Thermal reduction of 7H-benz[d,e]anthracen-7-one and related ketones under hydrogen-transfer conditions

Article abstract of DOI:10.1021/jo015637v

In the presence of hydrogen donor solvents and at elevated temperatures, aromatic ketones can be selectively deoxygenated to the corresponding hydroaromatic compounds. The kinetics for reduction of 7H-benz[d,e]anthracen-7-one (benzanthrone, 6) into 7H-benz[d,e]anthracene (benzanthrene, 1) in 9,10-dihydroanthracene (3) solvent has been investigated in detail. The relatively slow hydrogenation of 6 is due to reversibility of the initial hydrogen-transfer step according to a reverse radical disproportionation (RRD). The dynamics could well be rationalized using the energetics of species computed by density functional theory (DFT). The application of hydrogen donors such as 1 as a hydrogen-transfer agent, although favorable in terms of a low benzylic carbon-hydrogen bond dissociation enthalpy, is limited due to the slow self-hydrogenation, which in case of 1 gives 5,6-dihydro-4H-benz[d,e]anthracene (7).

Full text of DOI:10.1021/jo015637v

J . Org. Chem. 2001, 66, 6611-6619
6611
Th er m a l Red u ction of 7H-Ben z[d ,e]a n th r a cen -7-on e a n d Rela ted
Keton es u n d er Hyd r ogen -Tr a n sfer Con d ition s
Peter Mulder,*,† Sven Hemmink,† Martine I. De Heer,† Marco Lupo,† Danilo Santoro,† and
Hans-Gert Korth*,‡
Leiden Institute of Chemistry, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands,
and the Institut fu¨r Organische Chemie, Universita¨t Essen, D-45117 Essen, Germany
p.mulder@chem.leidenuniv.nl
Received March 20, 2001
In the presence of hydrogen donor solvents and at elevated temperatures, aromatic ketones can be
selectively deoxygenated to the corresponding hydroaromatic compounds. The kinetics for reduction
of 7H-benz[d,e]anthracen-7-one (benzanthrone, 6) into 7H-benz[d,e]anthracene (benzanthrene, 1)
in 9,10-dihydroanthracene (3) solvent has been investigated in detail. The relatively slow
hydrogenation of 6 is due to reversibility of the initial hydrogen-transfer step according to a reverse
radical disproportionation (RRD). The dynamics could well be rationalized using the energetics of
species computed by density functional theory (DFT). The application of hydrogen donors such as
1 as a hydrogen-transfer agent, although favorable in terms of a low benzylic carbon-hydrogen
bond dissociation enthalpy, is limited due to the slow self-hydrogenation, which in case of 1 gives
5,6-dihydro-4H-benz[d,e]anthracene (7).
Sch em e 1
In tr od u ction
At elevated temperatures, hydroaromatic solvents with
weakly bonded (benzylic-type) hydrogen atoms can act
as thermal reducing agents.1 For example, it has been
demonstrated that styrenes,1a fullerenes,1a quinones,1a
aromatic nitro and nitroso compounds,1a substituted
naphthalenes,1b,c and native lignins1d can be reduced in
this way. In the first step, a hydrogen atom is shuttled
from the donor solvent (RH) to an unsaturated moiety
(eq 1). This chemical transformation is known as the
reverse radical disproportionation, RRD, and plays an
important role in the liquefaction of coal. Subsequently,
a second hydrogen atom is transferred, according to eq
2, yielding the hydrogenated product.
donors, 7H-benz[d,e]anthracene (benzanthrene, 1 (Scheme
1)) contains a benzylic BDE(C-H)2a,b of as low as 64 kcal
mol-1, compared to 83 kcal mol-1 for a less efficient donor
solvent such as 1,2,3,4-tetrahydronaphthalene (tetralin,
2).2c This implies that, provided eq 1 is the rate-
determining step (i.e., v2 . v-1), the reduction rate at
300 °C by 1 will be more than 107 times faster than 2.
Furthermore, in the reduction of, e.g., R-methylstyrene
with a mixture of 1 and 9,10-dihydroanthracene (3), 1
may serve as a hydrogen-transfer agent (catalyst)3 since
it is not consumed but recycled by the auxiliary donor
solvent.4
R1R2CdCH2 + RH f R1R2C(•)CH3 + R•
R1R2C(•)CH3 + RH f R1R2CHCH3 + R•
(1)
(2)
In principle, the activation enthalpy for eq 1 is gov-
erned by the reaction enthalpy, because the reverse
reaction, a radical-radical disproportionation, is a near
diffusion-controlled process. The temperature at which
the thermal hydrogenation occurs at an appreciable rate
depends strongly on the thermodynamic properties of
both the hydrogen donor and hydrogen acceptor mol-
ecules. The relative RRD capability of donor molecules
can be directly derived from the respective benzylic
carbon-hydrogen bond dissociation enthalpy, BDE(C-
H). For example, one of the most reactive hydrogen
However, the efficiency of 1 appears to be less than
anticipated on the basis of enthalpic considerations. In
our study on the cleavage of R-phenoxyacetophenone at
around 300 °C, eqs 3-4, we observed that the addition
(2) (a) Bausch, M. J .; Gostowski, R.; J irka, G.; Selmarten, D.; Winter,
G. J . Org. Chem. 1990, 55, 5805-5806. (b) Bordwell, F. G.; Zhang,
X.-M. Acc. Chem. Res. 1993, 26, 510-517. (c) Laarhoven, L. J . J .;
Mulder, P. J . Phys. Chem. B 1997, 101, 73-77. (d) Malhotra, R.;
McMillen, D. F. Energy Fuels 1990, 4, 184-193.
(3) Morgenthaler, J .; Ru¨chardt, C. Eur. J . Org. Chem. 1999, 2219,
9-2230.
(4) The 9,10-dihydroanthracenyl radicals decay through dispropor-
tionation. For benzanthrenyl radical, without a reactive hydrogen atom
at the â-position, disproportionation with 9,10-dihydroanthracenyl
recycles 1.
† Leiden University.
‡Universita¨t Essen.
(1) (a) Ru¨chardt, C.; Gerst, M.; Ebenhoch, J . Angew. Chem., Int. Ed.
Engl. 1997, 36, 1406-1430 and references therein. (b) Arends, I. W.
C. E.; Mulder, P. Energy Fuels 1996, 10, 235-242. (c) Dorrestijn, E.;
Laarhoven, L. J . J .; Arends, I. W. C. E.; Mulder, P. J . Anal. Appl. Pyrol.
2000, 54, 153-192. (d) Dorrestijn, E.; Kranenburg, M.; Poinsot, D.;
Mulder, P. Holzforschung 1999, 53, 611-616.
10.1021/jo015637v CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/08/2001

6612 J . Org. Chem., Vol. 66, No. 20, 2001
Mulder et al.
F igu r e 1. Conversion of benzanthrone (6) and 9,10-dihydroanthracene (3) at 300 °C with [6]0 ) 0.42 M and [3]0 ) 4.27 M. Solid
lines: 6 (2), benzanthrene (1, 9), and 5,6-dihydro-4H-benz[d,e]anthracene (7, b). Dashed lines: 3 (4), anthracene (11, 0), and
1,2,3,4-tetrahydroanthracene (12, O).
of 1 to 3 (with a BDE(C-H)2d of 77 kcal mol-1) afforded
only a modest accelerating effect on the rate of disap-
pearance of the substrate.5
of donor solvents has been investigated to obtain insight
into the dynamics of their reduction pathways. To assess
the thermokinetic behavior of the (intermediate) species,
the benzylic BDE(C-H)s need to be known with some
precision. Since pertinent and consistent experimental
information is scarcely available, we have computed the
BDEs for a large set of compounds and related radicals
by density functional theory (DFT) employing the B3LYP/
6-31G(d,p) method.
PhC(O)CH2OPh + RH f PhC(•)OHCH2OPh + R•
(3)
PhC(•)OHCH2OPh f PhC(O)CH3 + PhO• (4)
The mechanistic and dynamic aspects of RRD for the
hydrogenation of olefinic moieties have been well docu-
mented for a range of donor and acceptor molecules.1a
For the transfer hydrogenation of carbonyl groups,
however, these features are less well understood. For
example, in the thermal reduction of benzophenone (4),
the expected product (benzhydrol) is formed only in
minute amounts. Instead, the dehydroxylated compound
(diphenylmethane) emerges as the most prominent prod-
uct.6 The formation of the alcohol can be easily rational-
ized analogously to eqs 1-2 and involves the formation
of an intermediate ketyl radical. However, the precise
mechanism for the loss of the hydroxyl in a nonpolar
environment remains somewhat obscure. A mechanism
has been postulated in which the intermediate alcohol
dimerizes to an ether with loss of water. Next, hydrogen
atom abstraction from the ether provides the diphenyl-
methane and benzophenone.6 Recently, we have demon-
strated that the quantitative conversion of o-hydroxyl
benzylic alcohol, dissolved in 1, 3, or 9,10-dihydro-
phenanthrene (5), into o-hydroxytoluene proceeds through
an ionic intermediate. An acid-catalyzed removal of the
hydroxyl group affords the corresponding benzylic car-
bocation, which abstracts a hydride from the hydroaro-
matic solvent.7
Resu lts
Ben za n t h r on e (6) in 9,10-Dih yd r oa n t h r a cen e
(3): P r od u cts. The product distribution for the temporal
conversion of 6 at 300 °C, with 3 as the hydrogen-donor
solvent (molar ratio of 6/3 ) 1:10) is displayed in Figure
1.
The product distributions under various reaction con-
ditions are summarized in Tables S1 and S2 of the
Supporting Information. The reduction product is ben-
zanthrene (1) with m/z 216 (M+), which at prolonged
reaction time decays into a hydrogenated product with
m/z 218 (M+), according to GC-MS analysis. To isolate
this compound for the structural identification, a thermal
reaction on a preparative scale was carried out with 6 in
tetralin (a solvent that can be easily removed by distil-
lation) at 365 °C. The isolated and purified product was
analyzed by NMR spectrometry (including a simulation
analysis) and identified as 4,5-dihydro-6H-benz[d,e]an-
thracene (7), and not its isomer 2,3-dihydro-1H-benz[d,e]-
anthracene (8) (Scheme 2). In the Supporting Informa-
tion, the NMR details are provided.
Some minor products (each with yields of less than
0.5%) were observed as well: a compound with a m/z 232
(M+) assigned as 7H-benz[d,e]anthracen-7-ol (benzan-
threnol, 9), an isomer of 1, and at longer reaction times
a hydrogenation product of 7 with m/z 234 (M+). Through-
out the experiments with 6 (m/z 230, M+), a product with
m/z 234 (M+) (100) and m/z 215 (M+-19) (20) was
obtained. The molecular mass of this compound, which
could be removed from the reaction mixture by aqueous
alkaline extraction, suggests the addition of four hydro-
gens to 6 with retention of oxygen. By comparing reten-
tion times of analogue compounds, this product is ten-
On the basis of the foregoing, a strong reducing reagent
such as benzanthrene (1) could well be obtained in situ
from the readily available 7H-benz[d,e]anthracen-7-one
(benzanthrone, 6). In this study, the reactivity of ben-
zanthrone and related aromatic ketones in the presence
(5) Dorrestijn, E.; Hemmink, S.; Hulstman, G.; Monnier, L.; Schep-
pingen, van W.; Mulder, P. Eur. J . Org. Chem. 1999, 607, 7-616
(correction Eur. J . Org. Chem. 1999, 1267, 7).
(6) Choi, C.; Stock, L. M. J . Org. Chem. 1984, 49, 2871-2875.
(7) Dorrestijn, E.; Kranenburg, M.; Ciriano, M. V.; Mulder, P. J . Org.
Chem. 1999, 64, 3012-3018.

Thermal Reduction of 7H-Benz[d,e]anthracen-7-one
J . Org. Chem., Vol. 66, No. 20, 2001 6613
Sch em e 2
sion. With the data from Table S1 and S2 (experiments
1.1-1.3 and 2.1-2.10), a plot was constructed of ln([6]/
[6]0)([3]0/[3]) versus ([6]0 - [3]0)t. A quite satisfying linear
regression line (r2 ) 0.96) was obtained with a slope of
kexp,6 ) 7.1 × 10-5 M-1 s-1 8b
.
Isotop e Effect on th e Hyd r ogen a tion of Ben za n -
th r on e (6). Experiments were conducted with benzan-
throne and mixtures of anthracene or perdeuterated
anthracene and 9,10-dihydroanthracene (molar ratio of
1:1) to explore the kinetic isotope effect on the RRD.9
After 60 min and at 300 °C, the conversion of 6 reached
the same level of 24% for both experiments, indicating
the absence of a (primary) isotope effect. The molar ratio
between 9,10-dihydroanthracene and anthracene dropped
from 1:1 to 0.7:1 during reaction.
Integrated MS spectra were disentangled to quantify
the H/D distributions in reactants and products. The
assumption was made that the degree of the most
predominant fragmentation (loosing hydrogen or deute-
rium) remains constant for alike compounds.
Six isotopomers could be identified for 9,10-dihydroan-
thracene, denoted as (H/D)8An(H/D)4 with (H/D4) as the
hydrogens or deuteriums attached to the positions C-9
and C-10. The product ratios were as follows: H8AnH4/
H8AnH3D/H8AnH2D2 ) 0.53:0.33:0.14 and D8AnH4/D8-
AnH3D/D8AnH2D2 ) 0.05:0.41:0.54. The overall molar
ratio between H8An(H/D)4 and D8An(H/D)4 compounds
was 0.93.
In addition, six isotopomers were detected for an-
thracene with ratios for H8AnH2/H8AnHD/H8AnD2 of
0.85:0.13:0.02 and for D8AnH2/D8AnHD/D8AnD2 of 0.21:
0.27:0.52, together with a molar ratio between H8An(H/
D)2 and D8An(H/D)2 compounds of 1.03.
tatively assigned as a phenolic compound, 5,6-dihydro-
7-hydroxy-4H-benz[d,e]anthracene (10). The product ratio
(1 + 7)/10 remained rather constant (ca. 30). In prelimi-
nary experiments with nonpurified 6, the yield of 10 could
reach a level of 20%. The same result was obtained when
the reduction of 6 was carried out in the presence of a
large amount of pyridine (see Table S2, experiment no.
2.12).
Next to the reduction of 6, the self-hydrogenation of 3
yielded anthracene (11) and 1,2,3,4-tetrahydroanthracene
(12) in equimolar amounts. The mass balances for 6 (sum
of remaining 6 and the products 1, 7, and 10) and for 3
(sum of the remaining 3 and the products 11 and 12)
were always better than 93%. The overall hydrogen
balance could be well described by 2[1] + 3[7] + 2[10] )
[11] - [12].
Benzanthrene is an unstable compound at room tem-
perature in the presence of air, where it is slowly
reoxidized to benzanthrone.
Ben za n th r on e (6) in 9,10-Dih yd r oa n th r a cen e (3):
Kin etics. The concentration of 6 was varied in such a
way that the ratio 6/3 ranged from 1:5 to 1:80. The
conversion of 6 remained constant at around 70%; only
the rate of hydrogenation of 1 into 7 was slowed when
applying a lower initial concentration of 6. When the
concentration of 3 was lowered by dilution with diphenyl
ether, the rate of conversion of 6 decreased and became
zero in the absence of a hydrogen donor. At first glance,
it seems somewhat surprising that the rate of reduction
of 6 depends on 3 and not on the in situ generated
benzanthrene (1), which is a superior hydrogen donor.
For example, after 60 min 0.24 M of 1 is obtained when
starting with mixture of 4.27 M 3 and 0.42 M 6.8a
Addition of an excess of anthracene (a reaction product)
retarded the hydrogenation of 6. The results are collected
in Table S2.
After reaction, the aromatic nuclei containing either
H or D are equally distributed in both 9,10-dihydroan-
thracene and anthracene. Besides, the overall H/D ratio
for the (benzylic) C-9 and C-10 positions is around three,
identical with that of the starting mixture. The finding
underscores that under these reaction conditions the
equilibrium H8AnH4 + D8AnD2 h H8AnH2 + D8AnD2H2
is firmly established. In the absence of benzanthrone, the
degree of H/D exchange is still kinetically determined.10
The deuterium incorporation in benzanthrene at the
benzylic C-7 position, Bz(H/D)2, was higher compared to
that in 9,10-dihydroanthracene. A considerable H/D
exchange has occurred in the aromatic nucleus as well.
Four isotopomers were present with a ratio for BzHD/
BzD2/DBzD2/D2BzD2 of 29:62:8:1; nondeuterated benzan-
threne (BzH2) could not be detected.
Finally, in the remaining benzanthrone some 18% H/D
exchange had taken place.
Ben za n th r on e (6) in Oth er Hyd r ogen Don or s. For
comparison, the hydrogenation of benzanthrone was
carried out in three other hydrogen-donor solvents:
tetralin (2), 9,10-dihydrophenanthrene (5), and 1,2-
Without detailing the mechanism, the experimental
rate constant for the disappearance of 6 (kexp,6) was
retrieved assuming v6 ) kexp,6[6][3] as the kinetic expres-
(8) (a) Molar densities of the reagents used in this study are not
known at the temperatures applied. By volumetric analysis, we could
estimate a value of 4.9 M for 3 used in this study at around 300 °C.
For other compounds we have assumed the same molar densities and
no mixing volumes. With an initial mixture of 0.1967 mmol 6, 2.0161
mmol 3, and 0.10 mmol 11, their concentrations at the reaction
temperature are computed as 0.42 M 6, 4.27 M 3, and 0.21 M 11. After
a reaction time of 60 min, the absolute amounts were determined as
(in mmol): 0.062 (6), 0.1118 (1), 1.6283 (3), and 0.4250 (11), or with a
constant reaction volume (in M) 0.13 (6), 0.24 (1), 3.4 (6), and 0.9 M
(11) at 300 °C (see Table S1, experiment 1.2). The latter values are
(9) Since perdeuterated anthracene is readily available, thermal H/D
exchange reaction between 3 and 11 is applied to generate in situ
deuterated 9,10-dihydroanthracene. A disadvantage, however, is that
the concentration of the deuterated hydrogen donor will increase with
reaction time. Hence, the kinetic isotope effect found can only serve
as a qualitative indication.
(10) A separate experiment with only a mixture of 3 and D8AnD2
(molar ratio of 1:1) showed a reduced degree of H/D scrambling. The
results are as follows: H8AnH4/H8AnH3D/H8AnH2D2 ) 0.80:0.16:0.04
and D8AnH4/D8AnH3D/D8AnH2D2 ) 0.06:0.37:0.57, with H8An(H/D)4/
D8An(H/D)4 ) 2.3; H8AnH2/H8AnHD/H8AnD2 ) 0.92:0.04:0.04, and D8-
AnH2/D8AnHD/D8AnD2 ) 0.09:0.04:0.87, with H8An(H/D)2/D8An(H/D)2
) 0.46.
regarded as
a typical composition and are used in the further
thermokinetic calculations. (b) In this analysis we did not use experi-
ments at extended reaction times (experiments 1.4-1.6, Table S1),
since the contribution of the consumption of 3 by self-hydrogenation
becomes significant.

6614 J . Org. Chem., Vol. 66, No. 20, 2001
Ta ble 1. Red u ction of Ben za n th r on e (6) in Va r iou s Hyd r ogen Don or Solven ts a t 300 °Ca
Mulder et al.
yield (%)
b
expt no.
t (min)
solvent
6
1
7
10
Sc
vrel
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
60
60
60
tetralin (2)
98.4
31.1
91.0
85.2
86.7
4.2
1.6
56.0
8.6
14.0
11.7
33.2
47.6
17.6
0.0
10.6
0.2
0.7
1.6
60.0
22.1
9.4
0.0
2.4
0.2
0.0
0.0
2.6
3.3
0.0
99
83
97
30
98
63
85
0
0.01
1
9,10-dihydroanthracene (3)
9,10-dihydrophenanthrene (5)
1,2-dihydronaphthalene (13)
tetralin
9,10-dihydroanthracene
9,10-dihydrophenanthrene
1,2-dihydronaphthalene
0.08
0.14
0.02
0.55
0.22
0.05
60
300
300
300
300
27.0
72.9
a
Yields normalized to 100%; products benzanthrene (1), 5,6-dihydro-4H-benz[d,e]anthracene (7), 7-hydroxy-5,6-dihydro-4H-benz[d,e]an-
b
thracene (10). Initial molar ratios (6/hydrogen donor 2, 3, 5, or 13) ) 15.2, 10.2, 12.2, 15.4. Relative rates for the disappearance of 6
according to first-order kinetics, with experiment 4.2 as the reference. c Yield of the donor solvent (S) due to (de)hydrogenation, in
experiments 4.8, 1,2-dihydronaphthalene was completely converted into a 1:1 mixture of naphthalene and tetralin.
Ta ble 2. Red u ction of Va r iou s Ar om a tic Keton es in 9,10-Dih yd r oa n th r a cen e (3) a t 300 °Ca
yields (%)
rel rates
e
g
expt no
ketone
product
vrel
vrel
5.1
5.2
5.3
5.4
5.5
5.6
benzophenone (4)
benzanthrone (6)
anthrone (14)
fluorenone (15)
xanthone (16)
90
36
62
83
100
10
diphenylmethane
benzanthrene (1)
9,10-dihydroanthracene (3)
fluorene
10
52b
c
0.10
1
0.0054
1
0.0066
1.27
0.0007
4.6
0.38
0.14
0
17
anthraquinone (17)
anthrone
79d
1.13f
a
Reaction time 60 min, data from competition experiments with equimolar amounts (ca. 0.2 M) of ketone in 3 and always in the
presence of 6. The degree of conversion of 6 was not affected by the presence of other ketones. For other products see Table S1. c The
b
d
reduction product may be 9,10-dihydroanthracene (3) or anthracene (11) and cannot be quantified in 3. Anthrone is partially reduced,
see experiment no. 5.3. e Relative rates (to 6) for the disappearance according to first-order kinetics. f Per carbonyl group. Predicted
g
relative rates using the ∆RRDH’s according to Table 3 and the rate constant for RRD according to ref 23.
dihydronaphthalene (13). As with 3, only 1 and 7 were
obtained as the major reaction products. Details are given
in Table 1.
The documented self-hydrogenation of donor solvent
13 into equimolar amounts of naphthalene and tetralin
is faster than the hydrogenation of 6.11 Interestingly,
although the RRD capacity of 2 and 5 are expected to be
the same, conversion of 6 differs markedly in these
solvents, i.e., after 300 min 13 vs 73%. This discrepancy
may be related to the presence of impurities with
different hydrogen-donating properties. For all practical
purposes, the synthesis of 1 should preferably be per-
formed in tetralin at more elevated temperatures. In this
solvent, the conversion of 6 increases from 13% at 300
F igu r e 2. Second-order plot for hydrogenation of 7H-benz-
[d,e]anthracene (1) into 5,6-dihydro-4H-benz[d,e]anthracene
(7) at 300 °C with [1]0 ) 0.42 M and [3]0 ) 4.27 M.
°C and 300 min to 99% at 365 °C and 180 min with
selectivities to 1 and 7 of 90 and 10%, respectively.
Ben za n th r en e (1) in 9,10-Dih yd r oa n th r a cen e (3).
Since 1 is converted into 7 during the hydrogenation of
6, this reaction has been studied separately by dissolving
1 in 3 at 300 °C. The temporal product distribution is
presented in Table S3. Analysis of the rate data reveals
that the decay of 1 is perfectly described by a second-
order kinetic expression: - d([1])/dt ) kexp,1[1]2. From the
plot of [1/[1] - 1/[1]0] versus reaction time (see Figure 2)
yields a linear regression line (r2 ) 0.998) with a slope
Sch em e 3
of kexp,1 ) 3.81 × 10-4 M-1 s-1
.
Ben zop h en on e (3), An th r on e (14), F lu or en on e
(15), Xa n th on e (16), a n d An th r a qu in on e (17) in 9,-
10-Dih yd r oa n th r a cen e. The behavior of some other
aromatic ketones in 3 was scrutinized as well by employ-
ing competition experiments and in the presence of
always the same amount of benzanthrone. Xanthone
appeared to be inert, and with the other ketones only the
deoxygenated analogues (hydroaromatics) were obtained
(see Table 2).
The hydrogenation of fluorenone (15) has been studied
previously (Scheme 3). With 15 in 3, and at 400 °C, the
reduction into fluorene is considerably slower (14%,
versus 17% at 300 °C, see Table 2).6 Another study has
revealed that fluorenone in 3 is hydrogenated into
(11) Gavalas, G. R.; Allen, D. T. Int. J . Chem. Kinet. 1983, 15, 219-
233.

Thermal Reduction of 7H-Benz[d,e]anthracen-7-one
J . Org. Chem., Vol. 66, No. 20, 2001 6615
Ta ble 3. Exp er im en ta l a n d B3LYP /6-31G(d ,p )//B3LYP /
6-31G(d ,p ) Com p u ted Ben zylic C-H Bon d Dissocia tion
En th a lp ies (k ca l m ol-1) in Hyd r ogen -Don or Molecu les
a n d Ra d ica ls a n d BDE(O-H) in Ketyl Ra d ica ls a t 298 K
fluorenol (80%) and fluorene (20%), with a total yield of
50% at 350 °C within 60 min. Upon increasing the
polarity of the solvent, the conversion of 15 diminishes,
while the product ratio changes in favor of fluorene.12
DF T-Com p u ted BDE(C-H)s in Hyd r oa r om a tic
Com p ou n d s a n d Der ived Ra d ica l Sp ecies. To facili-
tate the thermokinetic interpretations of our results, the
bond dissociation enthalpies, BDE(C-H) and BDE(O-
H), for a broad range of compounds and related radicals
have been computed by DFT using the B3LYP/6-31G-
(d,p)//B3LYP/6-31G(d,p) methodology. In this way, an
internally consistent set of values is obtained. The
computational results together with available experi-
mental data obtained from literature are presented in
Table 3. It should be emphasized that the experimental
data are stemming from both gas and liquid-phase
studies. Accepting the inherent experimental uncertain-
ties, the results in Table 3 demonstrate that the discrep-
ancy between theory and experiment (with some excep-
tions) is about 2 kcal mol-1. This underscores the internal
reliability of DFT at this level of theory for the computa-
tion of BDE’s. For the thermochemical assessments
throughout the discussion section, the computed values
are used unless specified otherwise.
BDEexp
-
compd
BDEexp
BDEDFT BDEDFT
toluene
ethylbenzene
cumene
cumyl
diphenylmethane
fluorene
89.6 ( 1.0a
87.0b
88.6
85.7
84.0
49.1
79.7
79.5
81.0
1
1.3
-0.9
-2.0
2.3
83.2c
47.1c
82 ( 1d
80 ( 1d
0.5
1,2-dihydronaphthalene
(1-H) (13)
1,2-dihydronaphthalene
(2-H) (13)
76.3
1,4-dihydronaphthalene
72.9
83.6
1,2,3,4-tetrahydronaphthalene 83.0 ( 1.5e
(tetralin) (2)
-0.6
1,2-dihydronaphthalen-2-yl
1,2-dihydronaphthalen-1-yl
32.4
27.7
75.2
81.0
76.7
72.1
82.3
9,10-dihydroanthracene (3)
1,2-dihydroanthracene (1-H)
1,2-dihydroanthracene (2-H)
1,4-dihydroanthracene
1,2,3,4-tetrahydroanthracene
(1-H) (12)
9,10-dihydroanthracenyl (19)
9,10-dihydrophenanthrene (5)
9,10-dihydrophenanthrenyl
Phenalene (27)
2,3-dihydrophenalene
2,3-dihydrophenalen-1-yl
7H-benz[d,e]anthracene (1)
6H-benz[d,e]anthracene (22)
4H-benz[d,e]anthracene (23)
3H-benz[d,e]anthracene (24)
1H-benz[d,e]anthracene (25)
5,6-dihydro-4H-benz[d,e]-
anthracene (4-H) (7)
77f
2.5
43.2f
42.9
84.2
31.9
59.6
83.3
46.3
65.4
60.3
61.3
53.5
55.5
83.7
0.3
2.4
Discu ssion
62 ( 3g
64 ( 3g
Red u ction of Ben za n th r on e in to Ben za n th r en e.
The formation of the ketyl radical (18) from benzanthrone
(6) may occur by RRD with 9,10-dihydroanthracene (3)
or the in situ formed benzanthrene (1), eqs 5-6, and
simultaneously producing 9,10-dihydroanthracenyl (19)
or benzanthrenyl (20) radicals (Scheme 4). Benzanthrene
(1) and not benzanthrenol (9) is the product of hydroge-
nation suggesting a fast dehydroxylation under the
applied experimental conditions (see below).
For the hydrogen transfer to afford the intermediate
9, three reaction channels are available: hydrogen
abstraction from 3 or 1 (eqs 7-8) or disproportionation
with 19, eq 9. The thermokinetic features of eqs 5-9 need
to be established to assess their relative contributions.
The H/D isotope-exchange experiments have shown
that the concentration of 9,10-dihydroanthracenyl radical
(19) can be obtained from the equilibration between 3
and anthracene (11):
-1.4
5,6-dihydro-4H-benz[d,e]-
anthracene (6-H) (7)
5,6-dihydro-4H-benz[d,e]-
anthracen-4-yl (26)
83.2
45.8
7H-benz[d,e]anthracen-7-ol (9)
anthrone (14)
xanthene
benzophenone ketyl
anthrone ketyl
xanthone ketyl
anthraquinone ketyl
fluorenone ketyl
benzanthrone ketyl (18)
9,10-dihydro-anthrol-9-yl
59.4
71.7
72.3
38.5
36.9
34.9
45.0
42.1
44.4
43.6i
75.5 ( 1.5d
41.6 ( 1.5h
3.2
3.1
3 + 11 h 19 + 19
(10)
Reference 13a. Reference 13b. c Reference 13c. Reference
a
b
d
2b. Reference 2c. f Reference 2d. g Reference 2a. Reference 13d.
e
h
The benzanthrenyl radical concentration (20) is dic-
tated by a similar equilibrium:
i
BDE for the C10-H bond in the radical.
abstractions (eqs 7-8) are relatively slow due to the
endothermicity of these reactions. Moreover, the reverse
of eq 6 is at least 2 orders of magnitude faster than the
product-forming step.15 This implies that the first step
(RRD) in the hydrogenation of benzanthrone with 1 is
reversible under these conditions. The rate expression
1 + 11 h 20 + 19
(11)
From the product distributions in Table S1 and S2
(Supporting Information), together with the enthalpic
data from Table 3, average concentrations for 19 and 20
at 300 °C can be estimated as 1.2 × 10-6 and 4.7 × 10-4
M, respectively.14 The kinetic analysis reveals that dis-
proportionation with 9,10-dihydroantracenyl (eq 9) is the
most important product-forming step; the hydrogen atom
(14) With ∆10S ) ∆11S ) 0 and using ∆10H ) 32.3, and ∆11H ) 22.5
kcal mol-1 from Table 3. At 300 °C, K10 is 4.8 × 10-13, and combined
[11] ) 0.9 M,8a and [3] ) 3.4 M,8a yields [19] )1.2 × 10-6 M; with K11
) 2.6 × 10-9, and [1] ) 0.24 M8a provides [20] ) 4.7 × 10-4 M.
(15) The kinetic parameters suggested by Savage have been used
for hydrogen atom abstraction reactions: A ) 108 M-1 s-1(per
hydrogen) Ea ) 16 + 0.35∆rH (exothermic), and Ea ) 16 + 0.65 ∆rH
(endothermic) kcal mol-1, with ∆rH as the reaction enthalpy.16 With
k7 ) 3.8 × 10-2 (∆7H ) 15.8 kcal mol-1), k8 ) 5.1 (∆8H ) 6.0 kcal
mol-1), and estimating that k9 ) k-6 ) 108 M-1 s-1 at 300 °C combined
with the concentrations of 3, 1, 19, and 20 in refs 8a and 14, provides
the relative rates of v7/v8 /v9/v-6 ) 1.1 × 10-3:1.0 × 10-2:1:250.
(12) Ebenhoch, J . Ph.D. Dissertation, Universita¨t Freiburg, 1996.
(13) (a) Tsang, W. In Energetics of Organic Free Radicals; Simo˜es,
J . A. M., Greenberg, A., Liebman, F., Eds.; Black Academic and
Professional: London, 1996; pp 22-58. (b) Meot-Ner, M. J . Am. Chem.
Soc. 1982, 104, 5-10. (c) Laarhoven, L. J . J .; Born, J . G. P.; Arends, I.
W. C. E.; Mulder, P. J . Chem. Soc., Perkin Trans. 2 1997, 2307-2312.
(d) Arnaut, L. G.; Caldwell, R. A. J . Photochem. Photobiol, A: Chem.
1992, 65, 1992, 15-20.

6616 J . Org. Chem., Vol. 66, No. 20, 2001
Mulder et al.
Sch em e 4
ficiently fast. We can estimate that with the inherently
elevated concentration of benzanthrenyl (see above),
which is a direct consequence of the low BDE(C-H) in
1, a pseudo-first-order rate constant of 105 s-1 or higher
for the incipient radical species is required to circumvent
the reversibility. These findings also explain why the rate
for reductive cleavage of R-phenoxyl-acetophone (see
Introduction) is accelerated only modestly when adding
1 to 3. The actual RRD by 1 is faster (eq 3) but the
following fragmentation of the ketyl radical into phenoxyl
and acetophenone (eq 4) is too slow to compete effectively
with the back hydrogen atom transfer to the benzanthre-
nyl radical. In the past, we have estimated a rate
constant for eq 4 as about 106 M-1 s-1 24
Now, an upper
.
for disappearance of benzanthrone (6) can now be pre-
limit of 104 M-1 s-1 seems to be more in line with the
current insights and can be related to the high intrinsic
activation barrier for the reverse of eq 4, the addition of
a phenoxyl radical to a double bond.25
sented as:
-d[6]/dt ) [k9K6K10/K11][3][6]
(12)
With the enthalpic data from Table 3, eq 12 is reduced
Hyd r ogen a tion of Ben za n th r en e. The hydrogena-
tion of benzanthrene (1) follows a second order (in 1)
reaction mechanism. At the high temperatures employed,
an appreciable radical concentration can be envisaged to
establish a dynamic equilibrium between 1 and its radical
(20). Delocalization of the unpaired electron26 in 20
creates a number of reactive sites and hydrogen transfer
yields an isomeric mixture of 7H-benz[d,e]anthracene (1),
6H-benz[d,e]anthracene (21), 4H-benz[d,e]anthracene (22),
3H-benz[d,e]anthracene (23), and 1H-benz[d,e]anthracene
(24). It seems reasonable to postulate that the concentra-
tions of the isomers are established by an overall
thermodynamic equilibrium (eq 14, Scheme 5), which is
corroborated by the degree of H/D exchange in benzan-
threne when a deuterated solvent is used.
Provided that the entropy changes are negligible, the
relative concentrations can be approximated by using the
benzylic BDE(C-H)s from Table 3 as 1/21/22/23/24 ) 3.5
× 104:3.9 × 102:9.4 × 102:1:5.8 at 300 °C. This ratio
agrees with the general feature that phenanthrene
derivatives are thermodynamically more stable than the
isomeric compounds with a (linear) anthracene backbone.
Compound 7, the product of hydrogenation of 1, emerges
through two distinct RRD reactions: hydrogen transfer
to -d[6]/dt )1.8 × 10-12k9[3][6].17 Combination with the
experimentally determined kexp,6 of 7.1 × 10-5 M-1 s-1
,
furnishes a rate constant for disproportionation of 9,10-
dihydroanthracenyl (19) with ketyl 18 of k9 ) 3.9 × 107
M-1 s-1at 300 °C.
In the absence of significant steric interactions, the rate
for disappearance of carbon-centered radicals by radical-
radical interaction is at the diffusion-controlled limit. The
rate constant (kdis) merely reflects the transport proper-
ties of the solvent rather than the nature of the encoun-
tering species.18,20 Typically, kdis is about 2 × 1010 M-1
s-1 at 300 °C. The two product channels are recombina-
tion (kr) and disproportionation (kd); the ratio (kd/kr)
strongly depends on the type of radicals involved.21
linear relationship between ln(kd/kr) and the reaction
enthalpy for disproportionation (∆dH) has been sug-
gested, which can be rewritten as eq 13.21,22
A
ln(kd) ) 13.6 - 0.16∆dH
(13)
With ∆9H of -16.5 kcal mol-1, eq 13 yields a k9 () kd)
of 1 × 107 M-1 s-1, in good agreement with our experi-
mental value of 3.9 × 107 M-1 s-1
.
The reaction entropy for disproportionation is close to
zero, and on the basis of microscopic reversibility, eq 13
can also be used to calculate the preexponential factor
for the RRD, kRRD ) kdexp(-Ea,RRD/RT). This implies that
the variation in the reactivity of a hydrogen donor cannot
be derived exclusively on an activation enthalpy consid-
eration.23
(23) Using eq 13 to derive the ARRD implies that both ARRD and Ea,RRD
depend on the reaction enthalpy. This phenomenon is commonly
referred to as the isokinetic relationship but may also be imposed by
a
small enthalpic barrier for disproportionation (Evans-Polanyi
principle). Indeed, the factor of 0.16 is quite common for exothermic
hydrogen atom abstraction reactions.20 The rate constant for RRD is
presented by kRRD ) ARRDexp(-Ea, RRD/RT); with ln(ARRD) ) ln(kd) )
13.6 + 0.16∆RRDH and Ea,
) ∆RRDH + 3 with ∆RRDH the reaction
RRD
enthalpy for RRD and a diffusional enthalpy barrier of 3 kcal mol-1
Thus, benzanthrene can be applied as a hydrogen
transfer agent only when the subsequent step is suf-
for the reverse reaction.18
(24) Huang, Y.; Page´, D.; Wayner, D. D. M.; Mulder, P. Can. J .
Chem. 1995, 73, 2079-2085.
(16) Savage, P. E. Energy Fuels 1995, 9, 590-598.
(25) (a) In this particular experiment, the 1/3 ratio was 1/16,5 and
according to ref 23 the relative v3,RRD is expected to be 70. Experimen-
tally, a rate enhancement of only a factor of 5 was found. (b) With a
(17) Assuming ∆6S ) 0 and with ∆6H ) 21.0 kcal mol-1 from Table
3, renders K6 ) 9.8 × 10-9 at 300 °C (see also ref 14).
(18) The kdis for, e.g., cumyl radicals, measured in tert-butylbenzene
∆4H of 12 kcal mol-1 24 log(A4) ) 13.4,25c and a rate constant, k4, of 104
,
as the solvent, is determined as 4 × 1011exp(-3.4/RT), i.e., at 300 °C
s-1 leads to Ea4 ) 25 kcal mol-1, thus Ea-4 ) 13 kcal mol-1 the reverse
reaction: the addition of the phenoxyl radical to the styrene derivative
(the enol form of acetophenone). This high activation barrier for
phenoxyl may be associated with the extended delocalozation of the
free electron into the aromatic ring. In the recombination reaction with
methyl, primarily the carbon-carbon coupling occurs,25d addition of a
phenoxyl to styrene with the spin on carbon yields an adduct with a
weak carbon-carbon bond, which will decompose to reactants. (c)
Autrey, S. T.; Alnajjar, M. S.; Nelson, D. A.; Franz, J . A. J . Org. Chem.
1991, 56, 2197-2202. (d) Arends, I. W. C. E.; Louw, R.; Mulder, P. J .
Phys. Chem. 1993, 97, 7914-7925.
kdis ) 2 × 1010 M-1 s-1
.
The temperature dependence for a diffusion-
19a
controlled rate constant is associated with the change in the solvent’s
viscosity (η). The viscosity of pure anthracene is 0.39 cP at 300 °C.19b
(19) (a) Korthe, T.; Marque, S.; Martschke, R.; Popov, M.; Fischer,
H. J . Chem. Soc. Perkin Trans 2, 1998, 1553-1559. (b) Landolt-
Bo¨rnstein, 6. Auf. Zweiter Band, 5. Teil, Bandteil a (Transportpha¨-
nomene I); Scha¨fer, K., Ed; Springer-Verlag: Berlin, Heidelberg, New
York, 1969; p 174.
(20) Arends, I. W. C. E.; Mulder, P.; Clark, K. B.; Wayner, D. D. M.
J . Phys. Chem. 1995, 99, 8182-8189.
(21) Manka, M. J .; Stein, S. E. J . Phys. Chem. 1984, 88, 5914-5919.
(26) According to our DFT calculations, the Mulliken spin densities
at positions C-7, C-6, C-4, C-3, and C-1 in the benzanthrenyl radical
(20) are 0.29, 0.23, 0.23, 0.10, and 0.15, respectively, in good agreement
with an EPR study to provide 0.31, 0.22, 0.21, 0.13, and 0.13.27
(22) Equation 13 is derived from ln(kd/kr) ) -10.1 - 0.16∆dH.21 Since
kdis kr > kd, it can be rewritten as ln(kd) ) -10.1-0.162 ∆dH + ln-
≈
(kdis) ) 13.6 - 0.16∆dH, with kdis from ref 18.

Thermal Reduction of 7H-Benz[d,e]anthracen-7-one
J . Org. Chem., Vol. 66, No. 20, 2001 6617
Sch em e 5
Sch em e 7
Sch em e 6
Sch em e 8
from 1 (eq 15) or 21 (eq 16) to the unsaturated moiety
(Scheme 6).28 Hydrogen abstraction by the intermediate
radicals from, e.g., 3, yields 7 and recycles 1. With kexp,1
of 3.81 × 10-4 M-1 s-1, and the equilibrium concentration
for 21, the RRD rate constants for eq 15 and eq 16 can
now be derived as 1.0 × 10-2 and 0.26 M-1 s-1, respec-
tively. There is no way to differentiate between these two
pathways, since the concentrations of 1 and 21 are
equilibrated. Interestingly, as a result of the hydrogena-
tion of 1, the RRD capacity is greatly diminished. The
BDE(C-H) of 83.4 kcal mol-1 in 7 creates a donor solvent
which is as (un)reactive as tetralin.
presence a weak carbon-oxygen bond. However, an
assessment of the reaction rates shows that the contribu-
tion of this pathway is negligible.29
Both routes include the formation of a carbocation.
These species may be trapped when an electron-rich
aromatic scavenger, e.g. phenol, is added to the reaction
medium. In this way, we have demonstrated that a
carbocation is the reactive intermediate in the conversion
of o-hydroxybenzylic alcohol into o-hydroxytoluene.7 How-
ever, adding phenol in about a 1:1 ratio to 3 in the
hydrogenation of 6 did not yield the expected electrophilic
addition products; it merely served as an inert diluent.
Stock arrived at this mechanistic rationale for the
conversion of benzophenone via benzhydrol to diphenyl-
methane by studying the reactivity of benzhydrol under
various conditions. Ether formation is a bimolecular rate-
determining reaction, and occurs at an appreciable rate
only when a fairly high concentration of the alcohol is
The isomers 21 to 24 can be regarded as derivatives
of phenalene, 26. The self-hydrogenation by RRD of 26
follows the same mechanism with a reported rate con-
stant of 9.8 × 104exp(-17.9/RT) or 1.45 × 10-2 M-1 s-1
at 300 °C.1a
The RRD enthalpy change for eq 1628 (15.9 kcal mol-1
)
is somewhat higher compared to that between two
phenalene molecules (13.3 kcal mol-1), while the rate
constant for reaction 16 derived in this work is about 20
times higher. With phenalene, the agreement between
the computed and experimental enthalpy of reaction is
quite satisfactory (13.3 vs 14.9 kcal mol-1). The discrep-
ancy is mainly caused by the reported and unexpected
low A factor, while according to eq 13 a value of 7 × 106
M-1 s-1 is predicted. A low experimental A-factor may
arise when the kinetic measurement is obscured by a
reversibility of the investigated reaction.
Con ver sion of Ben za n th r en ol in to Ben za n th r en e.
So far, it has been postulated that the deoxygenation of
benzanthrenol (9) into benzanthrene (1) is a fast process.
For this transformation two mechanistic options are
feasible: a hydride transfer (Scheme 7) mechanism or
the ether mechanism as proposed by Stock (Scheme 8).6
An alternative route for the decomposition of ether 29
(Scheme 8, eq 22) may be the direct homolysis due to the
(29) (a) The ether 29 may also decompose into 20 and an alkoxyl
radical which by a fast elimination of a hydrogen atom yields the initial
ketone, 6. According to the our DFT methodology, the BDE(C-H) for
the secondary carbon in propane is computed as 97.1 kcal mol-1. By
comparing this with the BDE(C-H) in 1 of 65.4 kcal mol-1 from Table
3, a resonance stabilization enthalpy (RSE) of 31.7 kcal mol-1 is
obtained. The BDE(C-O) in diisopropyl ether can be derived as 85.9
29b
kcal mol-1
.
Applying the RSE, the BDE(C-O) in 29 becomes 54.2
kcal mol-1. Subsequently, with
homolysis in the liquid phase of 1015.2 s-1
a
preexponential factor for bond
29c and with the BDE(C-O)
,
as a measure for the enthalpy of activation, a rate constant for
homolysis of 29 is computed as 3.3 × 10-6 s-1 at 300 °C. The pseudo
first-order rate constant for eq 22 is given by k22[19]. According to refs
14 and 15, [19] ) 1.2 × 10-6 M, and k22 ) 2 × 104 M-1 s-1 [∆22H )
-15.8 kcal mol-1 from Table 3, and using the BDE(C-H) of 9] provides
k22[19] ) 2.4 × 10-2 s-1. Hence, the rate of disappearance of 29 will
be entirely governed by eq 22. (b) Afeefy, H. Y.; Liebman, J . F.; Stein,
S. E. Neutral Thermochemical Data. in NIST Chemistry WebBook,
NIST Standard Reference Database Number 69; Linstrom P. J .,
Mallard, W. G., Eds.; J uly 2001, National Institute of Standards and
Technology, Gaithersburg MD, 20899 (http://webbook.nist.gov). (c)
Pratt, D. A.; Heer de, M. I.; Mulder, P.; Ingold, K. U. J . Am. Chem.
Soc. 2001, 123, 5518-5526.
(27) Lewis, I. C.; Singer, L. S. Magn. Reson. Chem. 1985, 23, 698-
704.
(28) In our calculation we have assumed that compounds 21 and
22 are (thermodynamically) identical, and that the ratio 1/(21 + 22)
) 1/21 ) 26, with an average BDE(C-H) for 21 of 61.7 and ∆16H )
15.9 kcal mol-1
.

6618 J . Org. Chem., Vol. 66, No. 20, 2001
Mulder et al.
Sch em e 9
present. We were unable to detect any significant amount
of benzanthrenol (9), which, combined with the fact that
the rate of conversion of 6 is independent of its initial
concentration, disfavors the ether mechanism under our
reaction conditions.
The nature of the acidic species to create the carboca-
tion remains unclear, the presence of (nonsolvated)
protons in a nonpolar environment is counter intuitive.
Ketyl radicals in solution are known to be more strongly
acidic species than the corresponding alcohols. Further,
a substantial difference in pKa (in water) values has been
found for alkyl and aryl ketyls; for example, for m-
hydroxy benzophenone ketyl and dimethyl ketyl, Me2C-
(•)OH, pKas of 6.5 and 12, respectively, have been
reported.30 The variation of almost 6 pKa units can be
associated with the better delocalization of the oxygen’s
lone pair due to the aromatic system. The ketyl derived
from benzanthrone is probably an even more acidic
species, in accordance with our observation that benzyl
alcohol when reacted in the presence of 6 is dehydroxy-
lated for about 50%.31 The addition of a strong base
(pyridine) is expected to reduce substantially the con-
centration of free ketyl as a result of deprotonation or
hydrogen bonding. Consequently the overall reduction
rate should be retarded.32 This, however, appears not to
be the case. Clearly, more experimental work is required
to unravel the nature of the putative reactive acidic
species present in the reduction of aromatic ketones.
the degree of conversion of 1. By addition of pyridine,
this ratio decreases to three, with a concomitant increase
in the overall conversion of 6 from 69 to 82%. In that
case, the ratio of the intermediate compounds 9/(31 +
32) will be shifted in favor of the thermodynamically
more stable hydrogen-bonded phenolic compounds.35
These results strongly indicate that formation of 31 is
reversible under standard reaction conditions. The rate
constant for hydrogenation of the olefinic moieties in 31
by benzanthrene (eq 26) is most likely identical to the
one determined for the hydrogenation of 21 into 7 (eq
15): i.e., 1 × 10-2 M-1 s-1. From our data (see Table 1), it
can be inferred that, at longer reaction times, the
apparent rate of hydrogenation of 6 is retarded. This is
probably the consequence of the build-up of anthracene,
which enhances the reformation of the ketyl radical, by
a RRD with 9 or 31. Thus, in the presence of large
amounts of anthracene, both reactions 9 and 25 are
reversible. Indeed, by addition of anthracene to the
reaction mixture, the product ratio increases to 57 and
the conversion of 6 drops to 24%. We have used these
data to compute the global rate constant for the conver-
sion of 9 into 1 (eq 28).
That 9 is an intermediate in the hydrogenation of 6
can be concluded from the formation of the phenolic
byproduct, 10. Analogously to the reduction of 1 into 7,
the disproportionation of the ketyl radical (18) yields a
number of distinct isomeric compounds: 9, 6H-benz[d,e]-
anthracen-7-ol (31) and 4H-benz[d,e]anthracen-7-ol (32)
(Scheme 9).33
The BDE(C-H) in 9 is lower than in 31 or 32, and a
thermal equilibration would render a ratio of 9/31/32 of
1:2.2:5.3, implying that the phenolic intermediates pre-
dominate under these conditions.34 The final product ratio
(1 + 7)/10 remains rather constant at around 30 and is
independent of the initial concentrations of 6, 3 and 1 or
9 f 1
(28)
(30) Bhasikuttan, A. C.; Singh, A. K.; Palit, D. K.; Sapre, A. V.;
Mittal, J . P. J . Phys. Chem. A 1999, 103, 4703-4711.
(31) The acidity of such a reaction mixture is underscored by the
observation that benzyl alcohol when dissolved in 9,10-dihydroan-
thracene appeared to be completely inert at 300 °C and 60 min reaction
time. However in the presence of 6 (ca. 0.3x the amount of alcohol),
50% dehydroxylation of benzyl alcohol into toluene occurred, while the
reactivity of 6 was not affected.
(32) (a) For example, benzophenone ketyl forms a strong hydrogen
bond with triethylamine,32b a solvent with the same hydrogen bond
accepting properties as pyridine.32c The reported equilibrium constant
at 25 °C for a 1:1 complex between benzophenone ketyl and triethy-
lamine is around 100 M-1. (b) Abe, T.; Kawai, A.; Kajii, Y.; Shibuya.
K.; Obi, K. J . Phys. Chem. A 1999, 103, 1457-1462. (c) Abraham, M.
H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J . J .; Taylor, P. J .
J . Chem. Soc., Perkin Trans. 2, 1989, 699-711. (d) Recently we have
shown the dramatic effect of hydrogen bonding on hydrogen atom
abstractions: Snelgrove, D. W.; Lusztyk, J .; Banks, J . T.; Mulder, P.;
Ingold, K. U. J . Am. Chem. Soc. 2001, 123, 469-477.
For an fully equilibrated system, the product ratio (1
+ 7)/10 is given by {(k28K9)/(k26K25)[1]), and substitition36
affords a value k28 ) 0.43 s-1 at 300 °C, which demon-
strates that the reverse reactions will be significant even
at low anthracene concentrations.37 The above analysis
underlines the complex dynamics of the reduction by a
hydrogen-donor solvent.
(35) With the hydrogen bond acidities and basicities from Abraham’s
list, the equilibrium constant for the hydrogen bonding between
pyridine and 1-naphthol (a model for 31 or 32) is calculated as 58 at
25 °C or 8 M-1 at 300 °C.32c With 3 M of pyridine in the reaction
mixture, the ratio between the hydrogen bonded complex and the free
phenol is 24 at 300 °C.
(33) According to our DFT calculations, the Mulliken spin densities
at positions C-7, C-6, C-4, C-3, and C-1 in the ketyl radical 18 are 0.29,
0.26, 0.25, 0.10, and 0.10, respectively, quite similar to the spin
distribution in the benzanthrenyl radical.26
(36) Under conditions of full equilibration for reactions 9 and 25,
the product ratio (1+7)/10 is given by [(k28K9)/(k26K25)[1]]. With ∆9S )
∆
25S ) 0, ∆9H ) -16.5, ∆25H ) -18.8 kcal mol-1 (see ref 34) to give
K9 ) 1.96 × 106 and K25 ) 1.48 × 107. With k26 ) k15 ) 1 × 10-2
,
(34) It has been assumed that the BDE(C-H)s in 31 and 32 are
identical to the ones in 21 and 22, since the substituent effect (by OH)
on benzylic C-H bonds is negligible.13c The ratio is computed assuming
no difference in entropy between the species, and with the enthalpic
data from Table 3. In our calculation we have assumed that compounds
31 and 32 are (thermodynamically) identical, with a (31 + 32)/9 )
31/9 ) 7.5, and an average BDE(C-H) for 31 of 61.7 kcal mol-1 and,
M-1s-1, combined with [1] ) 0.10 M (see Table S2, exp. 2.11) and (1+7)/
10 ) 57 (see Table S2, experiment 2.11) yields k28 ) 0.43 s-1
.
(37) For example, with [1] of 0.24 M,8a and k26 ) 1.0 × 10-2 M-1
s-1, the rate of hydrogenation of 31 is about 30 times slower than the
RRD with anthracene (11) with a k-25 of 0.08 M-1 s-1 and [11] ) 0.9
M.8a For the RRD of 9 with 11, with k-9 ) 0.41 M-1s-1, a pseudo-first
rate constant of 0.37 s-1 is calculated, which is comparable with k28
0.43 s-1, the global rate constant for formation of 1 from 9.36
)
hence, ∆25H ) -18.8 kcal mol-1
.

Thermal Reduction of 7H-Benz[d,e]anthracen-7-one
J . Org. Chem., Vol. 66, No. 20, 2001 6619
solved in toluene and 2-methoxynaphthalene was added as an
external standard. The experimental temperature was mea-
sured by means of a thermocouple placed through a septum
into a Pyrex tube filled with n-eicosane, in proximity of the
reaction tubes, and followed by a thermogram. The starting
time was taken from the moment a value 5 °C below the final
temperature was reached. The overall time to reach the
reaction temperature was about 5 min. The temperature
stability was (1 °C. Typically, the reproducibility of the high-
temperature liquid-phase experiments was around 2-4%.
The reaction model as outlined for benzanthone may
well be valid for other aromatic ketones. If the relative
rates are dictated by the variation in ∆RRDH only, the
predicted relative rates are shown in Table 3. By compar-
ing the experimental and predicted values, it becomes
obvious that in this way the differences in reactivity of
the various aromatic ketones can be explained only in a
semiquantitative fashion.
An a lysis. The analyses were performed on an HP 5890A
gas chromatograph (CP-Sil-5-CB capillary column, 50 m, 0.32
mm i.d., 0.4 µm film thickness, and hydrogen as the carrier
gas). Unknown products were identified on a GC-MS (HP
5890A-HP 5972) with a similar column and helium as the
carrier gas.
Con clu sion
The selective reduction of aromatic ketones into the
hydroaromatic compounds by hydrogen donor solvents
has been rationalized by reverse radical disproportion-
ation (RRD) reactions. The dehydroxylation of the inter-
mediate alcohol may occur through a proton transfer from
the acidic ketyl radical.
With the aid of DFT-computed enthalpies, providing
a consistent set of thermodynamic data, the dynamics of
the main reaction pathways for the reduction of benzan-
throne have been quantified. Although exhibiting favor-
ably low C-H bond dissociation enthalpies, application
of benzanthrene and analogues as hydrogen transfer
agents may suffer from reversibility and self-hydrogena-
tion.
Quantitative analyses were performed with GC-response
factors obtained from calibration with authentic samples or
standard estimation methods.
Syn th esis of 4,5-Dih yd r o-6H-ben z[d ,e]a n th r a cen e (7).
A Pyrex tube of 15 mL was charged with 1.4 g (6 mmol) of 6
and 2.6 g (20 mmol) tetralin and subsequently evacuated and
sealed as described above. The tube was placed in a thermostat-
controlled oven at T ) 365 °C for 3 h, cooled to room
temperature, and decapped. The reaction mixture was trans-
ferred with toluene, and the solvents including naphthalene
were removed by vacuum distillation (the apparatus was
heated with an IR lamp). The raw product was purified by
sublimation, yielding 1.0 g (4.6 mmol, 75%) of 7 (purity 97%
as determined by GC). 1H and 13C NMR spectra (300 MHz) in
(solvent CDCl3) were recorded with a Bruker DPX300.
Exp er im en ta l Section
Ma ter ia ls. Compound 1 was synthesized from 6 by reduc-
tion with LiAlH4 according to a literature procedure.38 Puri-
fication by sublimation afforded a compound with 94.5% purity
(the remaining 5.5% was 6). Commercially available 6 was
sublimed twice. The raw material has a purity of ca. 90%, the
nature of the impurity could not be identified, since the GC
analysis did not show any other compound. However, upon
dissolving 6 in, e.g., toluene, it could be noticed that some black
(carbonaceous) insoluble material was present. The color of
purified 6 is bright yellow. All other chemicals were com-
mercially available and were purified to g99% by distillation
or sublimation. The applied 9,10-dihydroanthracene contained
5% of anthracene.
Liqu id -P h a se Exp er im en ts. The procedure has been
described in detail previously.5 In short, a Pyrex tube of ca. 2
mL was charged for about one-fourth (in volume) with sub-
strate, hydrogen-donor solvent, and dibenzofuran as an inter-
nal standard. Prior to sealing of the tube, air was replaced by
helium in three freeze-pump-thaw cycles. After heating in
a thermostat-controlled (GC) furnace, the tube was cooled
rapidly and decapped. Then, the reaction mixture was dis-
Den sity F u n ction a l Th eor y Ca lcu la tion s. The calcula-
tions were performed using Gaussian 98 rev. A.7 package on
an IBM RS6000 computer. The B3LYP method with 6-31G-
(d,p) basis set was employed for the geometry optimization and
the frequency routines. The ZPVE were scaled by a factor of
0.9805, and the enthalpy of the species were computed at T )
298 K.
Ack n ow led gm en t. The authors wish to thank F.
Lefebre for recording all NMR spectra and the Hoch-
schulrechenzentrum Essen for the generous allotment
of computer resources.
Su p p or tin g In for m a tion Ava ila ble: Tables containing
data for the temporal conversion and product distribution
under various conditions for 6 (Tables S1 and S2) and the
temporal conversion of 7 (Table S3). NMR (simulation) data
for 7, Cartesian coordinates and electronic energies of the DFT
computations. This material is available free of charge via the
Internet at http://pubs.acs.org.
(38) Gerst, M.; Morgenthaler, J .; Ru¨chardt, C. Chem. Ber. 1994, 127,
691-696.
J O015637V