Compound 3 can be dechlorinated by treatment
with copper powder to give fulvene derivative 4. Using flash
vacuum pyrolysis, the thermal conversion of trienes 1 and 2
to hexachlorobenzene (5) occurs at higher temperatures be-
tween 600–1000 °C, likely via perchlorinated 1,3-cyclohexa-
diene (6) as an intermediate. The elimination of molecular
Cl2 from 3 and 6 requires very high activitation energies in
agreement with calculations.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
pathways have been postulated for the formation of these
chloroaromatic compounds: (1) by so-called de novo syn-
thesis from particulate C and (2) by inorganic Cl and/or the
formation from chlorinated precursors (“precursor
theory”).[4] A variety of chloroaliphatic, four-C and six-C
compounds has been identified by this method, and rel-
evant mechanisms have been suggested.[5] These growing re-
actions can start with simple chloromethanes and involve
perchlorinated alkenes like ethylene, propene, butadiene,
and the coupling products thereof. Though the formation
of several of these perchloroalkenes has been observed, the
mechanisms for their formation and transformation to
hexachlorobenzene are not clear. Whereas the chemistry of
the corresponding hydrocarbons is well-understood, one
must expect that the steric and electronic effects of the Cl
substituents strongly affect the stability of reactants, prod-
ucts, and transient intermediates. The interpretation of the
reaction mechanisms of perchlorinated alkenes in analogy
to simple hydrocarbon analogs may thus be misleading.
Our particular focus here is on the cyclization reactions of
perchloro-1,3,5-hexatriene (1), whose cyclization leads to a
methylenecyclopentene product instead of the expected cy-
clohexadiene.[6] Triene 1 represents a key intermediate in the
Introduction
The thermal formation of various classes of chloroaro-
matic compounds like chlorobenzenes, chlorophenols, and
chlorinated dibenzodioxins and furans (PCDD/F) has been
observed in technical incineration processes like municipal
waste incineration and other thermal processes.[1] These
micropollutants enter the environment by the emitted gases,
and some of the products are toxic, like PCDD/F.[2] There-
fore, the thermal formation pathways of PCDD/F have
been intensively investigated in the last two decades, and it
was shown that these compounds are mainly formed in the
postcombustion zone of the incinerators at temperatures of
300–450 °C on the surface of fly ash particles.[3,4] Two main
[a] Institut für Organische Chemie der Universität Mainz,
Duesbergweg 10-14, 55099 Mainz, Germany
E-mail: detert@uni-mainz.de
[b] Institut für ökologische Chemie, Helmholtz-Zentrum München,
Postfach 1129, 85778 Neuherberg bei München, Germany
E-mail: lenoir@helmholtz-muenchen.de
[c] Department Chemie und Biochemie, LMU München,
Butenandtstrasse 5-13, 81377 München, Germany
E-mail: zipse@cup.uni-muenchen.de
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2009, 1181–1190
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1181

H. Detert, D. Lenoir, H. Zipse
FULL PAPER
growing process, and its cyclization may thus be part of the Roedig et al.[8e] with KOH in acetone to yield trans- and
main reaction pathway leading to hexachlorobenzene.[5,7] cis-octachloro-1,3,5-hexatriene in a 4:1 ratio. The dechlori-
Pattern analysis has also been used to elucidate the mecha- nation with triphenylphosphane in diethyl ether at 0 °C re-
nism.[8a] In order to obtain a detailed picture of the cycliza- sulted in a 8:1 mixture of 2/1 (see Exp. Sect.). The thermal
tion/elimination reactions leading from 1 to hexachloroben- isomerization of 2 to trans isomer 1 had been performed by
zene, we have now studied the thermal reactivity of 1 in the Roedig et al. either by heating it to 250 °C[8d] or by boiling
condensed phase and in the gas phase using a combination it in fuming nitric acid and adding concentrated HCl.[8e]
of experimental and theoretical techniques.
We observed several by-products when following the latter
procedure. Therefore, we heated the mixture of the trans
and cis trienes (1:4) to 220 °C until the ratio reached 1:1.
We detected only traces of other products 3–5 by GC. The
isolation and purification of 1 was possible by crystalli-
zation from methanol. All the products were characterized
by IR, MS, and 13C NMR spectroscopy. Additionally, we
obtained single-crystal X-ray structures from the solid trans
triene 1, perchloromethylenecyclopentene 3, and the fulvene
derivative 4 (see Figures 1, 2, and 3).
Results and Discussion
Experimental Cyclization of Perchlorohexatrienes 1 and 2
There are 12 structural isomers with the formula C6Cl8,
but most isomers with three- or four-membered rings are
too strained to exist as stable species in higher temperature
regimes. Until now, only five C6Cl8 isomers have been pre-
pared and characterized.[8] Some of these isomers are most
The X-ray diffraction of crystals of triene 1 shows a
likely precursors for hexachlorobenzene as the thermody- unique disordered packing of the molecules (Figure 1). We
namically most stable product under the high-temperature found two superimposed molecules to occupy identical po-
conditions of incineration processes. They can be formed sitions for the eight Cl atoms and the two central C atoms,
by so-called “growing reactions” from dichloroacetylene as but the peripheral vinyl groups are disordered. For both
the starting compound, which was shown experimentally molecular units, the positions of Cl-1B/Cl-1BЈ are identical,
using a tubular furnace reactor.[5] Scheme 1 describes sev- but the positions of Cl-2/Cl-2Ј and Cl-1A/Cl-1AЈof the first
eral pathways for the formation of hexachlorobenzene (5) unit are adopted by Cl-1A/Cl-1AЈ and Cl-2/Cl-2Ј of the sec-
starting from trans-octachloro-1,3,5-hexatriene (1). These ond unit. We found two independent sets of positions for
involve the C6Cl8 isomers 2, 3, 6, and 7 as potential inter- the C atoms C-1,C-2/C-1Ј,C-2Ј with crossed double bonds.
mediates, as well as perchlorofulvene (4)[8f] as the product The length of the central double bond (1.330 Å) is similar
of Cl elimination from 3.
to that found for tetrachloroethylene (1.372 Å),[11] while the
The syntheses of trans- and cis-octachloro-1,3,5-hexatri- peripheral double bonds are slightly shorter (1.318 Å). Also
ene (1 and 2, respectively), octachloromethylenecyclopen- the lengths of the C–Cl bonds (1.715–1.731 Å) are in good
tene (3), and the fulvene derivative 4 have been described agreement with the value found for tetrachloroethylene
by the groups of Prins[9] and of Roedig[8c–8f] previously. (1.724 Å). The three vinylidene groups are connected by
Further studies by Simonov et al. concerned the electro- single bonds, each with a length of 1.480 Å. A bond angle
chemical properties of these compounds.[10] We followed the of 115.5° at the terminal Cl–C–Cl group reflects the re-
procedure given by Prins[9b] for the synthesis of deca- duced degree of steric compression compared with that of
chloro-1,5-hexadiene, which was dechlorinated according to C2Cl4 (113.3°). The structural data obtained from the X-
Scheme 1. Isomerization and cyclization reactions of hexatrienes 1 and 2.
1182
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1181–1190

Thermal Rearrangements of Perchlorohexatrienes
Figure 1. X-ray structure of 1. Selected bond lengths and bond
angles [values calculated at the B3LYP/6-31G(2d) level are given in
square brackets]: C(1)–C(2): 1.318 Å [1.342], C(1)–Cl(1B): 1.723 Å
[1.724], C(1)–Cl(1A): 1.715 Å [1.732], C(2)–C(3): 1.480 Å [1.471],
C(2)–Cl(2): 1.727 Å [1.745], C(3)–C(3Ј): 1.330 Å [1.341], C(3)–
Cl(3): 1.731 Å [1.752], C(2)–C(1)–Cl(1B): 122.7°, C(2)–C(1)–
Cl(1A): 121.8°, Cl(1A)–C(1)–Cl(1B): 115.5°, C(1)–C(2)–C(3):
123.0°, C(1)–C(2)–Cl(2): 120.9°, C(3)–C(2)–Cl(2): 116.1°, C(3Ј)–
C(3)–C(2): 126.8°, C(3Ј)–C(3)–Cl(3): 119.8°, C(2)–C(3)–Cl(3):
113.2°, Cl(1B)–C(1)–C(2)–C(3): –178.6°, Cl(1A)–C(1)–C(2)–C(3):
1.4°, Cl(1B)–C(1)–C(2)–Cl(2): –1.2°, Cl(1A)–C(1)–C(2)–Cl(2):
178.8°, C(1)–C(2)–C(3)–C(3Ј): –74.1° [–84.0], Cl(2)–C(2)–C(3)–
C(3Ј): 108.3°, C(1)–C(2)–C(3)–Cl(3): 111.1°, Cl(2)–C(2)–C(3)–
Cl(3): –66.5°.
Figure 2. X-ray structure of 3. Selected bond lengths and bond
angles [values calculated at the B3LYP/6-31G(2d) level are given in
square brackets]: Cl(1)–C(2): 1.818 Å [1.819], Cl(2)–C(2): 1.796 Å,
Cl(3)–C(3): 1.667 Å, Cl(4)–C(4): 1.706 Å [1.709], Cl(5)–C(5):
1.724 Å, Cl(6)–C(5): 1.766 Å, Cl(7)–C(6): 1.728 Å [1.724], Cl(8)–
C(6): 1.719 Å, C(1)–C(6): 1.320 Å [1.342], C(1)–C(2): 1.448 Å,
C(1)–C(5): 1.461 Å [1.518], C(2)–C(3): 1.587 Å [1.501], C(3)–C(4):
1.309 Å [1.338], C(6)–C(1)–C(2): 122.7°, C(6)–C(1)–C(5): 122.7°,
C(2)–C(1)–C(5): 114.3°, C(1)–C(2)–C(3): 102.4°, C(1)–C(2)–Cl(2):
118.2°, C(3)–C(2)–Cl(2): 108.5°, C(1)–C(2)–Cl(1): 115.2°, C(3)–
C(2)–Cl(1): 99.5°, Cl(2)–C(2)–Cl(1): 110.7°, C(4)–C(3)–C(2):
104.7°, C(4)–C(3)–Cl(3): 127.7°, C(2)–C(3)–Cl(3): 121.5°, C(3)–
C(4)–Cl(4): 124.7°, C(1)–C(5)–Cl(5): 123.4°, C(1)–C(5)–Cl(5):
111.8°, Cl(5)–C(5)–Cl(6): 110.0°, C(1)–C(6)–Cl(8): 121.8°, C(1)–
C(6)–Cl(7): 123.8°, Cl(8)–C(6)–Cl(7): 114.9°, C(6)–C(1)–C(2)–C(3):
–178.3°, C(5)–C(1)–C(2)–C(3): 6.6°, C(6)–C(1)–C(2)–Cl(2): 62.6°,
C(5)–C(1)–C(2)–Cl(2): –112.5°, C(6)–C(1)–C(2)–Cl(1): –71.5°,
C(5)–C(1)–C(2)–Cl(1): 113.4°, C(1)–C(2)–C(3)–C(4): –4.4°, Cl(2)–
C(2)–C(3)–C(4): 121.4°, Cl(1)–C(2)–C(3)–C(4): –123.0°, C(1)–
C(2)–C(3)–Cl(3): 179.1°, Cl(2)–C(2)–C(3)–Cl(3): –55.1°, Cl(1)–
ray analysis of trans triene 1 and calculations at the B3LYP/
6-31G(2d) level are in good agreement. The measured C–
Cl bond lengths of 1.723/1.715/1.727 Å are slightly shorter
(Յ1%) than the calculated values of 1.723/1.732/1.745 Å.
The same holds for the C–C distances of 1.318/1.480/
1.330 Å compared to the theoretical values of 1.342/1.471/ C(2)–C(3)–Cl(3): 60.5°, C(2)–C(3)–C(4)–Cl(4): –171.6°, Cl(3)–
C(3)–C(4)–Cl(4): 4.6°, C(6)–C(1)–C(5)–Cl(5): –66.4°, C(2)–C(1)–
1.341 Å. We found the three vinyl units to be nearly planar
C(5)–Cl(5): 108.7°, C(6)–C(1)–C(5)–Cl(6): 68.4°, C(2)–C(1)–C(5)–
in the X-ray structure and completely planar in the calcu-
Cl(6): –116.5°, C(2)–C(1)–C(6)–Cl(8): –1.5°, C(5)–C(1)–C(6)–Cl(8):
173.2°, C(2)–C(1)–C(6)–Cl(7): –176.1°, C(5)–C(1)–C(6)–Cl(7):
lated structure. We found slightly larger deviations for the
dihedral angles between the central and lateral vinyl units
with 74.1° (X-ray) and 84.0° (calculation). Still, both ap-
proaches indicate the complete decoupling of the triene sys-
tem. Calculations indicate this also to be true for the cis
isomer 2. This is in strong contrast to the fully planar non-
chlorinated parent system C6H8.[23a]
–1.4°.
lated structural data gives good agreements for the C–Cl
The X-ray analysis of crystalline perchloromethylenecy- bond lengths. We obtained larger diversions for the C
clopentene (3) shows unsymmetrical molecules. The re- framework. The values calculated for the exocyclic double
duced symmetry can result from disordered packing of the bond (1.343 Å), the adjacent single bonds (1.518 Å), and
molecules in the crystal, a well-known problem in the deter- the endocyclic double bond (1.338 Å) are significantly
mination of crystal structures of highly chlorinated com- larger than those from the X-ray analysis (1.320/1.448/
pounds.[12] The deviations from planarity are in the range 1.309 Å).
of 0.9 to 8.4°. We found the length of the exocyclic double
The crystals of fulvene 4 are built from completely planar
bond to be 1.320 Å, and the C–Cl bonds of 1.719/1.728 Å molecules with dihedral angles of 4° or less. The molecular
length are close to the values found for the terminal double structure of 4 is well-described by the Lewis formula in
bonds of 1. We found strongly reduced bond lengths for the Scheme 1, featuring short exocyclic and endocyclic C–C
endocyclic double bond (1.309 Å) and the corresponding double bonds of 1.354/1.346 Å and longer lengths of 1.483/
C–Cl bonds (1.667/1.706 Å). C–C single bonds connecting 1.487 Å for the formal C–C single bonds. Similar values
the endocyclic double bond and the adjacent dichlorometh- have been found in the X-ray crystal structure of perchlo-
ylene C are significantly longer (1.587 Å) than bonds con- roheptafulvalene,[13] indicative of the highly localized
necting the latter with the exocyclic vinylidene group (1.448/ double bonds in 4. The structural data from the X-ray
1.461 Å). These deviations most likely result from disor- analysis are in excellent agreement with the calculated data;
dered packing as the calculated gas-phase structure is fully the calculated values of the C–Cl bonds are slightly higher
C2v symmetric. A comparison of the measured and calcu- (≈ 1%) than these obtained from X-ray diffraction.
Eur. J. Org. Chem. 2009, 1181–1190
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1183

H. Detert, D. Lenoir, H. Zipse
FULL PAPER
3 rose to 71%, and the ratio of the hexatrienes remained
constant at 2.5:1. The addition of azobisisobutyronitrile as
a radical-chain initiator or CuCl to the mixture of the pure
compounds did not have a significant effect on reaction
rates or the composition of the reaction mixture. The ques-
tion of the thermodynamic equilibrium of the isomers 1 and
2 will be further discussed below.
The origin of the catalytic effect of 3 on the cis/trans
isomerization and the autocatalysis of the cyclization to 3
is probably the fission of one of the allylic C–Cl bonds in
3, generating a Cl· and a planar pentadienyl radical. The
weakness of these C–Cl–bonds in 3 results in the formation
of traces of fulvene 4 (≈ 3%) if 3 is heated for 3 h to 220 °C.
In solution, 3 is quantitatively dechlorinated to 4 with cop-
per powder. The reduced rate of the cis/trans isomerization
and the suppression of the ring closure to 3 in n-octade-
Figure 3. X-ray structure of 4. Selected bond lengths and bond
angles [values calculated at the B3LYP/6-31G(2d) level are given in
square brackets]: Cl(1)–C(1): 1.701 Å [1.722], Cl(2)–C(2): 1.704 Å cane, together with the formation of high-boiling hydro-
[1.713], Cl(3)–C(3): 1.700 Å, Cl(4)–C(4): 1.704 Å, Cl(5)–C(6):
carbons as by-products, indicate a trapping of intermediate
1.704 Å [1.726], Cl(6)–C(6): 1.712 Å, C(1)–C(2): 1.345 Å [1.357],
radicals. We further illustrated the pivotal role of Cl in these
C(1)–C(5): 1.483 Å [1.485], C(2)–C(3): 1.456 Å [1.460], C(3)–C(4):
reactions by saturating a trans/cis mixture (1:8) with Cl2 at
1.346 Å, C(4)–C(5): 1.481 Å, C(5)–C(6): 1.354 Å [1.357], C(2)–
ambient temperature and heating the mixture to 220 °C in
an open tube. Within 15 min, the trans/cis ratio increased
to a value of 2.7:1, and 37% of the hexatrienes underwent
ring closure to 3. Like Cl2, the presence of Br2 strongly
accelerated the cis/trans isomerization. Within 1 h at 220 °C,
C(1)–C(5): 108.2°, C(2)–C(1)–Cl(1): 123.1°, C(5)–C(1)–Cl(1):
128.6°, C(1)–C(2)–C(3): 109.5°, C(1)–C(2)–Cl(2): 127.1°, C(3)–
C(2)–Cl(2): 123.4°, C(4)–C(3)–C(2): 109.1°, C(4)–C(3)–Cl(3):
127.1°, C(2)–C(3)–Cl(3): 123.8°, C(3)–C(4)–C(5): 108.5°, C(3)–
C(4)–Cl(4): 123.1°, C(5)–C(4)–Cl(4): 128.4°, C(6)–C(4)–C(4):
127.8°, C(6)–C(5)–C(1): 127.6°, C(4)–C(5)–C(1): 104.7°, C(5)–
C(6)–Cl(6): 123.9°, Cl(5)–C(6)-Cl(6): 112.3°; all torsion angles are catalytic amounts of Br2 in the trans/cis mixture or in pure
smaller than |3.6°|.
trans-perchlorohexatriene (1) were sufficient to establish the
trans/cis equilibrium of 2.55:1. Contrary to Cl2, Br2 did not
initiate the cyclization. In Scheme 2 we propose a mecha-
nism for the thermal behavior of these compounds, which
is derived from calculations (see below).
Thermal Isomerization in Condensed Phase and in Solution
The thermal rearrangements of 1 and 2 have been
studied by Roedig et al. before.[8f] It was reported that 2
isomerized to 1 within 40 min at 250 °C; an isomerization
of 1 to the cis isomer 2 was not mentioned. Both isomers
were irreversibly converted to 3 within 5 min by heating to
290 °C or, in the presence of an excess of Cl2, within 10 min
at 150 °C.
In our hands, heating the mixture of 1 and 2 neat, in
1,2,4-trichlorobenzene, or in n-octadecane led to the isom-
erization of 2 to 1. At 220 °C, the ratio of 1:2 rose from 1:8
to 1:3 in the neat compounds within 7 h, in trichloroben-
zene within 5 h, and in n-octadecane within 13 h. Contrary
to the solutions, small amounts (Ͻ3%) of 3–5 formed by
the heating of the neat mixture for 7 h at 220 °C. The com-
position of this mixture changed completely upon pro-
longed heating (9 h), whereas the relative amount of 3 in-
creased to 71%, accompanied by 10% of 4 and 6% of 5,
only 9% of 1 and 5% of 2 remained. The formation of 3
was retarded in trichlorobenzene and suppressed in octade-
Scheme 2. Radical-induced cyclization reactions of hexatrienes 1
cane. The addition of 3 (5 mol-%) to a solution of 1 and 2
in trichlorobenzene had a strong accelerating effect on the
cis/trans isomerization; within 2 h at 220 °C, the trans/cis
and 2.
A halogen radical attack on C(3) of perchlorohexatriene
ratio reached 1:2. Starting with pure 1, heating to 220 °C 1 generates an allylic radical and enables the molecule to
for 1 h was sufficient to produce a 3:1 ratio of 1/2 together rotate around the C(3)–C(4) bond. The elimination of a Cl·
with 3 (14%). After 5 h, the composition had changed to a results in trans/cis equilibration. If the terminal C-6 atom
ratio of 53% of 1, 21% of 2, and 26% of 3. At this point, of the allylic radical attacks the π-orbital of the C(1)–C(2)
the cyclization to 3 became fast. Within 1 h, the amount of bond in an exo-trig process, ring closure can occur. Splitting
1184
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1181–1190

Thermal Rearrangements of Perchlorohexatrienes
off the Cl· from C(2) completes the formation of 3 and al-
lows for the propagation of the isomerization/cyclization
process. Cl attack is equally possible at the C(1) or C(2)
positions of hexatriene 1, but the conversion of the primary
adducts to product 3 will then require additional intramo-
lecular Cl migration steps. For the sake of simplicity we will
thus concentrate on the mechanism shown in Scheme 2.
In addition to the thermal isomerization of the trienes 1
and 2 neat and in solution, we performed flash vacuum
pyrolysis experiments. The relative composition of the py-
rolysis products is strongly dependent on the temperature
(Figure 4).
K(1/2) = exp[–∆G(1/2)/r.t.]
(1)
(2)
∆G(1/2) = ∆H(1/2) – T∆S(1/2)
We obtained the free energies of 1 and 2 at different tem-
peratures by combining gas-phase total energies obtained
at the G3(MP2)B3 level with thermal corrections obtained
at the B3LYP/6-31G(2d) level. The latter are based on the
rigid-rotor/harmonic oscillator model and standard
thermodynamic formulae for enthalpies and entropies. As
can be seen in Figure 5, the 1:2 ratio varies between 5.5
and 6.5 over the temperature range considered here, always
favoring the trans isomer 1 by a small margin. The very
small influence of the temperature observed here is due to
the very small enthalpy difference between 1 and 2, which
amounts to 0.42–0.46 kJmol–1 over the respective tempera-
ture range. From the vacuum pyrolysis experiments on mix-
tures of 1 and 2 described above, it can be seen that the
rate of isomerization accelerates in the temperature region
between 400 and 500 K. The almost 1:1 mixture of 1 and 2
obtained at 500 °C may thus correspond to the equilibrium
mixture at this particular temperature. A comparison of the
experimentally and theoretically predicted isomer ratio at
500 °C thus allows us to estimate the accuracy of the theo-
retical approach taken here to Ϯ 7 kJmol–1. The thermody-
namically most stable C6Cl8 isomer, 3, which appears to be
the final product in the condensed phase (below 250 °C) is
only formed in trace amounts. Contrary to the condensed
phase, an intramolecular reaction has to be assumed in the
vacuum pyrolysis experiments. The activation barrier for
the monomolecular cyclization of 1 and 2 must be much
higher than for their cis/trans isomerization. Only at very
high temperatures (800–1100 °C) we observed nearly exclus-
ive formation of hexachlorobenzene (5).
Figure 4. Flash vacuum pyrolysis of mixture of 1 and 2.
In the temperature regime below 650 °C, mainly cis/trans
isomerization occurs. The trans isomer 1 appears to be the
preferred product in the 500–600 °C range. At 500 °C, the
formation of hexachlorobenzene 5 becomes significant and
strongly increases in the 650–700 °C range reaching a pla-
teau of about 96% between 900 and 1100 °C. Traces of
hexachlorocyclopentadiene and hexachlorobutadiene have
been identified by GC as by-products in varying ratios
mainly between 750 and 900 °C. Analogously, starting with
pure 1, we observed mainly isomerization to a mixture of 1
and 2 at temperatures below 600 °C, whereas fragmentation
and the formation of 5 dominated above 700 °C.
The results show that while the isomerization of C6Cl8 1
and 2, other thermal reactions like the formation of
hexachlorobutadiene and hexachlorocyclopentadiene only
occur at more elevated temperatures between 600 and
900 °C. The formation of the latter two products presum-
ably results from retro-reactions.[1b]
trans-Octachlorohexatriene (1) appears to be thermody-
namically more stable than its cis isomer at higher tempera-
tures. The calculated enthalpy difference between 1 and 2 is
Figure 5. Temperature dependence of the 1/2 ratio, as calculated at
the G3(MP2)B3 level of theory.
If this transformation proceeds by cyclization of 1/2 to 3
very small at room temperature, but this difference may as the initial step, only relatively low activation barriers
change in favor of trans 1 at higher temperature. The equi- would separate 3 from 5, since 3 is detected only in minor
librium distribution [K(1/2)] of isomers 1 and 2 has been amounts. Considering that we detected Cl2 and fulvene 4
calculated over the temperature range 278.15–900 K from as by-products, it is possible that fulvene 4 can be another
the free energy difference between these two species intermediate which rearranges to 5 at high temperatures.
[∆G(1/2) = G(1) – G(2)] according to Equation (1). The free According to the results reported by Roedig et al.,[8f] the
energies can be dissected further into enthalpy and entropy photochemical activation of 4 in benzene results in quanti-
contributions according to Equation (2).
tative transformation into 5.
Eur. J. Org. Chem. 2009, 1181–1190
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1185

H. Detert, D. Lenoir, H. Zipse
FULL PAPER
An alternative pathway for the formation of hexachloro- Table 1. Relative energies and enthalpies at 298.15 K of stationary
points on the C6Cl8 potential energy surface (units: kJmol–1).
benzene is the disrotatory, pericyclic, ring closure of the 6π-
electron system to perchloro-1,3-cyclohexadiene 6 followed
by aromatization through the elimination of Cl. This com-
pound had been observed as part of a mass spectrum of a
mixture.[14a] Attempts to prepare intermediate 6 as a pure
compound by the reaction of o-chloranil and PCl5 failed,
yielding an inseparable mixture of 6 and hexachlorobenzene
(5) instead.[14b] The chlorination of p-chloranil resulted in a
mixture of 5 and octachloro-1,4-cyclohexadiene.[14c] Heat-
ing the 1,4-diene to 150 °C resulted in a fast and quantita-
tive formation of 5 by dechlorination and aromatiza-
tion.[14b]
Entry
Structure
∆H298
∆H298
∆H298
(B3LYP)[a]
(MP2)[b]
(G3MP2B3)[c]
1
2
3
4
5
6
7
8
1
0.0
+2.9
–47.5
+51.2
–73.9
–23.4
+41.8
+182.3
+216.7
+180.1
+133.9
+163.5
+190.1
+222.6
+227.7
0.0
+1.9
–92.3
+84.9
–59.0
–86.1
0.0
–0.5
–78.4
+83.3
–44.8
–69.6
+11.6
+171.0
+159.6
+167.7
+119.9
+272.2
+155.0
+130.0
+149.3
2
3
4 + Cl2
5 + Cl2
6
7
–0.2
TS1/7
TS2/3
TS2/6
TS2/7
TS5/6
TS2/8
8
+153.6
+124.9
+130.7
+94.7
+224.2
+119.4
+105.2
+116.21
9
10
11
12
13
14
15
Cyclization Reactions of Hexatrienes 1 and 2 Ϫ Theoretical
Aspects
TS3/8
[a] RB3LYP/6-311+G(3df)//RB3LYP/6-31G(2d). [b] MP2(FC)/
G3MP2large//RB3LYP/6-31G(2d). [c] G3MP2B3 level based on
RB3LYP/6-31G(2d) geometries.
In order to understand the experimental results described
above in more detail, we investigated several conceivable
pathways for the conversion of perchlorohexatrienes 1 and
2. We performed geometry optimizations at the B3LYP/6- fluorinated cyclobutenes, in which strongly electronegative
31G(2d) level of theory and subsequently calculated im- substituents such as CF3 are found to prefer inward rota-
proved relative energies following three different ap- tion.
proaches: (1) the calculation of B3LYP/6-311+G(3df) sin-
All attempts to locate a transition state for direct trans/cis
gle-point energies and their combination with thermal cor- isomerizations of trienes 1 and 2 through rotation around
rections from B3LYP/6-31G(2d) calculations to derive rela- the central C(3)–C(4) bond failed, leading to either TS1/7
tive enthalpies at 298 K, (2) the calculation of MP2(FC)/ or TS2/7 instead.[23b] This is in marked contrast to the
G3MP2large single-point energies and their combination trans/cis isomerization reactions of unsubstituted alkenes,
with thermal corrections from B3LYP/6-31G(2d) calcula- in which this process represents the main reaction pathway
tions to derive relative enthalpies at 298 K; the G3MP2large under thermal reaction conditions.
basis set is closely related to the 6-311+G(2df) basis set and
A second electrocyclic process relevant to the fate of
has been optimized for use in the compound G3MP2 perchlorohexatriene 2 is the disrotatory 6π-electrocycliza-
method, and (c) the calculation of G3MP2B3 enthalpies tion to octachlorocyclohexadiene 6. This reaction is exo-
using a series of single-point calculations as suggested by thermic by 88.0 kJmol–1 but faces a reaction barrier of
Curtiss et al.[15] (Table 1). Of these three levels, the latter +128.8 kJmol–1. The structure of the corresponding transi-
one can be expected to be the most accurate but also the tion state TS2/6 is quite in line with expectations for this
most demanding. A survey of the results obtained at all process.[17] The elimination of Cl2 from 6 can potentially
three levels in Table 1 shows that the MP2 results are some- yield hexachlorobenzene 5, but this step is thermochemi-
what closer to those obtained at the G3MP2B3 level than cally endothermic by 27.1 kJmol–1 and features a very high
they are to the B3LYP results. Since G3MP2B3 calculations reaction barrier of over 310.3 kJmol–1. Aside from the 6π-
were not feasible for the larger open-shell systems discussed and 4π-electrocyclization reactions (forming 6 and 7,
below, the subsequent discussion will focus on the MP2 re- respectively) hexatriene 2 can also cyclize in a hitherto un-
sults exclusively.
known fashion to the five-membered-ring product 3. The
A first point concerns the relative stability of trans-per- formation of this latter compound is exothermic by
chlorohexatriene 1 and its cis isomer 2. These isomers are 94.2 kJmol–1, making 3 the thermochemical sink of the po-
found to be almost isoenergetic at all levels of theory con- tential energy surface shown in Scheme 1. The transition
sidered here (Table 1). One pathway for the interconversion state TS2/3 for the interconversion of 2 and 3 is a true tran-
of 1 and 2 leads through octachloro-3-vinylcyclobutene (7), sition state at the RB3LYP/6-31G(2d) level of theory, fea-
which is predicted to be isoenergetic with 1. Conrotatory, turing only one imaginary frequency corresponding to
electrocyclic, ring opening[16,17] of 7 can occur with the tri- 151 cm–1 in its vibrational frequency spectrum. Intrinsic re-
chlorovinyl substituent rotating outward with a barrier of action path calculations show that TS2/3 connects the min-
+153.8 kJmol–1 through transition state TS1/7, forming ima 2 and 3 without the intermediacy of any further species.
trans-hexatriene 1 as the product. Alternatively, ring open- The structure of TS2/3 has substantial structural similarity
ing can occur with the trichlorovinyl substituent rotating to what may be expected for a biradical intermediate such
inward with a barrier of +94.9 kJmol–1 through transition as 8, and smoothly combines two bond-making/bond-
state TS2/7, now forming cis-hexatriene 2. This “torquose- breaking processes: (i) the formation of a new C–C bond
lectivity”[16] is reminiscent of the reaction outcome in per- between C(2) and C(6) and (ii) Cl migration from C(2) to
1186
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1181–1190

Thermal Rearrangements of Perchlorohexatrienes
C(3) (all with respect to atom numbering in reactant 2). pressively long C–Cl bond adjacent to the radical center
The animation of the imaginary frequency as well as the (1.86 Å). The cleavage of this C–Cl bond through transition
inspection of the structure of TS2/3 shows that the C–C state TS17 is endothermic by 60.8 kJmol–1 and faces an
bond formation is well-advanced while Cl migration has equally large barrier. The reaction passes through the
hardly begun. In order to probe the involvement of biradi- weakly bound complex 18, located 12.9 kJmol–1 below 3 +
cal intermediates, we reinvestigated the potential energy Cl· (Table 2). Taken together, the radical-induced conver-
surface between 2 and 3 at the unrestricted UB3LYP/6- sion of hexatriene 1 through intermediates 14 and 16 to
31G(2d) level of theory. Biradical 8 now represents a true cyclopentene
3 faces a barrier of not more than
intermediate located 103.3 kJmol–1 above reactant 2. The +60.8 kJmol–1, and thus, provides a worthwhile alternative
transition state TS2/8 for the formation of 8 from hexatri- to the unimolecular pathways described in Scheme 1. The
ene 2 is energetically and structurally very similar to transi- highest barrier in this latter case amounts to
tion state TS2/3 located at the RB3LYP level and located +171.0 kJmol–1 if we include the cis/trans isomerization
only 14.2 kJmol–1 above biradical 8. Subsequent Cl mi- process between 1 and 2 or +155.5 kJmol–1 for the reaction
gration through transition state TS3/8 is slightly more facile starting from 2.
with a barrier of +11.0 kJmol–1. A comparison of the se-
quence obtained from RB3LYP geometry optimizations (2–
TS2/3–3) with that obtained at the UB3LYP level (2–TS2/
Table 2. Relative energies and enthalpies at 298.15 K of stationary
points on the Cl· + C6Cl8 potential energy surface (units: kJmol–1).
Entry
Structure
∆H298
∆H298
8–8–TS3/8–3) shows that the predicted reaction barriers are
rather similar in both cases and that the intermediacy of
biradical intermediates such as 8 can neither be strictly ex-
cluded nor confirmed at this stage. The elimination of Cl2
from 3 to yield perchlorofulvene 4 is strongly endothermic
by 177.2 kJmol–1, and the subsequent rearrangement to
perchlorobenzene 5 is exothermic by 143.9 kJmol–1. This
implies that the conversion of 3 to benzene 5 + Cl2 is endo-
thermic by 33.3 kJmol–1 at 298.15 K, highlighting the high
thermochemical stability of 3. The exothermic rearrange-
ment of 4 to 5 has been observed by Roedig et al.[8d] upon
the photolytic activation of fulvene 4.
In addition to the unimolecular isomerization pathways
shown in Scheme 1, it is also conceivable that the cycliza-
tion reactions of hexatrienes 1 and 2 are induced by external
species. One option involves the addition of Cl· atoms as
described in Scheme 2. Starting from weakly bound com-
plexes between Cl· and trienes 1 and 2, the actual barriers
for the addition step are rather low at +1.5 and
+5.3 kJmol–1, respectively. The adduct radicals 12 and 14
can only formally be classified as allyl radicals. A closer
analysis of the structures shows that the bond lengths along
the C–C backbone as well as the corresponding dihedral
angles are more in line with localized C–C double bonds
with little interaction with the adjacent radical center. Rota-
tion around the C(3)–C(4) bond in adduct radical 12 is
quite facile with a barrier of +9.2 kJmol–1. The radical-in-
(B3LYP)[a]
(ROMP2)[b]
1
2
3
4
5
6
7
8
1 + Cl·
10a
TS11a
12
TS13
2 + Cl·
10b
TS11b
14
0.0
–17.07
0.0
–22.51
–21.01
–48.32
–39.14
–2.07
–24.61
–19.30
–64.56
–57.68
–153.11
–112.81
–105.19
–92.27
+4.87
–17.89
+1.12
+2.87
–10.46
+9.08
–27.86
+4.29
–68.58
–45.95
–47.22
–47.46
9
10
11
12
13
14
TS15
16
TS17
18
3 + Cl·
[a] UB3LYP/6-311+G(3df,2p)//UB3LYP/6-31G(2d). [b] ROMP2-
(FC)/G3MP2large//UB3LYP/6-31G(2d).
The absolute barriers cited here depend, of course, sub-
stantially on the chosen level of theory. However, the dra-
matically lower reaction barriers for the radical-induced
isomerization mechanisms outlined in Scheme 2 as com-
pared to the unimolecular isomerization pathways shown in
Scheme 1 persist at any level of theory studied here.
Conclusions
The thermal behavior of C6Cl8 isomers 1–3 as well as the
duced cis/trans isomerization involving the sequence 1–10a– C6Cl6 isomers 4 and 5 were studied by heating the com-
TS11a–12–TS13–14–TS11b–10b–2 thus provides an at- pounds in neat form and in n-octadecane or trichloroben-
tractive, low-barrier alternative to the unimolecular isomer- zene solutions. The trans- and cis-octachloro-1,3,5-hexatri-
ization described in Scheme 1. The highest barrier along the enes 1 and 2 show thermal equilibration from both sides at
former sequence amounts to +45.3 kJmol–1, which com- 220 °C, and the rate is enhanced by catalytic amounts of
pares favorably to the barrier for the latter sequence of Br2, Cl2, or 3; cis-triene 2 can be regarded as the key com-
+153.6 kJmol–1 (both at the MP2 level of theory). The cy- pound en route to the cyclized 3 and 4 at 250 °C; at higher
clization of radical 14 to provide the five-membered-ring temperatures, between 600–1000 °C, the formation of
cyclopentenylmethyl radical 16 is equally facile, facing a hexachlorobenzene (5) is observed, likely via octachloro-
barrier of only +6.9 kJmol–1. Radical 16 is exceedingly 1,3-cyclohexadiene (6) as an intermediate. The calculations
stable from a thermochemical perspective and represents of all compounds and the corresponding transition states
the lowest energy point of the potential energy surface can explain and accommodate the experimental findings
shown in Scheme 2. The structure of 16 is fully in line with and highlight the catalytic effects of open-shell intermedi-
the Lewis structure shown in Scheme 2 and sports an im- ates. These reactions are fundamental for understanding the
Eur. J. Org. Chem. 2009, 1181–1190
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1187

H. Detert, D. Lenoir, H. Zipse
FULL PAPER
137.5 ppm. UV (ethanol): λ (logε) = 216 (4.97) nm. C6Cl8
(355.690): calcd. C 20.26%; found C 20.26%.
formation of chlorinated micropollutants such as 5, which
are observed in technical incineration processes and are an
important indicator and precursor for PCDD/F.
Hexachlorofulvene (4): Methanol (20 mL) was added to a warm
solution of 3 (3.56 g, 0.01 mol) in toluene (6 mL) while stirring.
Copper powder (1.42 g, 0.022 mol) was added, and the suspension
was stirred for 4 h at ambient temperature. After adding dichloro-
methane (30 mL), the mixture was filtered into water (40 mL), the
organic layer was separated, the aqueous phase was extracted with
dichloromethane, and the pooled organic solutions were washed
with water and dried with MgSO4. Purification by chromatography
on silica gel with cyclohexane as an eluent (Rf = 0.95) yielded 3.31 g
(93%) of product; m.p. 153–154, ref.[8d] m.p. 152–153 °C IR (KBr):
Experimental Section
General: Meltings points are uncorrected. All NMR spectra were
recorded with a Bruker AC 300 spectrometer, and TMS was used
as an internal standard. MS was performed with a Finnigan MAT
95 instrument. IR spectra were recorded with a Perkin–Elmer Para-
gon 500 FT-IR spectrometer. UV/Vis spectra were recorded with a
Perkin–Elmer Lambda 16 spectrometer. GLC conditions: Perkin–
Elmer DB 35MS (30 m) column, temperature program: start at
150 °C, then 2 °C/min to 180 °C, then 10 °C/min to 240 °C, then
240 °C for 5 min, FID detector, N2 carrier gas, Merck–Hitachi D
2500 Chromato-Integrator.
ν = 1552, 1567, 1279, 1257, 1223, 931, 709, 637 cm–1. 13C NMR
˜
(75 MHz, CDCl3): δ = 136.9, 131.0, 130.5, 117.5 ppm. UV (etha-
nol): λ (logε) = 206 (3.94), 298 (4.07), 309 (4.07), 423 (2.49) nm.
C6Cl6 (284.784): calcd. C 25.31%; found C 24.90%.
Flash Vacuum Pyrolysis Experiments: The sample (100 mg) was
placed in a flask (100 mL) directly connected to a quartz tube
(25 mm inner diameter and 60 cm length) placed in a horizontal
tubular furnace (heated zone: 35 cm) and connected to a cold trap
(–196 °C). Pyrolysis was performed at 10–2 Torr and temperatures
between 300 and 1100 °C. After the experiment, the cold trap was
flushed with N2, warmed to ambient temperature, and the products
were dissolved in chloroform and analyzed by GLC. The response
factors of the GC detector (FID) of the isomeric trienes 1 and 2
were identical and defined as 1. The relative sensitivity for
hexachlorobenzene 5 was 1.14, whereas the sensitivity for 3 was
only 0.94.
1,5-Decachlorohexadiene: The diene was prepared from hexachlo-
ropropene with cuprous chloride as the coupling agent according
the procedure given by Prins.[9b] IR (KBr): ν = 1518, 1060, 938, 892,
˜
858, 808, 780, 732, 644, 627 cm–1. 13C NMR (75 MHz, CDCl3): δ
= 94.5 (C-3,4), 128.1 (C-1,6), 131.2 (C-2,5) ppm. MS (FD): 426
(100%, Cl10 pattern) [M]+. C6Cl10 (426.596): calcd. C 16.89%;
found C 17.06%.
cis-1,3,5-Octachlorohexatriene (2): The compound was prepared
according to the procedure given by Roedig et al.,[8e] resulting in
a 1:4 mixture of the trans and the cis isomers as a colorless oil.
Alternatively, K2CO3 (2.1 g, 0.015 mol) was added to 1,5-decachlo-
rohexadiene (6.4 g, 0.015 mol) in anhydrous diethyl ether (30 mL)
and 2-propanol (1 mL). The mixture was cooled in a water/ice bath,
and triphenylphosphane (3.9 g, 0.015 mol) was added with stirring.
The mixture was stirred for 90 min at 0 °C, at which time the phos-
phane had been consumed (TLC), and water (30 mL) was added.
The organic layer was separated, and the aqueous phase was ex-
tracted with cyclohexane (50 mL). The pooled organic solutions
were washed with brine (20 mL), dried with Na2SO4, and concen-
trated. After chromatography on silica gel with cyclohexane as an
eluent, a 1:8 mixture of 1 and 2 was obtained as a colorless oil in
Thermolysis of trans- and cis-Octachlorohexatrienes: A mixture of
cis- and trans-1,3,5-octachlorhexatriene (85:15) was used. The tri-
enes (neat or as a 10% solution in 1,2,4-trichlorobenzene or octade-
cane) were heated in a sand bath, and samples were taken and
analyzed by GLC. For the Br2-catalyzed isomerisation, small
amounts of pure Br2 (ca. 30 µL) were added to the sample (1.2 g),
which was placed in an open flask at 220 °C. Within 20 min, the
color faded due to the evaporation of Br2, and the mixture was
analyzed. Three subsequent additions of Br2 were sufficient to es-
tablish an equilibrium of 1:2 = 2.6:1. Catalysis of the isomerization
was performed by bubbling Cl2 through the liquid at ambient tem-
perature and heating the Cl2-saturated mixture in an open flask.
Heating 3 to 220 °C for 3 h in a test tube led to the formation of
small amounts of 4 (Ͻ5%). For the study of the isomerization in
the condensed phase, a mixture of 1 and 2 [neat or as a 0.2  solu-
tion in 1,2,4-trichlorobenzene or n-octadecane] was heated to
220 °C, and samples, taken at 1 h intervals, were analyzed by GLC.
84% yield (4.5 g). IR (neat): ν = 1575, 1540, 1197, 1161, 1131, 970,
˜
951, 898, 860, 831, 808, 683, 628 cm–1. cis-2: 13C NMR (75 MHz,
CDCl3): δ = 130.51 (C-2), 127.24 (C-1), 123.00 (C-3) ppm. UV
(ethanol): λ (logε) = 216 (4.36), 255 (sh, 3.76) nm. C6Cl8 (355.690):
calcd. C 20.26%; found C 20.23%.
trans-1,3,5-Octachlorohexatriene (1): The compound was prepared
from the cis/trans mixture (1.2 g) by heating it for 1 h to 220 °C
and adding Br2 (3ϫ100 mg) until the composition of the mixture
reached a 1:2 ratio of 2.5:1. Purification was performed by repeated
crystallization from dichloromethane and methanol to yield 24%
(0.3 g) of colorless crystals; m.p. 74 °C, ref.[8e] m.p. 72 °C. IR (KBr):
X-ray Structure Determination of 1: Performed with a SMART
CCD Bruker Nonius instrument using a transparent, color-
less plate. Crystal data: C6Cl8
0.12ϫ0.24ϫ0.36 mm3, monoclinic, space group P21/c, Mo-Kα,
graphite monochromation: 0.71069 Å, 193 K, unit cell
M
=
355.7 gmol–1,
T
=
ν = 1592, 1168, 955, 849, 790, 611 cm–1. 13C NMR (75 MHz,
˜
dimensions: a = 6.3235(4), b = 13.0794(7), c = 7.5282(4) Å, β =
109.557(4)°, V = 586.7(7) Å3, Z = 2, dcalc = 2.013 gcm–3, absorp-
tion µ = 1.87 mm–1, the Θ range for data collection was 2° to 28.3°,
and the index ranges were –8ՅhՅ8, –17ՅkՅ17, and
–10ՅlՅ10. Number of reflections collected: 5956; independent re-
flections: 1456 [Rint = 0.0871]. The structure was solved by direct
methods (program SIR 92, refinement by SHELXL-97).[18]
CDCl3): δ = 127.9 (C-3), 126.4 (C-1), 123.2 (C-3) ppm. MS (FD):
356 (100%, Cl8-pattern) [M]+. UV (ethanol): λ (logε) = 212 (4.39),
257 (sh, 3.92) nm. C6Cl8 (355.690): calcd. C 20.26%; found C
20.08%. For an X-ray structure, the product was crystallized from
the same solvent mixture.
Octachloro-4-methylenecyclopent-1-ene (3): The compound was
prepared by heating 1 or a mixture of 1 and 2 for 5 h to 250 °C
and subsequent recrystallization of the product from chloroform/
methanol to yield colorless crystals; m.p. 183 °C, ref.[8e] m.p.
Structure refinement was performed by full-matrix least-squares on
71 parameters, weighted refinement: w = 1/[σ2(Fo2) + (0.0552ϫP)2]
with P = [max(Fo2,0) + 2ϫFo2]/3, and non-hydrogen atoms were
183 °C. IR (KBr): ν = 1638, 1599, 1551, 1168, 916, 840, 801, 630 improved with anisotropic refinement. Goodness-of-fit on S =
˜
cm–1. 13C NMR (75 MHz, CDCl3): δ = 81.8 (C-3, 5), 135.5, 16.5, 0.957, maximum change of parameters 0.001ϫe.s.d, final R in-
1188
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1181–1190

Thermal Rearrangements of Perchlorohexatrienes
dices: R1 = 0.0410, wR2 = 0.1029, the final difference Fourier map
showed minimum and maximum values of –0.62 and 0.62 eÅ–3,
respectively.
gies with the thermal corrections calculated at the (U)B3LYP/6-
31G(2d) level yielded the ∆H298(MP2) values listed in Tables 1 and
2. Finally, improved energies for the system described in Scheme 1
were also obtained with a slight modification of the G3(MP2)B3
compound method.[15] One deviation from the original recipe con-
cerns the use of B3LYP/6-31G(2d) [instead of B3LYP/6-31G(d)]
geometries as well as the use of the same thermochemical correc-
tions as before. Given the possibility of intermediates with some
open-shell character, the QCISD(T)/6-31G(d) single-point calcula-
tions in the original recipe were replaced by CCSD(T)/6-31G(d)
calculations. The use of CCSD(T) [instead of QCISD(T)] calcula-
tions has previously been found by several authors to yield signifi-
cant improvements in the calculations of open-shell systems.[19–21]
The final enthalpies obtained from this procedure are nevertheless
termed “H298(G3MP2B3)” in order to reflect the spirit of the origi-
nal recipe. All calculations have been performed with GAUSSIAN
03.[22]
X-ray Structure Determination of 3: Performed with a SMART
CCD Bruker Nonius instrument using a colorless block. Crystal
data: C6Cl8, M = 355.66 gmol–1, 0.25ϫ0.35ϫ0.45 mm3, mono-
clinic, space group P21/n, Mo-Kα, graphite monochromation:
0.71069 Å, T = 295 K, unit cell dimensions: a = 5.914(6), b =
13,8758(7), c = 7.0237(13) Å, β = 90.088(2)°, V = 592.9(1) Å3, Z =
2, dcalc = 1.992 gcm–3, absorption µ = 1.853 mm–1, the Θ range for
data collection was 2° to 28.3°, and the index ranges were
–7ՅhՅ7, –18ՅkՅ18, and –9ՅlՅ9. Number of reflections col-
lected: 7152; independent reflections: 1464 [Rint = 0.0638]. The
structure was solved by direct methods (program SIR 92, refine-
ment by SHELXL-97).[18]
Structure refinement was performed by full-matrix least-squares on
2
127 parameters, weighted refinement:
w
=
1/[σ2(Fo
)
+
(0.0456ϫP)2 + 0.07ϫP] with P = [max(Fo2,0) + 2ϫFc ]/3, and all
2
Supporting Information (see also the footnote on the first page of
this article): Theoretical data, structures and energies of minima
and transition states.
non-hydrogen atoms improved with anisotropic refinement. Good-
ness-of-fit on
S = 0.992, maximum change of parameters
0.001ϫe.s.d, final R indices: R1 = 0.0278, wR2 = 0.0763, the final
difference Fourier map showed minimum and maximum values of
–0.27 and 0.26 eÅ–3, respectively.
Acknowledgments
X-ray Structure Determination of 4: Performed with an Enraf–Non-
We would like to thank Dr. Dieter Schollmeyer, Johannes Guten-
berg-University of Mainz, for solving the X-ray structures.
ius Turbo-Cad4 instrument equipped with a rotating anode using
a
red needle. Crystal data: C6Cl6,
0.16ϫ0.16ϫ1.392 mm3, monoclinic, space group P21/c, Cu-Kα,
graphite monochromation: 1.54180 Å, 143 K, unit cell
M =
284.75 gmol–1,
[1] Reviews: a) D. Lenoir, A. Wehrmeier, K.-W. Schramm, A.
Kaune, R. Zimmermann, P. H. Taylor, S. S. Sidhu, Environ.
Eng. Science 1998, 15, 37–47; b) P. H. Taylor, D. Lenoir, Sci-
ence Total. Environment 2001, 269, 1–24.
[2] Reviews: a) K. Breivik, R. Alcock, Y. F. Li, R. E. Bailey, H.
Fiedler, J. M. Pacyna, Environ. Pollution 2004, 128, 3–16 2; b)
H. Fiedler, Chemosphere 2007, 67, 96–108.
[3] L. C. Dickson, D. Lenoir, O. Hutzinger, Environ. Sci. Technol.
1992, 26, 1822–1828.
[4] Review: K. Tuppurainen, A. Asikainen, P. Ruokojörvi, J. Ruus-
kanen, Acc. Chem. Res. 2003, 36, 652–658.
[5] a) A. Wehrmeier, D. Lenoir, S. Sidhu, P. H. Taylor, W. R. Ru-
bey, A. Kettrup, B. Dellinger, Environ. Sci. Technol. 1998, 32,
2741–2748; b) P. H. Taylor, A. Wehrmeier, S. S. Sidhu, D.
Lenoir, K. W. Schramm, A. Kettrup, Chemosphere 2000, 40,
1297–1303.
[6] K. Alder, V. Brachel, Justus Liebigs Ann. Chem. 1957, 608, 195.
[7] A. Wehrmeier, D. Lenoir, K.-W. Schramm, R. Zimmermann,
K. Hahn, B. Henkelmann, A. Kettrup, Chemosphere 1998, 36,
2775–2801.
[8] a) A. Wehrmeier, Dissertation Thesis, Technical University of
Munich, 1998; b) H. Hopf, R. Stamm, P. G. Jones, Chem. Ber.
1991, 124, 1291–1294; c) A. Roedig, F. Bischoff, B. Heinrich,
G. Märkl, Justus Liebigs Ann. Chem. 1963, 670, 8–22; d) A.
Roedig, Justus Liebigs Ann. Chem. 1950, 569, 161–183; e) A.
Roedig, G. Voss, E. Kuchinke, Justus Liebigs Ann. Chem. 1952,
580, 24–34; f) A. Roedig, K. Kiepert, Justus Liebigs Ann.
Chem. 1955, 593, 71–80.
[9] a) H. J. Prins, Recl. Trav. Chim. Pays-Bas. 1946, 65, 151–156;
b) H. J. Prins, Recl. Trav. Chim. Pays-Bas. 1949, 68, 419–425;
c) H. J. Prins, Recl. Trav. Chim. Pays-Bas. 1953, 72, 253–256.
[10] V. D. Simonov, V. V. Semenov, M. V. Fominykh, Zh. Obhch.
Khim. 1970, 40, 970–972.
T
=
dimensions: a = 16.318(2), b = 3,7721(9), c = 16.702(2) Å, β =
118.954(5)°, V = 899.6(3) Å3, Z = 14, dcalc = 2.103 gcm–3, absorp-
tion µ = 16.89 mm–1, corrected with 6 surfaces, the Θ range for
data collection was 2° to 73°, and the index ranges were 0ՅhՅ20,
–4ՅkՅ0, and –20ՅlՅ18. Number of reflections collected: 2170;
independent reflections: 1806 [Rint = 0.0407]. The structure was
solved by direct methods (program SIR 92, refinement by
SHELXL-97).[18] Structure refinement was performed by full-
matrix least-squares on 109 parameters, weighted refinement: w =
1/[σ2(Fo2) + (0.1084ϫP)2 + 2.10ϫP] with P = [max(Fo2,0) +
2
2ϫFc ]/3, and all non-hydrogen atoms improved with anisotropic
refinement. Goodness-of-fit on S = 1.120, maximum change of pa-
rameters 0.001ϫe.s.d, final R indices: R1 = 0.0511, wR2 = 0.1553,
the final difference Fourier map showed minimum and maximum
values of –1.01 and 0.97 eÅ–3, respectively.
CCDC-684678 (1), -177377 (3), -684679 (4) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/datarequest/cif.
Theoretical Methods: Geometry optimizations of all stationary
points shown in Scheme 1 were performed at the Becke3LYP/6-
31(2d) level of theory. The same level of theory was used to calcu-
late analytic second derivatives in order to verify the nature of all
stationary points and calculate thermochemical corrections to en-
thalpies at 298.15 K. Structures of open-shell character were opti-
mized at the unrestricted UB3LYP/6-31G(2d) level. Improved rela-
tive energies for all stationary points were then obtained through
single-point calculations at the (U)B3LYP/6-311+G(3df) level of
theory. A combination of these energies with thermal corrections
calculated at the (U)B3LYP/6-31G(2d) level yielded the
∆H298(B3LYP) values listed in Tables 1 and 2. Alternatively, single-
point energies were also calculated at the MP2(FC)/G3MP2large
level, with a restricted reference for both closed- and open-shell
systems. A combination of the (RO)MP2(FC)/G3MP2large ener-
[11] C. N. R. Rao, R. Livingston, Curr. Sci. 1958, 27, 330.
[12] a) F. H. Herbstein, Acta Crystallogr. 1965, 19, 590; b) F. H.
Herbstein, Acta Crystallogr., Sect. B 1979, 35, 1661.
[13] M. Ishimoro, R. West, B. K. Teo, L. F. Dahl, J. Am. Chem.
Soc. 1971, 93, 7101.
[14] a) E. S. Lahaniatis, W. Bergheim, D. Kotzias, G. Pilidis, Chemo-
sphere 1994, 28, 229–236; b) A. Wehrmeier, Dissertation Thesis,
Eur. J. Org. Chem. 2009, 1181–1190
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1189

H. Detert, D. Lenoir, H. Zipse
FULL PAPER
Technical University of Munich, 1998, p. 53–54; c) M. E. Bar-
ral, Bull. Soc. Chim. Fr. 1895, 13, 418–427.
[15] A. G. Baboul, L. A. Curtiss, P. C. Redfern, K. Raghavachari,
J. Chem. Phys. 1999, 110, 7650–7657.
[16] W. R. Dolbier Jr., H. Koroniak, K. N. Houk, C. Sheu, Acc.
Chem. Res. 1996, 29, 471.
[17] K. N. Houk, Y. Li, J. D. Evanseck, Angew. Chem. Int. Ed. Engl.
1992, 31, 682.
[18] G. M. Sheldrick, SHELXL97, Program for Crystal Structure
Refinement, University of Göttingen, Germany 1997.
[19] A. M. Mebel, K. Morokuma, M. C. Lin, J. Chem. Phys. 1995,
103, 7414–7421.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morok-
uma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzew-
ski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, GAUSSIAN 03,
Revision B.03, Gaussian, Inc., Wallingford CT, 2004.
[20] C. W. Bauschlicher, H. Partridge, J. Chem. Phys. 1995, 103,
1788–1791.
[21] a) P. M. Mayer, C. J. Parkinson, D. M. Smith, L. Radom, J.
Chem. Phys. 1998, 108, 604–615; b) D. Wetmore, D. M. Smith,
J. T. Bennett, L. Radom, J. Am. Chem. Soc. 2002, 124, 14054–
14065, and references therein.
[22] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
[23] a) J. Catalán, J. L. G. de Paz, J. Chem. Phys. 2006, 124, 034306–
034311, and references cited therein; b) W. v. E. Doering, W. R.
Roth, F. Bauer, R. Breuckmann, T. Ebbrecht, M. Herbold, R.
Schmidt, H. W. Lennartz, D. Lenoir, R. Boese, Chem. Ber.
1989, 122, 1263–1275, and references cited therein.
Received: November 1, 2008
Published Online: January 28, 2009
1190
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1181–1190