Probing mechanisms of aryl-aryl bond cleavages under flash vacuum pyrolysis conditions

Article abstract of DOI:10.1071/CH14171

Several biaryls have been subjected to flash vacuum pyrolysis (FVP) at 1100°C and 0.8-0.9hPa. Product compositions are reported for the FVP of 9-phenylanthracene (1), 2-bromobiphenyl (5), biphenyl (8), 1,10- diphenylanthracene (12), 9-(2-naphthyl)anthracene (17), and 9,9′- bianthracenyl (20). The experimental results have been used to evaluate four possible mechanistic pathways for the cleavage of aryl-aryl bonds under these conditions: (1) the 'explosion' of substituted phenyl radicals; (2) hydrogen atom attachment to an ipso-carbon atom of the biaryl followed by C-C bond cleavage; (3) direct homolysis; and (4) loss of a fragment as an aryne. None of these mechanisms by itself successfully accommodates all of the experimental facts. The data suggest that aryl-aryl bond cleavages under FVP conditions involve at least two different mechanistic pathways and that the relative contributions of the competing pathways probably vary from one biaryl to the next.

Full text of DOI:10.1071/CH14171

RESEARCH FRONT
CSIRO PUBLISHING
Aust. J. Chem. 2014, 67, 1279–1287
Full Paper
http://dx.doi.org/10.1071/CH14171
Probing Mechanisms of Aryl–Aryl Bond Cleavages
under Flash Vacuum Pyrolysis Conditions
A
A
A
Edward A. Jackson, Xiang Xue, Hee Yeon Cho,
A B
,
and Lawrence T. Scott
A
Merkert Chemistry Center, Boston College, Chestnut Hill, MA 02467, USA.
Corresponding author. Email: lawrence.scott@bc.edu
B
Several biaryls have been subjected to flash vacuum pyrolysis (FVP) at 11008C and 0.8–0.9 hPa. Product compositions
are reported for the FVP of 9-phenylanthracene (1), 2-bromobiphenyl (5), biphenyl (8), 1,10-diphenylanthracene (12),
9-(2-naphthyl)anthracene (17), and 9,90-bianthracenyl (20). The experimental results have been used to evaluate four
possible mechanistic pathways for the cleavage of aryl–aryl bonds under these conditions: (1) the ‘explosion’ of
substituted phenyl radicals; (2) hydrogen atom attachment to an ipso-carbon atom of the biaryl followed by C–C bond
cleavage; (3) direct homolysis; and (4) loss of a fragment as an aryne. None of these mechanisms by itself successfully
accommodates all of the experimental facts. The data suggest that aryl–aryl bond cleavages under FVP conditions involve
at least two different mechanistic pathways and that the relative contributions of the competing pathways probably vary
from one biaryl to the next.
Manuscript received: 26 March 2014.
Manuscript accepted: 23 April 2014.
Published online: 10 June 2014.
Introduction
Flash vacuum pyrolysis (FVP) often promotes molecular
transformations at high temperatures in the gas phase for which
there is little or no precedent in solution under ordinary labo-
ratory conditions.[1–13] Harnessing these reactions as a uniquely
powerful strategy to synthesize highly strained, bowl-shaped
polycyclic aromatic hydrocarbons (PAHs) began in the early
1990s[12,14–17] and has subsequently enabled syntheses of more
than two dozen of such geodesic polyarenes,[18] as well as the
first laboratory syntheses of an isomerically pure C50C10 hemi-
spherical end-cap of a [5,5]carbon nanotube[19] and fullerene-
C60 in isolable quantities.[20,21] In many of these syntheses, new
5- and/or 6-membered rings are formed by thermal cyclode-
hydrogenation reactions in which a new C–C bond is formed
and two hydrogen atoms are lost. However, the mechanisms by
which these cyclizations occur remain mostly obscure. One such
thermal cyclodehydrogenation that forms a new 6-membered
ring has been shown to occur by way of an initial electrocyclic
reaction before cleavage of either C–H bond.[22] We attempted
to probe the mechanism(s) by which new 5-membered rings are
formed in thermal cyclodehydrogenation reactions by studying
the FVP of several phenyl-substituted PAHs (Scheme 1).
Unfortunately, our plans were thwarted at the outset by the
unanticipated intervention of a significant side reaction, namely,
loss of the phenyl group and recovery of the parent PAH
(Scheme 1).[23] Intrigued by this ‘phenyl loss’ reaction, because
it cleaves a strong aryl–aryl bond, we carried out FVP experi-
ments with several phenyl-substituted PAHs and larger biaryls
to probe the mechanism(s) involved in this process.[23]
ϩ
Scheme 1.
1
ϩ
ϩ
1
2
ϩ
3
4
Scheme 2.
nitrogen carrier gas, and a pressure of 0.8–0.9 hPa, as described
in greater detail in prior publications[13,18] and in the Experi-
mental section. The material balance is never 100 %, so the
product ratios reported herein correspond to the ratios of vari-
ous components in the initial product mixtures as determined by
Results and Discussion
All of the FVP experiments reported in this paper were run using
the same apparatus operating at 11008C with a slow stream of
Journal compilation Ó CSIRO 2014
www.publish.csiro.au/journals/ajc

1280
E. A. Jackson et al.
ϭ 1
2
1
ϭ 2
ϭ 3
ϭ 4
4 1
2
3
3
2
2
4
2
4
4
9.3
9.0
8.7
8.4
8.1
7.8
7.5
7.2
[ppm]
Fig. 1. 1H NMR spectrum (CDCl3, 500 MHz) of the crude product mixture obtained from FVP of 9-phenylanthracene
(1) at 11008C and 0.8–0.9 hPa.
NMR integration and/or gas chromatography–mass spectro-
metry (GC–MS) analysis. In a typical example, FVP of
9-phenylanthracene (1) gives the recovered starting material (1),
benzo[a]fluoranthene (2), anthracene (3), and benz[a]acephe-
nanthrylene (4) in a ratio of 8.5 : 3.0 : 1.5 : 1.0 (determined by
NMR integration, Scheme 2 and Fig. 1).
ϩ
ϩ
ϩ
5
6
7
The majority (,60 %) of the material that passes through the
hot zone is unchanged starting material 1. Benzo[a]fluoranthene
(2) is formed by thermal cyclodehydrogenation, and anthracene
(3) is the result of phenyl loss. Benz[a]acephenanthrylene (4)
almost certainly represents a secondary rearrangement product
derived from a portion of the benzo[a]fluoranthene (2); we
showed many years ago that 5- and 6-membered rings in PAHs
can trade places at high temperatures via a benzene ring
contraction mechanism that simultaneously causes ring expan-
sion of the fused 5-membered ring (5/6 ring swap),[12,13,24–26]
and the present example appears to be nothing more than a
benzannulated case of the known thermal isomerization of
fluoranthene to acephenanthrylene.[24] Therefore, the combined
yields of 2 and 4 (,25 % of the mixture) represent the total
extent of thermal cyclodehydrogenation.
The formation of anthracene (3) raises several questions:
(1) What is the source of the hydrogen atom that ends up where
the phenyl group was originally attached? (2) Which occurs
first: C–C bond rupture or C–H bond formation? (3) What is
the fate of the phenyl group? It is very difficult for us to
determine with confidence whether or not any benzene is
formed in this reaction on account of its high volatility, so the
fate of the phenyl group is unknown. One intriguing possibi-
lity is that the phenyl group might fall apart into smaller pieces
in a process driven by entropy at 11008C (Fig. 2).[27–31]
ϩ
8
9
ϩ
10
11
Scheme 3.
generated the 2-biphenylyl radical by taking advantage of the
known[12,16,18,32–34] fragility of aryl bromides at 11008C. Three
major products were observed: acenaphthylene (6), biphenylene
(7), and biphenyl (8) in a GC–MS ratio of 6 : 3 : 1 (Scheme 3). No
other products accounted for more than 1 % of the pyrolysate,
but naphthalene (9) could be positively identified as ,0.5 % of
the mixture, and phenylacetylene (10) and phenylbutadiyne (11)
were both detected in lower amounts (,0.3 %).
Phenylacetylene (10) and phenylbutadiyne (11) are the
products predicted by the ‘phenyl radical explosion’ hypothesis
(Fig. 2), and their formation confirms that some benzene rings
must indeed fall apart into smaller pieces under these condi-
tions. However, the low abundance of these fragmentation
products clearly shows that ‘phenyl radical explosion’ is a very
minor pathway in this system at 11008C. The biphenyl (8)
presumably arises from scavenging of a hydrogen atom by the
2-biphenylyl radical. Biphenylene (7) probably comes from
ortho-cyclization of the 2-biphenylyl radical,[13,35] with loss of
To test if a substituted phenyl radical could ‘explode’ like
this, we subjected 2-bromobiphenyl (5) to FVP and analysed
the product mixture by GC–MS. To maximize chances of
observing the proposed fragmentation, we intentionally

Aryl–Aryl Bond Cleavages in FVP
1281
ϩ
ϩ
H
H
H
1,2-shift
of H
ϩ
ϩ
H
ϩ
ϩ
H
H
1,2-shift
of H
ϩ
ϩ
ϩ
ϩ
H
H
Fig. 2. Fragments expected from the ‘explosion’ of a substituted phenyl radical. Isomerizations of aryl radicals by 1,2-shifts of
hydrogen atoms under FVP conditions have been previously demonstrated experimentally.[32]
H
H•
12
ϩ
Scheme 4.
12
ϩ
ϩ
ϩ
a hydrogen atom, and acenaphthylene (6) is a known thermal
rearrangement product of biphenylene (7).[36]
FVP of the parent hydrocarbon, biphenyl (8), gives almost
entirely unchanged biphenyl (8). Acenaphthylene (6), the sec-
ond most abundant compound observed, was formed as ,1 % of
the product mixture (GC–MS); biphenylene (7) was found in
even lower amounts, and naphthalene (9) could also be identi-
fied, but no trace of either phenylacetylene (10) or phenyl-
butadiyne (11) could be detected. From these two FVP
experiments, we conclude that (1) the ‘phenyl radical explosion’
pathway really does occur but (2) it is insignificant as a
mechanism for phenyl loss.
An alternative mechanism for phenyl loss is outlined in
Scheme 4. The hydrogen atoms released from the slow build-
up of carbonaceous material on the internal walls of the hot zone
and from every gas phase cyclodehydrogenation that occurs in
these FVP reactions could serve as initiators of the aryl–aryl
bond cleavage. Attachment of a hydrogen atom to one of the
ipso-carbon atoms of the biaryl would set the stage for C–C bond
cleavage to produce an aryl radical and an isolable arene. The
high reactivity of the central ring of anthracene would favour
attachment of hydrogen to the anthracene ipso-carbon atom, in
1
14
ϩ
ϩ
ϩ
13
15
16
ϩ
ϩ
2
3
4
Scheme 5.

1282
E. A. Jackson et al.
17.51
100
95
90
85
80
75
70
65
60
55
50
ϭ 17.51 min
ϭ 25.92 min
ϭ 27.92 min
ϭ 42.78 min
27.92
ϭ 44.75 min
ϭ 68.42 min
45
40
35
30
25
20
15
10
5
25.92
44.75
44.97
68.42
42.78
45.12
45.20
45.36
68.75
68.87
11.82
35.66
60.21
60
79.22
80
86.46
2.63
0
0
20
40
100
Time [min]
Fig. 3. GC–MS of the crude product mixture obtained from FVP of 1,10-diphenylanthracene (12) at 11008C and
0.8–0.9 hPa.
preference to the ipso-carbon atom of the phenyl group, con-
sequently leading to the observed product, anthracene (3), and
a phenyl radical. The phenyl radical generated in this manner
could either attach to the carbonaceous material on the walls or
scavenge a hydrogen atom to produce benzene; in either case, no
additional products would be detected.
To evaluate this mechanism, we designed a competition
experiment to estimate the relative ease of phenyl loss from the
central and terminal rings of anthracene. The ipso-attachment
of hydrogen should be more favourable in the reactive central
ring of anthracene than in a terminal ring, thus facilitating
phenyl loss from the central ring of anthracene when compared
with that from a terminal ring via this mechanism. Accordingly,
we synthesized 1,10-diphenylanthracene (12)[37] that was then
subjected to FVP. Many products were formed. Extensive
chromatography of the pyrolysate allowed us to positively
identify rubicene (13),[38] a product of double cyclodehydro-
genation, the two products of phenyl loss, 9-phenylanthracene
(1) and 1-phenylanthracene (14), the recovered starting material
(12), benzo[a]fluoranthene (2), anthracene (3), benz[a]acephe-
nanthrylene (4), and compounds that we tentatively assigned
as phenyl-substituted benzo[a]fluoranthenes, resulting from a
single cyclodehydrogenation without any phenyl loss (15 and
16, Scheme 5). The most important piece of data, however, is
the GC–MS of the crude pyrolysis mixture (Fig. 3) that shows
the two products of phenyl loss, 9-phenylanthracene (1) and
1-phenylanthracene (14), formed in comparable amounts (less
than 2-fold difference in abundance). These data are clearly
incompatible with the proposed ipso-attachment of a hydrogen
atom as the sole mechanism for phenyl loss.

Aryl–Aryl Bond Cleavages in FVP
1283
Ј
Ј
Ј
ϩ
ϩ
ϩ
Scheme 6.
S
Scheme 8.
ϩ
ϩ
17
20
ϩ
17
3
20
3
Scheme 9.
ϩ
ϩ
as a major contributor to the cleavage of aryl–aryl bonds
under FVP conditions.
9
18
19
The results from the previous experiment strongly suggest
that the naphthyl unit is not initially released as the 2-naphthyl
radical when the aryl–aryl bond cleaves. This leads to the
hypothesis that one of the two aryl groups may be lost as an
aryne from a biaryl ortho-radical (Scheme 8). The fragment aryl
radical could then scavenge a hydrogen atom to give the parent
arene, but the aryne might simply react with the carbonaceous
material on the walls of the hot tube and never reach the trap as
the corresponding arene.
In the FVP of 9-(2-naphthyl)anthracene (17), only the
naphthyl unit has ortho-hydrogen atoms. Hence, cleavage by
this mechanism would necessarily produce naphthyne and the
9-anthracenyl radical. This could account for the greater abun-
dance of anthracene observed relative to that of naphthalene. To
test the aryne expulsion proposal, we studied the FVP of
9,90-bianthracenyl (20), which has no ortho-hydrogen atoms
on either aryl unit and, according to this mechanism, should
be stable towards aryl–aryl bond cleavage. However, when 9,90-
bianthracenyl (20) was subjected to FVP, we found that it
ruptured, producing anthracene (3), Scheme 9.
Based on NMR integration, anthracene (3) constituted
,18 % (ꢀ4 %) of the pyrolysis product mixture, and recovered
9,90-bianthracenyl (20) was the only other compound detected in
significant amounts (,82 %). Because 9,90-bianthracenyl (20),
which has no ortho-hydrogen atoms on either aryl unit, clea-
ves to produce anthracene (3) with an efficiency comparable
with that observed from FVP of 9-phenylanthracene (1) and
9-(2-naphthyl)anthracene (17), the aryne expulsion mecha-
nism (Scheme 8) seems unlikely to be a major fragmentation
pathway even for those biaryls that do have one or more ortho-
hydrogen atoms.
Based on these findings, none of the four proposed mechan-
isms (Fig. 2, Schemes 4, 6, and 8) satisfactorily explains all of
the FVP data reported here. In such circumstances, one must
immediately acknowledge the possible existence of yet another
mechanistic pathway that has not been considered. We cannot
dismiss this possibility; however, there also remains the possi-
bility that no single, universal mechanism operates to the
exclusion of all others for the cleavage of every aryl–aryl bond
in all biaryls. Even for a single biaryl, cleavage of the aryl–aryl
bond may occur by two (or more) competing simultaneous
mechanistic pathways. This latter scenario is, in fact, completely
plausible at 11008C, where ‘exclusive’ operation of a single
Scheme 7.
The most obvious possible mechanism for aryl–aryl bond
cleavage is direct homolysis of the C–C bond that generates
two aryl radicals. Both radicals would then be expected to
scavenge hydrogen atoms with similar efficiency, giving the
two constituent arenes as the observed products in comparable
amounts (Scheme 6). Although the great strength of aryl–
aryl bonds makes this hypothesis seem implausible at
first (bond dissociation energy of biphenyl: DH ¼ 478.6 ꢀ
6.3 kJ molꢁ1),[39] it must be remembered that the entropic
benefit of generating two species from one is significant at
11008C (1473 K) and can greatly reduce the free energy cost of
bond cleavage (DGz ¼ DHz ꢁ TDSz).
Because benzene is too volatile for us to detect reliably, we
chose to study the FVP of an asymmetrical biaryl in which both
aryl groups are larger than the phenyl ring, so that we could see
the products from both aryl groups. To this end, 9-(2-naphthyl)
anthracene (17)[40] was synthesized by Suzuki coupling of
9-bromoanthracene with naphthalene-2-boronic acid. Subse-
quent FVP of 17 gave mostly recovered starting material
(m/z 304) and a mixture of hexacyclic PAHs (m/z 302) that we
tentatively assigned as the two isomeric products of thermal
cyclodehydrogenation (dibenzo[a,j]fluoranthene (18) and
dibenzo[a,k]fluoranthene (19)) and a significant amount of
anthracene (3) that is one of the products of the aryl–aryl bond
cleavage. The other product of aryl–aryl bond cleavage, naph-
thalene (9), was only obtained in trace amounts (Scheme 7).
From the NMR spectrum of the crude product mixture, the
ratio of the recovered starting material (17) to anthracene (3)
was determined to be 4.4 : 1, indicative of a high extent of aryl–
aryl bond cleavage. However, surprisingly, NMR signals corre-
sponding to naphthalene (9) were not observed. Analysis of the
crude product mixture by GC–MS, which is more sensitive,
revealed a tiny peak above the noise level, corresponding to
naphthalene (9). However, the ratio of anthracene (3) to naph-
thalene (9) in the product mixture from FVP of 9-(2-naphthyl)
anthracene (17) was more than 10 : 1. Assuming that the two aryl
radicals would scavenge hydrogen atoms to give the two
constituent arenes with comparable efficiency, the data from
this experiment do not support the direct homolysis mechanism

1284
E. A. Jackson et al.
mechanism (.99 %), as the favoured pathway, would require
DGz of at least 52.0 kJ molꢁ1 lower than that for the second
lowest energy pathway (DDGz ¼ RTln99).
N
N
ϩ
Ϫ
Cl
Cl
Conclusions
Faced with our inability to formulate any stand-alone mecha-
nism that accommodates all of the experimental facts, we ten-
tatively conclude that aryl–aryl bond cleavages under FVP
conditions involve at least two different mechanistic pathways
and that the relative contributions of the competing pathways
probably vary from one biaryl to the next. Evidence has been
found for the ‘explosion’ of substituted phenyl radicals (Fig. 2).
However, the data indicate that this process represents no more
than a minor pathway for phenyl loss. The comparable ease of
phenyl loss from the central and terminal rings of 1,10-
diphenylanthracene (12) when subjected to FVP does not sup-
port the ipso-attachment of a hydrogen atom as the dominant
mechanism for phenyl loss in the current system (Scheme 4).
FVP of 9-(2-naphthyl)anthracene (17) gives anthracene (3) and
naphthalene (9) in very unequal amounts (3 : 9 . 10 : 1); the fate
of the naphthyl fragment is unknown, but it is almost certainly
not lost as the 2-naphthyl radical. Loss of the naphthyl fragment
as naphthyne (Scheme 8) would be consistent with the low
yield of naphthalene (9), but the anthracene (3) produced by
FVP of 9,90-bianthracenyl (20) cannot be formed by an aryne
expulsion mechanism. Direct homolysis (Scheme 6) and/or a
mechanism initiated by ipso-attachment of a hydrogen atom
(Scheme 4) could account for the formation of anthracene (3)
from FVP of 9,90-bianthracenyl (20).
Pd2(dba)3
Cs2CO3
dioxane
Cl
B(OH)2
Scheme 10.
Br
B(OH)2
Pd[PPh3]4, K2CO3, DMAc
Scheme 11.
1,4-dioxane (50 mL) was added using a syringe. The reaction
flask was lowered into the 808C bath, and the reaction mixture
was stirred at that temperature. After 2 days, GC–MS analysis
showed that the reaction was complete. The reaction mixture
was removed from the heating bath, allowed to cool, filtered
through a plug of silica, washed thrice with water, dried with
magnesium sulfate, and filtered. The solvent was removed by
rotary evaporation to give a crude oil that solidified overnight.
This material was redissolved in dichloromethane and crystal-
lized from dichloromethane and hexanes to give 440 mg (59 %)
of material that was determined to be suitable for FVP based on
NMR spectroscopy analysis (see Supplementary Material). The
NMR spectrum reveals the presence of residual water, dichlor-
omethane, and dioxane. No NMR spectra have been reported for
this compound in earlier publications reporting on the synthesis
of the compound.[37] Hence, these are included herein. dH
(CDCl3, 500 MHz) 8.52 (1H, s), 7.91 (1H, d, J 8.0), 7.69 (1H, t,
J 5.0), 7.65–7.36 (14H, m), 7.34 (1H, ddd, J 7.5, 6.5, 1.0). dC
(CDCl3, 125 MHz) 141.26, 140.42, 139.21, 137.36, 131.45,
131.40, 130.66, 130.39, 130.16, 130.14, 128.92, 128.55, 127.61,
127.53, 126.83, 126.73, 126.15, 125.70, 125.44, 125.17, 124.92.
Experimental
All commercially available reagents, catalysts, and chemicals
were used as received, without further purification. All solvents
were of reagent grade. Anhydrous dimethylacetamide (DMAc)
was obtained from a solvent purification system constructed by
Contour Glass. The 2-bromobiphenyl and 9-phenylanthracene
were commercially available and used in FVP as received. The
9,90-bianthracenyl was prepared according to a published pro-
cedure.[41] Proton and carbon NMR spectra were obtained using
a Varian 500 MHz NMR spectrometer. Chemical shifts are
reported in ppm downfield from tetramethylsilane with either
d-chloroform (dH 7.26 ppm, dC 77.0 ppm) or d2-dichloromethane
(dH 5.30 ppm) as the standard reference. Thin layer chroma-
tography was performed on Analtech Silica G TLC plates. For
preparative column chromatography, 40–63 mm silica gel was
used. Mass analyses were performed using a Thermo Electron
Corporation Finnigan Trace GC Ultra gas chromatograph unit
connected to a Thermo Electron Corporation Finnigan Trace
DSQ mass spectrometer.
9-(2-Naphthyl)anthracene (17)[39] (Scheme 11)
A silicon oil bath was heated to 1008C. To a clean dry 100-mL
round bottom flask were added a magnetic stirring bar, 500 mg
of 9-bromoanthracene (1.94 mmol, 257.13 g molꢁ1), 334.4 mg
of naphthalen-2-ylboronic acid (1.94 mmol, 1 equiv., 171.99 g
molꢁ1), 2.69 g of potassium carbonate (19.4 mmol, 10 equiv.,
138.21 g molꢁ1), and 449.4 mg of tetrakis(triphenylphosphine)
palladium (Pd(PPh3)4, 0.389 mmol, 0.2 equiv., 1155.56 g
molꢁ1). The flask was sealed with a rubber stopper, and the
system was evacuated and filled with nitrogen three times.
Anhydrous dimethylacetamide (48 mL) was added using a
syringe. The reaction flask was lowered into the 1008C bath, and
the reaction mixture was stirred at that temperature. After 1 day,
the reaction mixture was removed from the heating bath and
allowed to cool. Dichloromethane was added; the mixture was
filtered through a plug of silica, washed six times with water,
dried with magnesium sulfate, and filtered. Silica was added,
and the product mixture underwent rotary evaporation.
The product was purified by column chromatography using
1,10-Diphenylanthracene (12)[37] (Scheme 10)
A silicon oil bath was heated to 808C. To a clean dry 250-mL
round bottom flask containing 555 mg of 1,10-dichloroan-
thracene[42] (2.25 mmol, 247.12 g molꢁ1) were added a magnetic
stirring bar, 2.19 g of phenylboronic acid (18.0 mmol, 8 equiv.,
121.93 g molꢁ1), 11.7 g of caesium carbonate (36.0 mmol,
16 equiv., 325.82 g molꢁ1), 229 mg of 1,3-bis(2,6-diisopro-
pylphenyl)imidazolium chloride (0.539 mmol, 0.24 equiv.,
425.05 g molꢁ1), and 247 mg of tris(dibenzylideneacetone)
dipalladium (Pd2(dba)3, 0.270 mmol, 0.12 equiv., 915.72 g molꢁ1).
The flask was sealed with a rubber stopper, and the system was
evacuated and filled with nitrogen three times. Anhydrous

Aryl–Aryl Bond Cleavages in FVP
1285
Vacuum gauge
Nitrogen
flow
Pyrolysis
oven
Sublimation
oven
Liquid nitrogen traps
Vacuum
WARNING: Explosive liquid oxygen can condense in a trap that is cooled by liquid nitrogen while it is open to
the atmosphere for any significant length of time.
Fig. 4. Apparatus for flash vacuum pyrolysis.
silica and hexanes. The product fractions were precipitated
from dichloromethane and hexanes and filtered to give 274 mg
(46 %) of material that was determined to be suitable for FVP
based on NMR spectroscopy analysis (see Supplementary
flow is turned on. For our purposes, the additional oven was
only used for insulation. Once the pyrolysis oven reaches a
temperature of 11008C, the boat containing the starting material
is incrementally moved towards the pyrolysis oven (using the
pair of stirring bar magnets) at 2–3 cm every 6 min to achieve
complete and even sublimation of the starting material. For
molecules described herein, the molecular weights are suffi-
ciently low to achieve sublimation either at room temperature or
in the warmer regions of the first oven as the boat approaches
the aperture of the pyrolysis oven. No additional heating of the
sublimation oven is required. In cases beyond the scope of this
work, involving starting materials of higher molecular weights,
the sublimation oven temperature can be raised in conjunction
with or instead of movement of the boat towards the pyrolysis
oven. The sublimed material is transported uni-directionally by
the high purity nitrogen flow, and the pyrolysis materials are
sequestered in the trap cooled by liquid nitrogen.
1
Material). The H and 13C NMR spectra match those reported
in the literature.[39]
General Experimental Procedure for Flash
Vacuum Pyrolysis
An apparatus was constructed, as shown in Fig. 4, having a
quartz tube running through a sublimation oven and a pyrolysis
oven. Both ovens are equipped with a hinge so that the top half
of the heating element can be raised and the quartz tube can be
inserted or removed. During operation, the system is isolated
from the atmosphere and placed under a high vacuum, typically
less than 1 hPa. Nitrogen is delivered to the side closest to the
sublimation oven via a gas chromatography capillary (25 cm
long and 0.25 mm inner diameter) from a nitrogen gas reservoir
that is maintained at 103 hPa, by connection to a gas cylinder and
a bubbler. The length and diameter of the GC capillary repro-
ducibly determine the flow rate of nitrogen based on the Hagen–
Poiseuille equation. The stream of nitrogen flows through a
plug of glass wool placed directly inside the quartz pyrolysis
tube before closing the system. This creates turbulence so that
the nitrogen flow is uniform beyond the glass wool. Two traps
are situated at the end of the quartz pyrolysis tube, downstream
from the pyrolysis oven. During operation, the first trap is cooled
by a Dewar flask filled with liquid nitrogen so that the pyrolysis
products will condense and be unable to reach the vacuum
pump. For our purposes, only the first liquid nitrogen trap is used
because the amount of material submitted to pyrolysis is small
(200 mg or less). It is recommended to use both liquid nitrogen
traps for larger pyrolysis samples. A vacuum gauge is connected
to the system to monitor the vacuum strength. A second valve
provides a means of opening the system to the atmosphere, once
pyrolysis is complete. A specialised glass fused metal accordi-
on-like section connects the apparatus to the vacuum pump to
reduce vacuum pump vibrations in the system. Before closing
the system, the starting material is weighed out on a quartz boat
that is then inserted to the open end of the quartz tube on the
sublimation oven side. A strong magnetic stirring bar is inserted
in the quartz tube upstream of the quartz boat. This stirring bar,
guided by an external magnet, is used to push the boat (before
and after closing the system) so that it is surrounded by the
opening of the sublimation oven. The aforementioned glass
wool is then inserted before the system is closed and the nitrogen
Procedure 1
Generally, pyrolysis is carried out in the following order. The
boat containing the starting material is inserted into the quartz
tube, the system is closed, nitrogen flow is initiated, the vacuum
pump is turned on, and the pyrolysis oven is set to 11008C. Once
the set temperature is reached, liquid nitrogen is added to the
Dewar flask surrounding the first trap, and the starting material
is moved gradually towards the pyrolysis oven. Following
sublimation of all the starting material in the pyrolysis oven,
the pyrolysis oven is set to room temperature and allowed to cool
to 6008C or lower. The liquid nitrogen Dewar flask is then
removed, and the system is rapidly opened to air, while concur-
rently turning off the vacuum pump. The cold trap is removed
from the apparatus, and the contents are dissolved in the solvent
and sealed in a closed system within a few minutes.
Procedure 2
Some starting materials have such low molecular weights
that they sublime at room temperature under high vacuum
conditions. In these cases, it is not possible to use the pyrolysis
method as described in Procedure 1. The following alternative
sequence is used. The pyrolysis oven is set to 11008C and
allowed to reach that temperature in an open air atmosphere
with everything beyond the pyrolysis oven (e.g. cold traps,
vacuum) already fully connected. Once the set temperature is
achieved, the boat containing the starting material is inserted
into the quartz tube, the front end of the system is sealed, and
nitrogen flow is initiated. Liquid nitrogen is added to the Dewar

1286
E. A. Jackson et al.
Table 1. Details of the flash vacuum pyrolyses reported
Quantity
[mg]
Procedure no.
Vacuum pressure
[hPa]
Quartz boat movement
speed [cm minꢁ1
]
9-Phenylanthracene (1)
2-Bromobiphenyl (5)
20
200
200
200
20
1
4
4
1
3
1
880
870
890
870
910
870
0.4
0
Biphenyl (8)
0
1,10-Diphenylanthracene (12)
9-(2-Naphthyl)anthracene (17)
9,90-Bianthracenyl (20)
0.4
0.4
0.4
20
flask surrounding the first trap, and the vacuum pump is turned
on after no more than 30 s of cold trap equilibration to prevent
liquid oxygen from condensing. Following sublimation of all
the starting materials in the pyrolysis oven, the same subsequent
steps to those detailed in Procedure 1 are used.
[4] R. F. C. Brown, F. W. Eastwood, Synlett 1993, 9. doi:10.1055/
S-1993-22330
[5] R. F. C. Brown, Eur. J. Org. Chem. 1999, 3211. doi:10.1002/(SICI)
1099-0690(199912)1999:12,3211::AID-EJOC3211.3.0.CO;2-F
[6] C. Wentrup, Chimia (Aarau) 1977, 31, 258.
[7] C. Wentrup, Lect. Heterocycl. Chem. 1984, 7, 91.
[8] H. McNab, Contemp. Org. Synth. 1996, 3, 373. doi:10.1039/
CO9960300373
Procedure 3
[9] H. McNab, Aldrichim Acta 2004, 37, 19.
[10] L. T. Scott, Accounts Chem. Res. 1982, 15, 52. doi:10.1021/
AR00074A004
In some cases, the products include relatively volatile species
such as naphthalene and phenyl acetylene. In these cases, fol-
lowing sublimation of all the starting material in the pyrolysis
oven (according to the setup sequence in Procedure 1), the
system is rapidly opened to the atmosphere (pyrolysis oven
still at 11008C) while concurrently turning off the vacuum
pump. The liquid nitrogen Dewar flask is removed very quickly
thereafter to prevent liquid oxygen from condensing (30 s or
less). The cold trap is removed, and the contents are dissolved
in solvent within a few minutes. The pyrolysis oven is set to
room temperature and allowed to cool shortly thereafter. This
order of steps ensures that none of the relatively volatile species
is lost because the cold trap is not allowed to warm up until
after the system is at 103 hPa.
[11] L. T. Scott, Chem. Aust. 1987, 54, 298.
[12] L. T. Scott, Pure Appl. Chem. 1996, 68, 291.
[13] A. Necula, L. T. Scott, J. Anal. Appl. Pyrolysis 2000, 54, 65.
doi:10.1016/S0165-2370(99)00085-6
[14] L. T. Scott, M. M. Hashemi, D. T. Meyer, H. B. Warren, J. Am. Chem.
Soc. 1991, 113, 7082. doi:10.1021/JA00018A082
[15] L. T. Scott, M. S. Bratcher, S. Hagen, J. Am. Chem. Soc. 1996, 118,
8743. doi:10.1021/JA9621511
[16] S. Hagen, M. S. Bratcher, M. S. Erickson, G. Zimmermann, L. T. Scott,
Angew. Chem. Int. Ed. Engl. 1997, 36, 406. doi:10.1002/
ANIE.199704061
[17] L. T. Scott, P.-C. Cheng, M. M. Hashemi, M. S. Bratcher, D. T. Meyer,
H. B. Warren, J. Am. Chem. Soc. 1997, 119, 10963. doi:10.1021/
JA972019G
[18] V. M. Tsefrikas, L. T. Scott, Chem. Rev. 2006, 106, 4868. doi:10.1021/
CR050553Y
Procedure 4
In some cases, materials are of a sufficiently low molecular
weight to sublime at room temperature under high vacuum
conditions and the products include relatively volatile species
such as naphthalene and phenyl acetylene. In this situation, a
logical hybrid methodology of Procedures 2 and 3 is employed.
Table 1 summarizes the details of the FVP reported here.
[19] L. T. Scott, E. A. Jackson, Q. Zhang, B. D. Steinberg, M. Bancu,
B. Li, J. Am. Chem. Soc. 2012, 134, 107. doi:10.1021/JA209461G
[20] L. T. Scott, M. M. Boorum, B. J. McMahon, S. Hagen, J. Mack,
J. Blank, H. Wegner, A. de Meijere, Science 2002, 295, 1500.
doi:10.1126/SCIENCE.1068427
[21] L. T. Scott, Angew. Chem. Int. Ed. 2004, 43, 4994. doi:10.1002/
ANIE.200400661
[22] X. Xue, L. T. Scott, Org. Lett. 2007, 9, 3937. doi:10.1021/OL7015516
[23] X. Xue, Mechanistic Studies on the Thermal Cyclodehydrogenations
of Polycyclic Aromatic Hydrocarbons 2008, Ph.D. thesis, Boston
College, Chestnut Hill, Massachusetts, USA.
Supplementary Material
1H and 13C NMR spectra for 1,10-diphenylanthracene (12); 1H
spectra for 9-(2-naphthyl)anthracene (17) and for FVP products
that were positively identified by 1H NMR spectroscopy: 1, 2, 3,
9, 14, 20; and 1H NMR and GC–MS data from the FVP of 1, 5, 8,
12, 17, and 20 are available on the Journal’s website.
[24] L. T. Scott, N. H. Roelofs, J. Am. Chem. Soc. 1987, 109, 5461.
doi:10.1021/JA00252A025
[25] L. T. Scott, N. H. Roelofs, Tetrahedron Lett. 1988, 29, 6857.
doi:10.1016/S0040-4039(00)88459-2
[26] L. T. Scott, M. M. Hashemi, T. H. Schultz, M. B. Wallace, J. Am.
Chem. Soc. 1991, 113, 9692. doi:10.1021/JA00025A055
[27] R. D. Smith, Combust. Flame 1979, 35, 179. doi:10.1016/0010-2180
(79)90021-X
[28] A. Laskin, A. Lifshitz, in Proceedings of the 26th International
Symposium on Combustion 1996, pp. 669–675 (The Combustion
Institute: Pittsburgh, PA).
Acknowledgements
The authors thank the Department of Energy for financial support of this
work and the Division of Organic Chemistry of the American Chemical
Society for a Roche-funded graduate fellowship to H.Y.C.
References
[29] B. Negru, S. J. Goncher, A. L. Brunsvold, G. M. P. Just, D. Park,
D. M. Neumark, J. Chem. Phys. 2010, 133, 074302. doi:10.1063/1.
3473743
[1] R. F. C. Brown, Pyrolytic Methods in Organic Chemistry: Application
of Flow and Flash Vacuum Pyrolytic Techniques 1980 (Academic
Press: New York, NY).
[30] L. K. Madden, L. V. Moskaleva, S. Kristyan, M. C. Lin, J. Phys.
Chem. A 1997, 101, 6790. doi:10.1021/JP970723D
[31] H. Wang, A. Laskin, N. W. Moriarty, M. Frenklach, Proc. Combust.
Inst. 2000, 28, 1545. doi:10.1016/S0082-0784(00)80552-4
[2] R. F. C. Brown, Recl. Trav. Chim. Pays-Bas 1988, 107, 655.
doi:10.1002/RECL.19881071202
[3] R. F. C. Brown, Pure Appl. Chem. 1990, 62, 1981. doi:10.1351/
PAC199062101981

Aryl–Aryl Bond Cleavages in FVP
1287
1
[32] M. A. Brooks, L. T. Scott, J. Am. Chem. Soc. 1999, 121, 5444.
doi:10.1021/JA984472D
[33] M. Ladacki, M. Szwarc, J. Chem. Phys. 1952, 20, 1814. doi:10.1063/
1.1700308
[38] The H NMR spectral data for rubicene were found in V. Sachweh,
H. Langhals, Chem. Ber. 1990, 123, 1981. doi:10.1002/
CBER.19901231009
[39] Y.-R. Luo, Handbook of Bond Dissociation Energies in Organic
[34] M. Szwarc, D. Williams, P. Roy. Soc. A 1953, 219, 353. doi:10.1098/
RSPA.1953.0152
Compounds 2002 (CRC Press: Boca Raton, FL).
[40] For a recent synthesis of 9-(2-naphthyl)anthracene (17), see M. Xu,
X. Li, Z. Sun, T. Tu, Chem. Commun. 2013, 49, 11539. doi:10.1039/
C3CC46663A
[35] A. Necula, L. T. Scott, Polycyclic Aromat. Compd. 2010, 30, 260.
doi:10.1080/10406638.2010.524851
[36] R. F. C. Brown, N. Choi, K. J. Coulston, F. W. Eastwood,
U. E. Wiersum, L. W. Jenneskens, Tetrahedron Lett. 1994, 35, 4405.
doi:10.1016/S0040-4039(00)73368-5
[37] For a previous synthesis of 1,10-diphenylanthracene (12), see
S. C. Dickerman, D. De Souza, P. Wolf, J. Org. Chem. 1965,
30, 1981. doi:10.1021/JO01017A065
[41] F. Bell, D. H. Waring, J. Chem. Soc. 1949, 267. doi:10.1039/
JR9490000267
[42] I. Tanimoto, K. Kushioka, T. Kitagawa, K. Maruyama, Bull. Chem.
Soc. Jpn. 1979, 52, 3586. doi:10.1246/BCSJ.52.3586