Flash-vacuum-pyrolytic reorganization of angular 4phenylene

Article abstract of DOI:10.1039/b902648g

Flash vacuum pyrolysis of angular 4phenylene furnishes "biphenylene dimer" on route to coronene.

Full text of DOI:10.1039/b902648g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Flash-vacuum-pyrolytic reorganization of angular [4]phenylenew
Peter I. Dosa,a Zhenhua Gu,a Dominik Hager,a William L. Karneyb
and K. Peter C. Vollhardt*a
Received (in College Park, MD, USA) 9th February 2009, Accepted 27th February 2009
First published as an Advance Article on the web 10th March 2009
DOI: 10.1039/b902648g
Flash vacuum pyrolysis of angular [4]phenylene furnishes
‘‘biphenylene dimer’’ on route to coronene.
through a deprotonation–iodination sequence, evidently
facilitated by the relative acidity of the a hydrogens in
biphenylene (Scheme 1).9 The resulting 2 can be iodo-
debrominated by Cu-catalyzed halogen exchange10 intercepting
the published synthesis of 4, or elaborated directly to 3 and then
4 (16.6% overall yield from biphenylene; see the ESIw).
The phenylenes constitute a class of novel s- and p-activated
polycyclic aromatic hydrocarbons (PAHs) composed of
fused alternating benzene and cyclobutadiene rings.1 Their
high energetic content makes them attractive entry points
into carbon-rich hydrocarbon isomerization manifolds, the
mapping of which aids in the understanding of the mecha-
nisms of combustion and soot formation2 and provides
synthetic stimulus for the execution of pyrolytic routes to
new carbon arrays, in particular substructures of carbon
allotropes.3 With relevance to the present work, Scott
proposed a ‘‘benzocyclobutene walk’’ via four-ring scission
and two successive 1,5-H shifts to explain the formation of
fluoranthene from benzo[b]biphenylene via the intermediacy of
benzo[a]biphenylene,4 suggesting the possibility of establishing
topological equilibrations within a group of [N]phenylenes.
The information gleaned from such processes would be
valuable in the verification of computational estimates of heats
of formation.1,5 Moreover, their realization might allow access
to novel phenylene isomers. Indeed, we have reported on the
equilibration of linear with angular [3]phenylene (calculated
DHisomerization ¼ ꢀ2.4 kcal molꢀ1), prior to their reorganiza-
tion to several PAHs (C18H10).6 Consideration of 4 (C24H12)
under these conditions paints a potentially complicated
picture, as it has two close energy phenylene isomers [the
zig-zag (ꢀ0.2 kcal molꢀ1) and bent isomers (þ0.5 kcal molꢀ1)]
and a multitude of possible lower energy PAHs. We describe
the flash vacuum pyrolysis (FVP) of angular [4]phenylene 4,
which, at relatively low temperatures, proceeds remarkably
cleanly and with a surprising outcome.
With suitable quantities of 4 in hand, its FVP was under-
taken in batches ranging from 3–50 mg and at varying
pressures [10ꢀ7 Torr (turbomolecular vacuum pump) to 10ꢀ2
Torr (normal vacuum pump)] by sublimation through a hot
silanized quartz tube. The starting material sublimed
unchanged at 700 1C. Increasing the temperature to 800 1C
revealed the onset of slight conversion, and, at 900 1C, the
substantially clean formation of a new product was clearly
visible (Fig. 1). This compound is an unexpected isomer of 4:
‘‘biphenylene dimer’’ (cycloocta[1,2,3,4-def;5,6,7,8-d0e0f0]-
bisbiphenylene) 5 (Fig. 2),11 and is formed in 15% yield
(45% recovery of 4; 60% mass recovery). Its structure was
confirmed by comparison with an authentic sample. The low
field region of the spectrum depicted in Fig. 1 (see the ESIw)
exhibits only low intensity unresolved signals, indicating the
possible generation of PAHs. The exception is a singlet at
d ¼ 8.93 ppm, caused by the presence of coronene 6
(0.1%; Fig. 2), its identity again ascertained by comparison
with an authentic sample. Partial separation of the
components of the crude pyrolyzate was accomplished by
repeated preparative thin layer chromatography (PTLC) on
silica eluting with hexanes–CH2Cl2 (10
: 1). The fastest
The original nine-step approach to 4 (2.2% overall yield
from 1-bromo-2-iodobenzene) entailed a low-yielding and
capricious step on route to the required angularly functiona-
lized 1,2-diiodobiphenylene, namely the regioselective but
low-yielding (10–30%) CpCo-catalyzed cotrimerization of a
monosilylated o-diethynylbenzene with trimethylsilylacetylene
(TMSA).7 A greatly improved route to such building blocks
features the selective 1-iodination of 2-bromobiphenylene 18
a Department of Chemistry, University of California at Berkeley,
Berkeley, California 94720-1460, USA. E-mail: kpcv@berkeley.edu
b Departments of Chemistry and Environmental Science, University of
San Francisco, 2130 Fulton Street, San Francisco,
California 94117-1080, USA. E-mail: karney@usfca.edu
w Electronic supplementary information (ESI) available: Spectral
details for new compounds, selected 1H NMR spectra of pyrolysis
products, and absolute energies of computed molecules. See DOI:
10.1039/b902648g
Scheme 1 A new preparation of 4.
ꢁc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 1967–1969 | 1967

View Article Online
1
Fig. 1 Partial H NMR spectrum (400 MHz, CDCl3) of crude FVP
products of 4 at 900 1C.
Scheme 2 Proposed mechanism of conversion of 4 to 5. Numbers are
energies of 4 and 5 relative to coronene (kcal molꢀ1), computed
at B3PW91/6-311þG**//B3LYP/6-31G*. All energies are ZPE
corrected, and all structures have zero imaginaries.
collapse into product. The computed exothermicity of this
isomerization (B45 kcal molꢀ1) makes it unlikely that it is
reversible. While seemingly unprecedented, such 1,6-H shifts
may be operating in other pyrolytic PAH interconversions.3d
Scheme 3 portrays a tentative route to 7 that is a slightly
modified benzocyclobuta analog of the mechanism proposed
for the isomerization of angular [3]phenylene to benzo[ghi]-
fluoranthene,6 featuring an initial Stone–Wales type rearran-
gement13 (to A; topological change of central core fusion:
4/6/4 to 5/5/5), strained ring cleavage/H shuttle (to B), vinyl-
idene deinsertion–insertion (to C and E) and a hydrogen
shift/ring contraction (to D). The final stages propose a
‘‘benzocyclobutene walk’’ (to F), followed by final ring
closure, topologically related to Scott’s synthesis of 7.12a In
contrast to the [3]phenylene series, this pathway is minor
Fig. 2 FVP pyrolysis (5–7) and oxidation products (8) of 4.
moving band contained a major contributor to the otherwise
unidentifiable PAH mixture, namely benzo[a]corannulene 7
(Fig. 2), identified by comparison of its spectral data with
those in the literature.12
Subsequent bands contained 4, 6 and then 5. During this
laborious process, it was noted that the amount of recovered 4
decreased continuously, accompanied by the concomitant
appearance of yet another compound. That this was an
oxidative degradation product was confirmed by stirring pure
4 in CHCl3 in air for 72 hours to furnish dione 8 (56%; 14%
recovered 4; Fig. 2). Such oxidations, presumed to involve
singlet oxygen additions, are precedented for the phenylenes.1
Increasing the temperature of the FVP oven by 50 1C
intervals to 1050 1C led to increasing conversions of 4, but
also decreasing mass recovery, the ratio of 5 to 4 reaching a
maximum of 1 : 1. At the same time the proportion of
coronene 6 and traces of other PAHs increased. These findings
suggested that 5 might be the source of 6. Indeed, FVP of 5 at
950 1C greatly increased the proportion of 6 (30%) in the
mixture with starting material and trace PAHs (but no 7),
while, unfortunately, mass recovery decreased to 15%. At
1050 1C the proportion of 6 increased to 50%. Its independent
pyrolysis led to unchanged starting material. An attempt to
purify some of the PAHs by PTLC led to fractions (contami-
nated with some 5 and 6) containing two unknowns, 9 and 10,
in varying proportions, both giving rise to six doublets
(d ¼ 7.2.–8.6 ppm)z in their respective 1H NMR spectra
(see the ESIw). These compounds are new isomers (GCMS),
but could not be further characterized due to their trace
occurrence and sensitivity to air (vide infra).
We propose the following mechanisms to account for the
results. The initial relatively facile reorganization of 4 to 5
plausibly begins by four-membered ring cleavage as in the
‘‘benzocyclobutene walk’’ (Scheme 2).4,6 The resulting
biradical is barred from 1,5-H shifts and thus chooses two
consecutive 1,6-H shifts to reach the topology suitable for
Scheme 3 Proposed mechanism of conversion of 4 to 7. Numbers are
energies relative to coronene (kcal molꢀ1), calculated as in Scheme 2.
ꢁc
This journal is The Royal Society of Chemistry 2009
1968 | Chem. Commun., 2009, 1967–1969

View Article Online
This work was enabled by the NSF (CHE 0451241 and
CHE 0553402). We are grateful to Professor G. Mehta for
spectral data of 7 and to Elizabeth Noey for some computa-
tional results. We are indebted to Professor A. Rajca for a
supply of 5 and helpful interactions.
Notes and references
z 9: 1H NMR (400 MHz, CDCl3): d (assignments by COSY and
computation) ¼ 8.51 (d, J ¼ 8.6 Hz, 2 H5), 8.44 (d, J ¼ 7.8 Hz, 2 H4),
8.35 (d, J ¼ 8.6 Hz, 2 H6), 8.20 (d, J ¼ 7.8 Hz, 2 H3), 7.93 (d, J ¼ 5.2 Hz,
2 H2), 7.85 (d, J ¼ 5.2 Hz, 2 H1). 10: 1H NMR (400 MHz, CDCl3): d ¼
7.91 (d, J ¼ 8.0 Hz, 1 H), 7.60 (d, J ¼ 8.4 Hz, 1 H), 7.46 (d, J ¼ 8.0 Hz,
1 H), 7.37 (d, J ¼ 7.6 Hz, 1 H), 7.23 (AB quartet, 2H). 9: 1H NMR
(calculated with GIAO-B3LYP/6-311þG**//BHLYP/6-31G*) d ¼ 8.71
(d, J ¼ 7.9 Hz, H5), 8.58 (d, J ¼ 7.9 Hz, H6), 8.56 (d, J ¼ 7.3 Hz, H4), 8.25
(d, J ¼ 7.3 Hz, H3), 8.02 (d, J ¼ 5.1 Hz, H2), 7.93 (d, J ¼ 5.1 Hz, H1).
1 O. S. Miljanic and K. P. C. Vollhardt, in Carbon-Rich Compounds:
´
From Molecules to Materials, ed. M. M. Haley and
R. R. Tykwinski, Wiley-VCH, Weinheim, 2006, pp. 140–197.
2 For selected reviews, see: C. S. McEnally, L. D. Pfefferle,
B. Atakan and K. Kohse-Hoinghaus, Prog. Energy Combust.
¨
Sci., 2006, 32, 247–294; Z. A. Mansurov, Combust., Explos. Shock
Waves, 2005, 41, 727–744.
Scheme 4 Proposed mechanism of conversion of 5 to 6. Numbers are
energies relative to coronene (kcal molꢀ1), calculated as in Scheme 2.
3 See, e.g.: (a) V. M. Tsefrikas and L. T. Scott, Chem. Rev., 2006,
106, 4868–4884; (b) A. Necula and L. T. Scott, J. Anal. Appl.
Pyrolysis, 2000, 54, 65–87; (c) R. F. C. Brown, Eur. J. Org. Chem.,
1999, 3211–3222; (d) M. Sarobe, H. C. Kwint, T. Fleer, R. W. A.
Havenith, L. W. Jenneskens, E. J. Vlietstra, J. H. van Lenthe and
J. Wesseling, Eur. J. Org. Chem., 1999, 1191–1200.
because of competition by the trajectory in Scheme 2, leading
to phenylene 5 irreversibly.
Finally, a fairly straightforward, if speculative, route to
coronene 6 can be formulated, exploiting the relatively high
symmetry of 5 (Scheme 4). Thus, eight sequential hydrogen
shift/ring contractions, first as they occur during the FVP of
biphenylene itself4 (to G and H; topological change of
fusion: 6/4 to 5/5) and in related isomerizations (to I and J;
topological change of fusion: 6/5 to 5/6)3b,6 would engender
bis-as-indacenyl J. Two subsequent Stone–Wales shifts
(to K and 6) complete the sequence. The skeletal changes
depicted in Schemes 3 and 4 are heuristic and other pathways
may be actually followed, including radical- or H-catalyzed
alternatives.14 Scheme 4 has the benefit that it allows the
tentative assignment of ‘‘unknown’’ 9 as structure K, on the
basis of observed and calculated chemical shifts and coupling
constantsz and by comparison with those of 1,2 : 5,6-dibenzo-
pentafulvene,15 as well as 2D NMR correlations.
4 D. V. Preda and L. T. Scott, Org. Lett., 2000, 2, 1489–1492.
5 D. Bruns, H. Miura, K. P. C. Vollhardt and A. Stanger, Org. Lett.,
2003, 5, 549–552; J. M. Schulman and R. L. Disch, J. Am. Chem.
Soc., 1996, 118, 8470–8474; J. M. Schulman and R. L. Disch,
J. Phys. Chem. A, 1997, 101, 5596–5599.
6 P. I. Dosa, A. Schleifenbaum and K. P. C. Vollhardt, Org. Lett.,
2001, 3, 1017–1020; A. J. Matzger and K. P. C. Vollhardt, Chem.
Commun., 1997, 1415–1416.
7 P. I. Dosa, G. D. Whitener, K. P. C. Vollhardt, A. D. Bond and
S. J. Teat, Org. Lett., 2002, 4, 2075–2078; R. H. Schmidt-Radde
and K. P. C. Vollhardt, J. Am. Chem. Soc., 1992, 114, 9713–9715.
8 M. E. Cracknell, R. A. Kabli, J. F. W. McOmie and D. H. Perry,
J. Chem. Soc., Perkin Trans. 1, 1985, 115–120.
9 M. K. Shepherd, Cyclobutarenes: The Chemistry of Benzocyclo-
butene, Biphenylene and Related Compounds, Elsevier, New York,
1991, pp. 222–223.
10 A. Klapars and S. L. Buchwald, J. Am. Chem. Soc., 2002, 124,
14844–14845.
11 A. Rajca, A. Safronov, S. Rajca, C. R. Ross, II and J. J. Stezowski,
J. Am. Chem. Soc., 1996, 118, 7272–7279.
12 (a) L. Peng and L. T. Scott, J. Am. Chem. Soc., 2005, 127,
16518–16521; (b) G. Mehta and P. V. V. S. Sarma, Chem.
Commun., 2000, 19–20.
13 See: H. F. Bettinger, B. I. Yakobson and G. Scuseria, J. Am. Chem.
Soc., 2003, 125, 5572–5580 and references therein.
14 For selected illustrative papers, see: A. Stirling, M. Iannuzzi,
A. Laio and M. Parrinello, ChemPhysChem, 2004, 5, 1558–1568;
V. V. Kislow and A. M. Mebel, J. Phys. Chem. A, 2007, 111,
9532–9543; M. R. Nimlos, J. Filley and J. T. McKinnon, J. Phys.
Chem. A, 2005, 109, 9896–9903; R. W. Alder and J. N. Harvey,
J. Am. Chem. Soc., 2004, 126, 2490–2494; A. Violi, J. Phys. Chem.
A, 2005, 109, 7781–7787.
In summary, we have demonstrated that higher phenylenes
may be subject to FVPs that can give rise to other phenylene
topologies. In particular, the relatively clean isomerization of 4
to 5 raises the question whether its linear, bent and zig-zag
isomers will find the same trajectory or traverse their own to
yet other C24H12 isomers. In addition, 5 is the only known
fragment of an extended two-dimensional phenylene net,11
and the present synthesis suggests that it might be possible to
access more extended pieces from phenylenes larger than 4,
such as angular [6] and [8]phenylene.1 These avenues are under
exploration.
15 A. Escher, W. Rutsch and M. Neuenschwander, Helv. Chim. Acta,
1986, 69, 1644–1654.
ꢁc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 1967–1969 | 1969