1,2-shifts of hydrogen atoms in aryl radicals

Article abstract of DOI:10.1021/ja984472d

An energy barrier on the order of 60 kcal/mol is predicted for the 1,2- shift of hydrogen atoms in aryl radicals. Such rearrangements are, therefore, not expected to occur under ordinary laboratory conditions, but they should be prevalent in the aryl radicals formed during combustion, flash vacuum pyrolysis, and other high-temperature gas-phase processes. As a demonstration of this rearrangement, the 2-benzo[c]phenanthryl radical (1) was generated by flash vacuum pyrolysis of the corresponding aryl bromide (9). A 1,2-shift of hydrogen out of the sterically congested cove region of 1, followed by cyclization and rearomatization of the resulting radical (Scheme 1), is proposed to explain the observation of benzo[ghi]fluoranthene (4) as the dominant monomeric product formed. Under the same conditions, [ 1,3,4,5- ²H₄]-2-bromobenzo[c]phenanthrene (13) gives [ 1,2,3,4-²H₄]- benzo[ghi]fluoranthene (15) as the dominant monomeric product (Scheme 6), in accord with the expectation of a deuterium atom 1,2-shift.

Full text of DOI:10.1021/ja984472d

5444
J. Am. Chem. Soc. 1999, 121, 5444-5449
1,2-Shifts of Hydrogen Atoms in Aryl Radicals
Michele A. Brooks and Lawrence T. Scott*
Contribution from the Department of Chemistry, Merkert Chemistry Center, Boston College,
Chestnut Hill, Massachusetts 02467-3860
ReceiVed December 29, 1998
Abstract: An energy barrier on the order of 60 kcal/mol is predicted for the 1,2-shift of hydrogen atoms in
aryl radicals. Such rearrangements are, therefore, not expected to occur under ordinary laboratory conditions,
but they should be prevalent in the aryl radicals formed during combustion, flash vacuum pyrolysis, and other
high-temperature gas-phase processes. As a demonstration of this rearrangement, the 2-benzo[c]phenanthryl
radical (1) was generated by flash vacuum pyrolysis of the corresponding aryl bromide (9). A 1,2-shift of
hydrogen out of the sterically congested cove region of 1, followed by cyclization and rearomatization of the
resulting radical (Scheme 1), is proposed to explain the observation of benzo[ghi]fluoranthene (4) as the dominant
monomeric product formed. Under the same conditions, [1,3,4,5-2H4]-2-bromobenzo[c]phenanthrene (13) gives
[1,2,3,4-2H4]-benzo[ghi]fluoranthene (15) as the dominant monomeric product (Scheme 6), in accord with the
expectation of a deuterium atom 1,2-shift.
Every chemist has heard of the 1,2-shifts of hydrogen atoms
allowed”, whereas the same rearrangements in carbanions would
involve antiaromatic transition states, which makes them
“symmetry forbidden”. The transition states for 1,2-shift rear-
rangements in radicals are neither strongly stabilized by
aromaticity nor strongly destabilized by antiaromaticity; there-
fore, the activation energies for such process should be large,
but not necessarily prohibitively so. A simple computational
analysis (next section) suggests that aryl radicals should be good
candidates to exhibit 1,2-shift rearrangements of hydrogen
atoms.
and alkyl groups (Wagner-Meerwein rearrangements) that
pervade the chemistry of carbocations,1-3 but few know much
about the analogous rearrangements of organic free radicals4-6
and carbanions.7-10 Examples of the latter processes, in fact,
are comparatively rare, and in those cases where they do appear
to occur, nonconcerted mechanisms frequently prevail.11-14
A satisfying explanation for the special proclivity of car-
bocations for 1,2-shift rearrangements, and for the reluctance
of radicals and carbanions to behave similarly, can be found in
the Woodward-Hoffmann rules of orbital symmetry conserva-
tion. Such rearrangements are properly viewed as pericyclic
transformations (i.e., [1,2] sigmatropic reactions) that proceed
by way of three-centered transition states involving the cyclic
delocalization of two electrons for cations, three electrons for
radicals, and four electrons for carbanions. Thus, 1,2-shift
rearrangements in cations involve aromatic transition states (4N
+ 2 electrons in the cycle), which makes them “symmetry
Computational Results
For calculations involving the energies of aryl radicals, density
functional theory has proven superior to conventional unre-
stricted Hartree-Fock theory and gives results that agree quite
well with experiment.15,16 BLYP/6-311G** calculations (cor-
rected for zero-point vibrational energy, ZPE, differences), for
example, give a value of 106.9 kcal/mol for the homolytic C-H
bond dissociation energy of benzene,16 which compares favor-
ably with the recently reported experimental values of 109.8 (
0.8 kcal/mol17 and 112.0 ( 0.6 kcal/mol.18 We obtained a value
of 109.4 kcal/mol for the homolytic C-H bond dissociation
energy of benzene using BP/DN** density functional calcula-
tions19 (ZPE-corrected) and have performed the calculations
below at this level of theory.
(1) For a review on 1,2-shifts in carbocations, see: Vogel, P. Carbocation
Chemistry; Elsevier: New York, 1985; pp 323-372.
(2) See also: Shubin, V. G. Top. Curr. Chem. 1984, 116/117, 267-
341.
(3) See also: Kirmse, W. Top. Curr. Chem. 1979, 80, 89-124.
(4) For a review on 1,2-shifts in radicals, see: Wilt, J. W. Free Radicals;
Kochi, J. K., Ed.; Wiley: New York, 1973; Vol. I., p 333-501.
(5) See also: Nonhebel; Walton Free-Radical Chemistry; Cambridge
University Press: London, 1974; pp 498-552.
(6) See also: Friedlina, R. K. AdV. Free Radical Chem. 1965, 1, 211.
(7) For a review on 1,2-shifts in carbanions, see: Zimmerman, H. E.
Molecular Rearrangements, Part 1; de Mayo, P., Ed.; Wiley: New York,
1963; p 345.
(8) See also: Cram, D. J. Fundamentals of Carbanion Chemistry;
Academic Press: New York, 1965; Chapter VI.
In vinyl radical, the activation energy calculated for a 1,2-
shift of hydrogen (43.5 kcal/mol) actually exceeds the energy
(15) Barone, V.; Adamo, C.; Russo, N. Int. J. Quantum Chem. 1994,
52, 963-971.
(16) Cioslowski, J.; Liu, G.; Martinov, M.; Piskorz, P.; Moncrieff, D. J.
Am. Chem. Soc. 1996, 118, 5261-5264.
(9) See also: Buncel, E. Carbanions: Mechanistic and Isotopic Aspects;
Elsevier: Amsterdam, 1975.
(10) See also: Hunter, D. H.; Stothers, J. B. Rearrangements in Ground
and Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980;
Vol. 1, p 392.
(17) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994, 98,
2744-2765.
(18) Davico, G. E.; Bierbaum, V. M.; DePuy, C. H.; Ellison, G. B.;
Squires, R. R. J. Am. Chem. Soc. 1995, 117, 2590-2599.
(19) BP/DN** calculations were performed using the density functional
program implemented in the Spartan 5.0 software package (Wavefunction,
Inc., Irvine, CA 92612). Vibrational analyses at the BP/DN** level of theory
have confirmed that the transition-state species reported here are character-
ized by one and only one imaginary vibrational frequency and that species
corresponding to energy minima have no imaginary vibrational frequencies.
(11) For a review on the Stevens rearrangement, see: Lepley, A. R.;
Giumanini, A. G. Mech. Mol. Migr. 1971, 3, 297-440.
(12) See also: Pine, S. H. Org. React. 1970, 18, 403-464.
(13) For a review on the Wittig rearrangement, see: Johnstone, R. A.
W. Mech. Mol. Migr. 1969, 2, 249-266.
(14) For reviews on neophyl rearrangements, see refs 4-6.
10.1021/ja984472d CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/27/1999

1,2-Shifts of H Atoms in Aryl Radicals
J. Am. Chem. Soc., Vol. 121, No. 23, 1999 5445
Scheme 1
Figure 1. Energy required for â-scission and for 1,2-shift of hydrogen
in the vinyl radical (top) and phenyl radical (bottom). All calculations
performed at the BP/DN** level of theory19 and corrected for zero-
point vibrational energy (ZPE) differences.
Scheme 2
required for simple â-scission to give acetylene and a hydrogen
atom (40.9 kcal/mol) at the BP/DN** level of theory (both
values ZPE-corrected). Wang et al. have reached the same
conclusion using electron correlated ab initio computational
methods.20 In phenyl radical, on the other hand, â-scission
should be strongly disfavored, owing to the high strain energy
of the benzyne that would be generated. Indeed, we find the
calculated activation energy for a 1,2-shift of hydrogen in phenyl
radical (58.4 kcal/mol) to be considerably less than the energy
required for â-scission (82.2 kcal/mol) at the BP/DN** level
of theory (both values ZPE-corrected).21,22 Figure 1 summarizes
these computational results.
Clearly, hydrogen atom 1,2-shifts in aryl radicals must be
far less favorable energetically than the familiar [1,2]-sigma-
tropic rearrangements of carbocations, and they certainly would
not be expected to occur spontaneously under ordinary labora-
tory conditions. We are unaware of any prior examples, other
than our own, and herein offer what we believe to be the first
closely examined case.23-26
would be made available to the newly created aryl radical only
after the hydrogen has shifted.
The 2-benzo[c]phenanthryl radical (1) offers both of these
features. To begin with, a 1,2-shift of hydrogen out of the cove
region to the adjacent radical center ought to be accompanied
by a significant relief of steric congestion. We calculate an
exothermicity of 5.3 kcal/mol for this rearrangement at the BP/
DN** level of theory (ZPE-corrected).27 Equally importantly,
following the 1,2-shift of hydrogen, the new aryl radical
suddenly gains access to an irreversible cyclization reaction (2
f 3 f 4, Scheme 1).
To generate aryl radical 1 under conditions that would give
it enough energy to surmount a barrier of nearly 60 kcal/mol,
we decided to subject the corresponding aryl bromide (9) to
flash vacuum pyrolysis. The C-Br homolytic bond dissociation
energy of bromobenzene is 82.7 ( 1.1 kcal/mol,17 and thus any
temperature that is sufficient to rupture the C-Br bond in 9
should provide more than enough energy to promote the
subsequent 1,2-shift of hydrogen, assuming our calculations are
correct. The pyrolysis of aryl bromides has been used for many
years as a classic means of generating aryl radicals in the gas
phase.28-30
System Design
The rather substantial activation energy predicted by the above
calculations for a 1,2-shift of hydrogen in the phenyl radical
persuaded us that it would be prudent to design a system wherein
(a) the 1,2-shift of hydrogen could benefit from some thermo-
dynamic driving force and (b) a diagnostic reaction channel
(20) Coupled cluster ab initio calculations corrected for electron cor-
relation, using Dunning’s triple-ú polarization basis set (CCSD(T)/TZ2P),
as implemented in the ACES II program, give an activation energy of “at
least 47 kcal/mol” for the 1,2-shift of hydrogen in vinyl radical and a
dissociation threshold of 40.7 kcal/mol for the â-scission of vinyl radical
to acetylene and a hydrogen atom: Wang, J.-H.; Chang, H.-C.; Chen, Y.-
T. Chem. Phys. 1996, 206, 43-56.
(21) The only prior calculations we have found for the 1,2-shift of
hydrogen in phenyl radical, although now quite out of date, also predicted
a high energy barrier: Burmistrov, V. N.; Khudyakov, I. V.; Christensen,
A. IzV. Akad. Nauk, Ser. Khim. 1979, 460-2 (Chem. Abstr. No. 90:186104).
(22) See also van der Hart, W. J. J. Am. Soc. Mass Spectrom. 1995, 6,
513-5.
(23) Evidence for the migration of hydrogen atoms in dehydrobenzenes
(aryl diradicals) at elevated temperatures has been seen in the beautiful
work of R. F. C. Brown et al.: Brown, R. F. C.; Eastwood, F. W. Pure
Appl. Chem. 1996, 68, 261-266.
(27) Cioslowski et al. have calculated an exothermicity of 6.9 kcal/mol
for the difference in energy between 1 and 2 at the BLYP/6-311G** level
of theory (ZPE corrected).16
(28) Ladaki, M.; Szwarc, M. J. Chem. Phys. 1952, 20, 1814.
(29) Ladaki, M.; Szwarc, M. Proc. R. Soc. London, Ser. A 1953, A219,
341.
(30) Benson, S. W.; O’Neal, H. E. Kinetic Data on Gas-Phase Unimo-
lecular Reactions; NBS-NSRDS 21, U.S. Government Printing Office:
Washington, DC, 1970.
(24) See also: Banciu, M. D.; Brown, R. F. C.; Coulston, K. J.; Eastwood,
F. W.; Jurss, C.; Mavropoulos, I.; Stanescu, M.; Wiersum, U. E. Aust. J.
Chem. 1996, 49, 965-976.
(25) See also: Bapat, J. B.; Brown, R. F. C.; Bulmer, G. H.; Childs, T.;
Coulston, K. J.; Eastwood, F. W.; Taylor, D. K. Aust. J. Chem. 1997, 50,
1159-1182.
(26) See also: Brown, R. F. C.; Coulston, K. J.; Eastwood, F. W. Aust.
J. Chem. 1997, 50, 1183-1189.

5446 J. Am. Chem. Soc., Vol. 121, No. 23, 1999
Brooks and Scott
Scheme 3
Scheme 4
Before describing the experimental aspects of the present
project, it is relevant to note a prior discovery in our lab.
Heartened only by a hunch that the chemistry depicted in
Scheme 1 should work, but with no experimental precedent,
we developed an efficient synthesis of tribromobenzo[c]naphtho-
[2,1-p]chrysene (5) and subjected it to flash vacuum pyrolysis
at 1050 °C. As we had hoped, the C-Br bonds were broken,
and that initiated what appears to be a 3-fold version of the
cyclization reaction proposed in Scheme 1, giving us a marvel-
ous, bowl-shaped, C3-symmetric C30H12 “hemibuckminster-
fullerene” (Scheme 2).31 A desire to understand more thoroughly
the details of this remarkable transformation was, in fact, our
motivation for the present project.
Removal of the minor quantity of unwanted ortho-isomer turned
out to be easier at a later stage in the reaction sequence, and
thus the impure p-bromotoluene-d7 (10) was treated directly with
N-bromosuccinimide in refluxing carbon tetrachloride to give
p-bromobenzyl bromide-d6 (11), contaminated by a small
quantity of the ortho-isomer. Oxidation of this mixture with the
sodium salt of 2-nitropropane in ethanol35-38 gave the desired
p-bromobenzaldehyde-d5 (12), which was separated from the
ortho-isomer by chromatography. Wittig coupling of 12 with
the ylide derived from phosphonium salt 7 followed by
photolysis as before gave [1,3,4,5-2H4]-2-bromobenzo[c]phenan-
threne (13).
Syntheses of Aryl Radical Precursors
Pyrolysis Results and Discussion
Although the requisite 2-bromobenzo[c]phenanthrene (9) had
been synthesized before,32 we developed the convenient,
alternative, two-pot synthesis illustrated in Scheme 3. Thus, a
solution of the commercially available 2-bromomethylnaphtha-
lene (6) and triphenylphosphine in dimethylformamide was
heated to reflux to form the phosphonium salt (7). This
intermediate was not isolated but was used in situ for a Wittig
reaction by first cooling the solution to room temperature and
then adding p-bromobenzaldehyde and lithium ethoxide. Chro-
matographic separation of the cis and trans product isomers,
although possible, was found to be unnecessary, since the
isomers interconvert photochemically in the next step. Photo-
cyclization of the mixture of stilbenes (8) in the presence of
propylene oxide33 gave 2-bromobenzo[c]phenanthrene (9).32
As described in the next section, pyrolysis of 9 does indeed
give benzo[ghi]fluoranthene (4), the product expected from the
aryl radical generated by a 1,2-shift of hydrogen out of the cove
region (Scheme 1). We therefore prepared an isotopically labeled
derivative of 9 to see whether a deuterium in the cove region
would migrate to the site of the initial radical. Synthesis of the
deuterated analogue 13 was achieved by replacing p-bromobenz-
aldehyde in Scheme 3 with p-bromobenzaldehyde-d5 (12,
Scheme 4). The extra (“spectator”) deuterium atoms at positions
3, 4, and 5 in 13 that result from this synthetic strategy are not
involved in the hydrogen shift and thus do not affect the results.
As summarized in Scheme 4, the synthetic route to 13 began
with bromination of toluene-d8 in trifluoroacetic acid-d1.34
Flash vacuum pyrolyses were performed on 50-mg samples
of 2-bromobenzo[c]phenanthrene (9) at temperatures ranging
from 950 to 1100 °C, with a controlled influx of nitrogen to
serve as a carrier gas (final pressure ) 2.0-2.5 mmHg). The
product mixture consisted of unchanged 9, benzo[ghi]fluoran-
thene (4),39-41 and benzo[c]phenanthrene (14)32,42-45 (Scheme
5). The composition of the product mixture was determined by
1H NMR analysis and was found to vary with the pyrolysis
temperature (Table 1).
The presence of benzo[c]phenanthrene (14) as a minor
product in the pyrolysates reflects the well-known fact46,47 that
aryl radicals (1 and 2 in this reaction) are capable of abstracting
hydrogen atoms from other organic species in the gas phase or
on the walls of the pyrolysis tube. At 950 °C the product formed
by what appears to involve a hydrogen shift (4) outweighs the
product formed by hydrogen abstraction (14) by a factor of 25:6
(35) Modification of the method by Hass, H. B.; Bender, M. L. J. Am.
Chem. Soc. 1949, 71, 1767-1769.
(36) See also: Hass, H. B.; Bender, M. L. In Organic Syntheses; Rabjohn,
N., Ed.; 1963, Collect. Vol. 4, pp 932-934.
(37) See also: Yamamoto, K.; Ikeda, T.; Kitsuki, T.; Okamoto, Y.;
Chikamatsu, H.; Nakazaki, M. J. Chem Soc., Perkin Trans. 1 1990, 271-
276.
(38) See also: Yamamoto, K.; Sonobe, H.; Matsubara, H.; Sato, M.;
Okamoto, S.; Kitaura, K. Angew. Chem., Int. Ed. Engl. 1996, 35, 69-70.
(39) Campbell, R.; Reid, D. H. J. Chem. Soc. 1952, 3281-3284.
(40) Crombie, D. A.; Shaw, S. J. Chem. Soc. 1969, 2489-2490.
(41) Studt, P.; Win, T. Liebigs Ann. Chem. 1983, 519.
(42) Cook, J. W. J. Chem. Soc. 1931, 2525-2529.
(43) Szmuszkovicz, J.; Modest, E. J. J. Am. Chem. Soc. 1948, 70, 2542-
2543.
(44) Defay, N.; Zimmermann, D.; Martin, R. H. Tetrahedron Lett. 1971,
21, 1871-1874.
(45) Pataki, J.; Harvey, R. G. J. Org. Chem. 1982, 47, 20-22.
(46) Fahr, A.; Stein, S. E. J. Phys. Chem. 1988, 92, 4951-4955.
(47) Chen, R. H.; Kafafi, S. A.; Stein, S. E. J. Am. Chem. Soc. 1989,
111, 1418-1423.
(31) Hagen, S.; Bratcher, M. S.; Erickson, M. S.; Zimmermann, G.; Scott,
L. T. Angew. Chem., Int. Ed. Engl. 1997, 36, 406-408.
(32) Martin, R. H.; Moriau, J.; Defay, N. Tetrahedron 1974, 30, 179-
185.
(33) Liu, L.; Yang, B.; Katz, T. J.; Poindexter, M. K. J. Org. Chem.
1991, 56, 3769-3775.
(34) Brown, H. C.; Wirkkala, R. A. J. Am. Chem. Soc. 1966, 88, 1447-
1452.

1,2-Shifts of H Atoms in Aryl Radicals
J. Am. Chem. Soc., Vol. 121, No. 23, 1999 5447
Scheme 5
Scheme 6
Table 1. Flash Vacuum Pyrolyses of
2-Bromobenzo[c]phenanthrene (9)
relative amountsa
(°C) recovery (%)b 9 (%) 4 (%) 14 (%)
temp
mass
yield of purified
mixture (%)
950
1000
1050
1100
72
73
97
86
69
31
0
25
60
90
90
6
9
10
10
37
27
35
29
0
a Determined by integration of the H NMR spectra of the purified
mixture of monomeric products; error limits estimated at approximately
(5%. b Mass recovery (%) based on the theoretical yield of 4.
1
Since the optimum temperature for ring closure of 9 to 4
was found to be 1050 °C (Table 1), pyrolysis of the deuterated
analogue (13) was likewise performed at 1050 °C under the
same conditions (Scheme 6). 1H NMR analysis of the pyrolysate
(Figure 2B) indicates that the products of cyclization (15 + 16)
again outweigh the products of hydrogen abstraction (17 + 18),
although now in a ratio of only approximately 3:1. No starting
material (13) survived. The reduced dominance of the hydrogen-
shift pathway is consistent with an unfavorable kinetic isotope
effect, which could retard the deuterium 1,2-shift but should
not affect the rate of the hydrogen-abstraction reactions leading
to 17 and 18.
Comparing the 1H NMR spectrum of the pyrolysate from 13
(Figure 2B) with that from the pyrolysis of the nondeuterated
aryl bromide (9, Figure 2A), we see a new, small singlet in the
trough of the original doublet at 9.15 ppm (labeled Hc and Ha,
respectively, in Figure 2B). This signal comes from a hydrogen
in the cove region of a benzo[c]phenanthrene that must have a
deuterium at the adjacent position, since it lacks vicinal coupling.
We assign that singlet to compound 17, which presumably is
formed by a 1,2-shift of deuterium followed by a hydrogen
abstraction. Compound 17 should also contribute to the doublet
at 9.15 ppm (labeled Ha), since the other hydrogen in the cove
region of 17 does have a hydrogen neighbor.
The fact that doublet Ha in Figure 2B is larger than singlet
Hc, however, indicates that another compound must also
contribute to the doublet labeled Ha. The logical candidate is
18, the expected product of hydrogen abstraction by the original
aryl radical, prior to any 1,2-shift of deuterium. The singlet for
the hydrogen in 18 attached to the position vacated by the
bromine atom, unfortunately, is lost among the jumble of signals
at higher field. From the signals at 9.15 ppm, therefore, we
conclude that both 17 and 18 were formed; however, the signals
are small, and thus both are just minor products.
Figure 2. 1H NMR spectra of the pyrolysate from FVP of (A)
2-bromobenzo[c]-phenanthrene and (B) [1,3,4,5-2H4]-2-bromobenzo-
[c]phenanthrene.
or about 4:1 (Table 1). As the temperature is raised, the
hydrogen-shift pathway grows increasingly dominant (ca. 7:1
at 1000 °C and 9:1 at 1050 and 1100 °C).
Figure 2A shows the 1H NMR spectrum of the product
mixture obtained from pyrolysis of 2-bromobenzo[c]phenan-
threne (9) at 1050 °C. The small doublet at 9.15 ppm (labeled
Ha) comes from the hydrogen in the cove region of the minor
product (14), whereas the large doublet at 8.14 ppm (labeled
Hb) comes from the hydrogen in the five-membered ring bay
region of the major product (4), i.e., the hydrogen shown
explicitly on the structure of 4 in Scheme 1. The dominance of
the hydrogen-shift pathway is clearly apparent.
Following a 1,2-shift of deuterium, the newly formed radical
in the cove region has a choice between hydrogen abstraction
to give 17 (already identified above) and cyclization to give
15. Recalling the dominance of cyclization over hydrogen
abstraction in the nondeuterated radicals (Figure 2A), we would
expect 15 to be formed in greater amount than 17, and such
appears to be the case. Compound 15 is responsible for the major
portion of the large doublet at 8.14 ppm (labeled Hb in Figure
2B). The other contribution to the doublet labeled Hb in Figure

5448 J. Am. Chem. Soc., Vol. 121, No. 23, 1999
Brooks and Scott
Scheme 7
Scheme 8
2B is compound 16, the presence of which is revealed by the
small singlet at 8.14 ppm labeled Hd. Although the pathway by
which 16 is formed may not be obvious (discussed further
below), it should be recognized that 16 is a minor product
relative to 15, as are 17 and 18. The most important conclusion,
therefore, is that the single most abundant product (15) and
one of the three minor products (17) from the pyrolysis of 13
are those expected from a 1,2-shift of deuterium.
The minor cyclized product 16, which has no deuterium at
the site where bromine was originally attached, most likely
represents a secondary product, formed by cyclodehydrogenation
of 18. Pyrolysis of benzo[c]phenanthrene (14) is known to give
benzo[ghi]fluoranthene (4), usually accompanied by varying
amounts of the subsequent rearrangement product cyclopenta-
[cd]pyrene (19), depending on the temperature and the pressure
(Scheme 7).48-50
Under our conditions (1050 °C, nitrogen carrier gas, pressure
) 2.0-2.5 mmHg), the pyrolysis of 14 gives mostly unchanged
14 (78%) plus minor amounts of 4 (12%) and 19 (10%).51,52
This degree of conversion establishes a lower limit of what
would be seen during the pyrolysis of aryl bromides 9 and 13,
since the substantially higher concentrations of aryl radicals in
those experiments should promote the cyclodehydrogenation of
benzo[c]phenanthrene by exothermic hydrogen-atom abstraction
from the cove region. The unimolecular isomerization of 4 to
19, on the other hand, should remain relatively insignificant,
which presumably accounts for why we see none from the
pyrolysis of aryl bromides 9 and 13.
The alternative possibility that 16 might be formed from 13
by way of the benzyne intermediate 20 (Scheme 8) looks
attractive on the surface, but our results exclude this pathway.
Precedent abounds for the reversible ring contraction of ben-
zynes and for subsequent trapping of the resulting vinylidenes
by C-H bond insertion;23-26,53 however, the final 6/5-ring
switch rearrangement required to generate 16 from 22 would
be endothermic; cyclopenta[cd]pyrene is known to be the energy
sink on the C18H10 energy surface.48-50 The absence of any
detectable 22 among the pyrolysis products from 13 argues
against this benzyne mechanism as the source of 16.
studies provide further evidence that the major reaction pathway
involves migration of an adjacent hydrogen atom to the site
where the original radical was generated. The high temperatures
employed to induce homolytic cleavage of an aryl bromide bond
(flash vacuum pyrolysis at 950-1100 °C) gives the resulting
radical more than enough energy to surmount the barrier for
hydrogen migration. These results strongly suggest that the 1,2-
shift of hydrogen atoms at high temperatures should also occur
in simple aryl radicals, such as the phenyl radical, for which an
energy barrier of 58.4 kcal/mol has been calculated.
Aryl radicals are believed to play a major role in the chemistry
of flames, combustion, and soot formation. Kinetic models for
such processes should take into account the intramolecular
equilibrations of isomeric radicals, such as the 1- and 2-naphthyl
radicals, by reversible 1,2-shifts of hydrogens atoms.
Experimental Section
General. Diethyl ether was purified by distillation under nitrogen
from the sodium ketyl of benzophenone. All other solvents and
commercial chemicals were of the best available grade and were used
without further purification, unless otherwise stated. Proton NMR
chemical shifts are reported in ppm downfield from tetramethylsilane
with tetramethylsilane (δ ) 0.00 ppm) as the reference standard, unless
otherwise specified. Carbon NMR shifts are reported in ppm downfield
from tetramethylsilane with deuteriochloroform (δ ) 77.23 ppm) as
the reference standard. Preparative thin layer chromatographies were
performed on 20 cm × 20 cm Analtech Uniplate Taper plates, silica
gel GF. For column chromatographies, silica gel 60-200 mesh was
used. High-resolution mass spectrometry (HRMS) was performed by
the Mass Spectroscopy Laboratory, School of Chemical Sciences,
University of Illinois. Elemental analysis was performed by Robertson
Microlit Laboratories. Melting points are uncorrected.
(E)- and (Z)-1-(p-Bromophenyl)-2-(2-naphthyl)ethylene (8). A
suspension of 2-bromomethylnaphthalene54 (6, 0.44 g, 2.0 mmol) and
triphenylphosphine (0.52 g, 2.0 mmol) in 5 mL of dimethylformamide
was heated at 150 °C for 30 min. After the solution was cooled to
room temperature, p-bromobenzaldehyde (0.37 g, 2.0 mmol) and lithium
ethoxide (2.2 mL of a 1.0 M solution in ethanol, 2.2 mmol) were added,
and the resulting suspension was stirred at room temperature for 3.5 h.
The reaction mixture was diluted with 100 mL of diethyl ether and
washed successively with 50 mL each of 10% aqueous hydrochloric
Conclusion
From an aryl radical poised to relieve steric congestion by a
1,2-shift of hydrogen (1), the dominant product (4) has been
found to derive from the rearranged radical. Deuterium-labeling
(48) Necula, A.; Scott, L. T., unpublished results.
(49) Plater, M. J. Tetrahedron Lett. 1994, 35, 6147-6150.
(50) Sarobe, M.; Jenneskens, L. W.; Wiersum, U. E. Tetrahedron Lett.
1996, 37, 1121-1122.
(51) Tintel, C.; Cornelisse, J.; Lugtenburg, J. Recl. TraV. Chim. Pays-
Bas 1983, 102, 14-20.
(52) Jans, A. W. H.; Tintel, C.; Cornelisse, J.; Lugtenburg, J. Magn.
Reson. Chem. 1986, 24, 101-104.
(53) Brown, R. F. C.; Eastwood, F. W. Synlett 1993, 9-19 and references
therein.
(54) Purchased from Aldrich Chemical Co., Milwaukee, WI.

1,2-Shifts of H Atoms in Aryl Radicals
J. Am. Chem. Soc., Vol. 121, No. 23, 1999 5449
acid, water, and brine. The organic layer was separated and dried over
magnesium sulfate, and the solvent was evaporated under reduced
pressure. Purification by column chromatography on silica gel with
hexane as eluant yielded 0.51 g (82%) of a mixture of cis and trans
isomers (cis to trans ratio of ca. 2.5:1 by 1H NMR). For the subsequent
photochemical cyclization, the isomers were not separated; however,
for characterization purposes, partial separation could be achieved by
further column chromatography on silica gel with hexane as eluant.
(E)-1-(p-Bromophenyl)-2-(2-naphthyl)ethylene (8). Recrystalliza-
tion from acetone gave colorless plates: mp 188.5-189 °C; 1H NMR
(400 MHz, CDCl3) δ 7.87-7.81 (m, 4 H), 7.73 (dd, J ) 8.6 Hz, J )
1.8 Hz, 1 H), 7.52-7.46 (m, 4 H), 7.43 (d, J ) 8.8 Hz, 2 H), 7.27 (d,
J ) 16.4 Hz, 1 H), 7.16 (d, J ) 16.4 Hz, 1 H); 13C NMR (100 MHz,
CDCl3) δ 136.53, 134.65, 133.87, 133.35, 132.04, 129.71, 128.62,
128.25, 128.21, 127.93, 127.08, 126.64, 126.29, 123.58, 121.59; 13C
NMR (100 MHz, DMSO-d6, δ ) 39.51 ppm as internal standard) δ
136.42, 134.44, 133.26, 132.68, 131.68, 129.30, 128.51, 128.28, 127.92,
127.73, 127.64, 126.66, 126.56, 126.18, 123.60, 120.62; UV-vis
(hexane) λmax (ꢀ) 230 (23 000), 240 (sh, 13 000), 251 (12 000), 273
(27 000), 282, (30 000), 306 (sh, 31 000), 321 (40 000), 335 (sh, 30
000), 356 (sh, 5900); MS (EI, 70 eV) m/z (relative intensity) 310 (46,
M+), 308 (45, M+), 229 (64), 228 (100); HRMS, calcd for C18H13Br
(M+) 308.0200, found 308.0201. Anal. Calcd for C18H13Br: C, 69.92;
H, 4.24. Found: C, 69.72; H, 4.19.
(Z)-1-(p-Bromophenyl)-2-(2-naphthyl)ethylene (8). Recrystalliza-
tion from methanol gave fine white needles: mp 80-81 °C; 1H NMR
(400 MHz, CDCl3) δ 7.80-7.77 (m, 1 H), 7.75-7.72 (m, 1 H), 7.72
(s, 1 H), 7.67 (d, J ) 8.8 Hz, 1 H), 7.47-7.43 (m, 2 H), 7.35-7.30
(m, 3 H), 7.14 (d, J ) 8.0 Hz, 2 H), 6.80 (d, J ) 12.4 Hz, 1 H), 6.59
(d, J ) 12.0 Hz, 1 H); 13C NMR (100 MHz, CDCl3) δ 136.29, 134.62,
133.64, 132.83, 131.58, 131.16, 130.82, 129.46, 128.21, 128.15, 127.90,
127.85, 126.90, 126.35, 126.27, 121.27; UV-vis (hexane) λmax (ꢀ) 229
(44 000), 252 (17 000), 271 (20 000), 302 (17 000); MS (EI, 70 eV)
m/z (relative intensity) 310 (35, M+), 308 (36, M+), 230 (11), 229
(63), 228 (100); HRMS, calcd for C18H13Br (M+) 308.0200, found
308.0201. Anal. Calcd for C18H13Br: C, 69.92; H, 4.24. Found: C,
69.73; H, 4.05.
130.85, 130.73, 129.18, 129.14, 129.03, 128.93, 128.39, 127.55, 127.29,
127.19, 126.72, 126.67, 126.46, 125.12, 120.21; UV-vis (hexane) λmax
(ꢀ) 218 (39 000), 225 (sh, 35 000), 231 (33 000), 243 (12 000), 253
(sh, 14 000), 264 (sh, 26 000), 274 (52 000), 284 (72 000), 305 (sh, 12
000), 317 (10 000), 329 (sh, 3500); MS (EI, 70 eV) m/z (relative
intensity) 308 (51, M+), 306 (52, M+), 227 (45), 226 (100); HRMS,
calcd for C18H11Br (M+) 306.0044, found 306.0044. Anal. Calcd for
C18H11Br: C, 70.38; H, 3.61. Found: C, 70.31; H, 3.40.
Flash Vacuum Pyrolysis of 2-Bromobenzo[c]phenanthrene. Flash
vacuum pyrolyses were performed on 50-mg samples of 2-bromobenzo-
[c]phenanthrene (9) at temperatures ranging from 950 to 1100 °C, with
a steady flow of nitrogen carrier gas (final pressure: 2.0-2.5 mmHg),
as previously described.56 The crude pyrolysate was purified by flash
vacuum chromatography on silica gel with hexane as eluant to yield a
mixture of unchanged 9, benzo[ghi]fluoranthene (4) and benzo[c]-
phenanthrene (14). The product composition of the mixture varied with
1
the temperature of pyrolysis and was determined by H NMR (Table
1). Partial separation of the pyrolysis products 4 and 14 was possible
by bulb-to-bulb distillation under reduced pressure. Benzo[ghi]fluo-
ranthene (4): mp 148-149 °C (lit.39 147-149 °C); 1H NMR (400 MHz,
CDCl3) δ 8.14 (d, J ) 6.8 Hz, 2 H), 7.99 (d, J ) 8.8 Hz, 2 H), 7.95
(d, J ) 8.0 Hz, 2 H), 7.95 (d, J ) 8.4 Hz, 2 H), 7.70 (dd, J ) 8.0 Hz,
J ) 6.8 Hz, 2 H); 13C NMR (100 MHz, CDCl3) δ 137.55, 133.42,
133.25, 128.53, 127.87, 126.90, 126.85, 126.60, 125.25, 123.62; MS
(EI, 70 eV) m/z (relative intensity) 226 (100, M+).
Flash Vacuum Pyrolysis of 2-Bromobenzo[c]phenanthrene-d4
(13). Flash vacuum pyrolysis was performed on 50 mg of 2-bromoben-
zo[c]phenanthrene-d4 (13) at 1050 °C, with a steady flow of nitrogen
carrier gas (final pressure: 2.0-2.5 mmHg), as previously described.56
The crude pyrolysate (32 mg, 86% mass recovery) was purified by
flash vacuum chromatography through a short plug of silica with hexane
as eluant to yield 28 mg (76%)57 of a mixture of benzo[ghi]fluoranthene-
d4/d3 (15/16) and benzo[c]phenanthrene-d4 (17/18) in a ratio of 3:1;58
no starting material survived. Benzo[ghi]fluoranthene-d4/d3 (15/16): 1H
NMR (400 MHz, CDCl3) δ 8.14 (d, J ) 6.8 Hz, 1 H), 8.14 (s, small,
1 H), 7.99 (d, J ) 8.8 Hz, 1 H), 7.95 (d, J ) 8.0 Hz, 1 H), 7.95 (s, 1
H), 7.94 (d, J ) 8.8 Hz, 1 H), 7.70 (dd, J ) 8.0 Hz, J ) 7.2 Hz, 1 H);
the presence of 16 was detected by a small singlet at 8.14 ppm. By 1H
NMR integration, 15 and 16 formed in a ratio of 2.5:1. Benzo[c]-
phenanthrene-d4 (17/18): 1H NMR (400 MHz, CDCl3) δ 9.15 (d, J )
8.4 Hz, 1 H), 9.15 (s, small, 1H), 8.04 (dd, J ) 8.0 Hz, J ) 1.6 Hz, 1
H), 7.92 (d, J ) 8.4 Hz, 1 H), 7.85 (d, J ) 8.4 Hz, 1 H), 7.84 (s, 1 H),
7.70 (td, J ) 7.6 Hz, J ) 1.6 Hz, 1 H), 7.64 (td, J ) 7.4 Hz, J ) 1.2
Hz, 1 H); the presence of 17 was detected by a small singlet at 9.15
ppm.
2-Bromobenzo[c]phenanthrene (9). A solution of (E)- and (Z)-1-
(p-bromophenyl)-2-(2-naphthyl)ethylene55 (8, 0.50 g, 1.6 mmol), iodine
(0.63 g, 2.5 mmol), and propylene oxide33 (17 mL, 240 mmol) in 1.6
L of benzene was photolyzed in a quartz vessel for 8.5 h in a Rayonet
photochemical apparatus equipped with 13 of the 254-nm 35-W mercury
lamps. The reaction mixture was then concentrated to 150 mL by rotary
evaporation and washed with 15% aqueous sodium thiosulfate, water,
and brine. The organic layer was separated and dried over magnesium
sulfate, and the solvent was evaporated under reduced pressure.
Purification by column chromatography on silica gel with hexane as
eluant yielded 0.39 g (78%) of an off-white solid. Recrystallization
from methanol gave 2-bromobenzo[c]-phenanthrene (9) as fine white
Flash Vacuum Pyrolysis of Benzo[c]phenanthrene (14). Flash
vacuum pyrolysis was performed on 50 mg of benzo[c]phenanthrene
at 1050 °C, with a steady flow of nitrogen carrier gas (final pressure:
2.0-2.5 mmHg), as previously described.56 The crude pyrolysate (47
mg, 94% mass recovery) was purified by preparative thin-layer
chromatography on silica gel with hexane as eluant to yield 28 mg
(56%) of a mixture of benzo[ghi]fluoranthene (4, 12%), cyclopenta-
[cd]pyrene (19, 10%)51 and unchanged benzo[c]phenanthrene (78%).58
1
needles: mp 87-88 °C (lit.32 mp 86-87.5 °C); H NMR (400 MHz,
CDCl3) δ 9.29 (s, 1 H), 9.05 (d, J ) 8.4 Hz, 1 H), 8.04 (d, J ) 7.6 Hz,
1 H), 7.93 (d, J ) 8.4 Hz, 1 H), 7.90 (d, J ) 8.8 Hz, 1 H), 7.87 (d, J
) 8.8 Hz, 1 H), 7.85 (d, J ) 8.0 Hz, 1 H), 7.83 (d, J ) 8.4 Hz, 1 H),
7.75 (t, J ) 7.6 Hz, 1 H), 7.72 (d, J ) 8.4 Hz, 1 H), 7.66 (t, J ) 7.4
Hz, 1 H); 13C NMR (100 MHz, CDCl3) δ 133.71, 132.14, 131.74,
131.62, 130.38, 130.25, 129.23, 128.91, 128.37, 127.57, 127.55, 127.21,
126.91, 126.86, 126.53, 126.36, 120.79; 13C NMR (100 MHz, DMSO-
d6, δ ) 39.51 ppm as internal standard) δ 133.15, 131.73, 131.29,
Acknowledgment. Financial support from the Department
of Energy is gratefully acknowledged. We are also indebted to
Richard P. Johnson and James E. Jackson for valuable discus-
sions.
(55) The ultraviolet light that causes photocyclization also interconverts
the (E)- and (Z)-stilbene isomers; therefore, either isomer (or a mixture of
isomers) can be used for the photochemical ring closure.
(56) For a more detailed description of the apparatus and the procedure,
see: Scott, L. T.; Bratcher, M. S.; Hagen, S. J. Am. Chem. Soc. 1996, 118,
8743-8744.
Supporting Information Available: Experimental proce-
dures for the syntheses of [1,3,4,5-2H4]-2-bromobenzo[c]-
phenanthrene (13) and benzo[c]phenanthrene (14). NMR spectra
of the synthetic intermediates, pyrolysis precursors, and pyrolysis
products (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
(57) Yield of purified mixture (%) after separation from higher molecular
weight byproducts by flash vacuum chromatography.
1
(58) Determined by integration of the H NMR spectra of the purified
mixture of monomeric products.
JA984472D