Mulder et al.
SCHEME 11
23 the lowest energy structure of the hydrogen atom
adduct appears to be a π-complex with the chlorine atom
having shifted from C-2 in the parent compound to C-1
in the complex. The corresponding HA values are -35.0
kcal mol-1 for 22 and -34.5 kcal mol-1 for 23, suggesting
an almost identical reactivity of both isomers, in clear
contrast with the HA values of -31.3 and -26.3 kcal
mol-1 for C-1 and C-2 in naphthalene.20
When comparing 3b (without chlorine) and 6b (with
chlorine), the computed HA decreases from -40.5 to
-54.5 kcal mol-1. However, the DFT calculation predicts
for the 6b + H• reaction as a lowest energy product not
the cyclohexadienyl radical 3-σ-Cl but rather a chlorine-
arene π-complex 3-π-Cl (Scheme 11) with a chlorine-
arene distance of about 2.5 Å. A similar situation holds
for all 2- and 4-chloro-substituted tautomers. The differ-
ence in ∆fH between the π and σ chlorine-benzene
radical adduct has been estimated to be around 7 kcal
mol-1 in favor of the σ species.22 Apparently, DFT
identifies a post-transition-state complex as the lowest
energy ensemble, stabilized by π-type interaction (and
additional hydrogen bonding between Cl and the adjacent
hydroxyl or amine group for the 2-chloro-substituted
systems).23 Although obscured by this characteristic of
the DFT method, the HA values for the enones a -2 of all
2-substituted phenols studied are almost invariant (4,
-58.8 kcal mol-1; 7, -58.1 kcal mol-1; 8, -58.3 kcal
mol-1; 9, -58.3 kcal mol-1). Furthermore, the change in
HA between 4a -2 and 6b is about the same as for the
two tautomers 3a , 3b of phenol. Similar to the chlori-
nated phenols and anilines, the computed HA values for
the carbon carrying the chlorine in 1-chloro- and 2-chlo-
ronaphthalene (22, 23) refer to π-complexes, with the
chlorine atom interacting preferably with C-1. Thus, for
Discu ssion
The foregoing results show that chlorines attached to
the 2- or 4-position in aromatic alcohols or amines, e.g.
phenols or anilines, can be removed with a remarkably
high selectivity in the presence of hydrogen donor
solvents such as 9,10-dihydroanthracene (1). Under
similar conditions elimination of Cl at C-3 does not occur.
At elevated temperatures and in the presence of 1 and
(in situ formed) anthracene (24) an appreciable station-
ary concentration of the 9,10-dihydroanthracenyl radical
(2) is present in the reaction mixture (see Scheme 4).7
Since, in principle, the thermal decomposition of 2 may
act as a source for hydrogen atoms, it might be postulated
that these species are responsible for the observed
hydrodechlorination. Earlier, it has been demonstrated
that in the gas phase and at around 1000 K, the
hydrodechlorination rate for chlorophenols by an addi-
tion/elimination mechanism is exclusively governed by
the rate of the addition of the hydrogen atom. Relative
rates for 4, 5, and 6 have been established as 1.6:1:1.5.1b
This hydrogen atom chemistry leads to the elimination
of other substituents as well. According to gas-phase
kinetic parameters, the relative rates for hydrodesubsti-
tution of chlorobenzene, phenol, and aniline are quite
comparable (1:0.3:0.4 at 630 K).24 In the present study,
however, chlorobenzene from chlorophenols and chloroa-
nilines or 1-chloronaphthalene (22) from 36 was never
observed as a product. Therefore, it can be concluded that
free hydrogen atom chemistry is not of any importance
for the present liquid phase reaction conditions and that
RRD is most likely the ruling mechanism. The rate of
dechlorination of the chloroaromatics is dramatically
increased as a result of the introduction of an OH or NH2
group. Besides, the product distribution is quite different.
From, e.g., 4-chloro-1-naphthylamine (36) only 1-naph-
thylamine (35) is found, whereas with 1-chloronaphtha-
lene (22) the formation of an array of condensation
products predominates. All these findings indicate that
tautomerization prior to RRD plays a pivotal role in the
hydrodesubstitution mechanism for a defined category
of aromatic compounds. In the following sections we will
discuss some dynamic aspects.
(21) (a) The bond dissociation enthalpies (BDE) of the O-H and
N-H bonds for the parent compounds have been computed as well by
our DFT method, and they are presented in Table 5 as scaled BDEs,
using the experimental values for the BDE(O-H) in phenol and the
BDE(N-H) in aniline as references. Although the BDE(O-H) in 3,
which is experimentally determined as 87.2 kcal mol-1 (ref 21b), is
underestimated by DFT at this level of theory by about 4 kcal mol-1
,
the computed change in BDE(O-H), ∆BDE(O-H), upon substitution,
is quite independent of the methodology applied and agrees with the
experiment (ref 21c). This is reinforced by the present work: the
experimental and DFT-computed ∆BDE(O-H) values for 1-naphthol
(33) are almost identical (-5.9 versus -6.2 kcal mol-1) (ref 21d). For
10 DFT predicts a BDE(N-H) of 87.9, which is 2 kcal mol-1 below the
best experimental value (ref 21e). As for the phenols, it may be
assumed that the ∆BDE(N-H) values for the aromatic amines are
sufficiently reliable (ref 21f). The change in BDE(O-H) or BDE(N-
H) in phenyl derivatives upon ring substitution can be mainly
attributed to the degree of delocalization of the unpaired electron in
the corresponding phenoxyl and anilinyl radicals (refs 20c and 20f).
This also appears to be valid when expanding the aromatic system
from phenyl to anthracenyl. The effect is more profound for O-H than
for N-H bonds, because of the higher intrinsic spin delocalization in
aromatic oxyl than in aminyl radicals. (b) Wayner, D. D. M.; Lusztyk,
E.; Page´, D.; Ingold, K. U.; Mulder, P.; Laarhoven, L. J . J .; Aldrich, H.
S. J . Am. Chem. Soc. 1995, 117, 8737-8744. (c) Pratt, D. E.; De Heer,
M. I.; Mulder, P.; Ingold, K. U. J . Am. Chem. Soc. 2001, 123, 5518-
5526. (d) Bordwell, F. G.; Cheng, J .-P. J . Am. Chem. Soc. 1991, 113,
1736-1743. (e) MacFaul, P. A.; Wayner, D. D. M.; Ingold, K. U. J .
Org. Chem. 1997, 62, 3413-3414. (f) Pratt, D. A.; Dilabio, G. A.;
Valgimigli, L.; Pedulli, G. F.; Ingold, K. U. J . Am. Chem. Soc. 2002,
124, 11085-11092.
Hyd r od ech lor in a tion : Rela tive Ra tes. A general
expression has been developed for the rate constant for
RRD, kRRD, and requires only the knowledge of the
reaction enthalpy, ∆RRDH (eq 1), with ln(ARRD) ) 13.6 +
kRRD ) ARRD exp[- (∆RRDH + 3)/RT]
(1)
0.16∆RRDH, and ∆RRDH in kcal mol-1 6
predicts that when the hydrogen transfer becomes less
.
This equation
(22) Benson, S. W. J . Am. Chem. Soc. 1993, 115, 6969-6974.
(23) At this level of theory we noted a similar deviation for the
hydrogen abstraction by an alkoxyl radical from phenol, RO• + PhOH
f ROH + PhO•, DFT assigns a hydrogen bonded post transition-state
complex as the lowest energy product instead of the two separated
species: De Heer, M. I.; Mulder, P.; Korth, H.-G.; Ingold, K. U.;
Lusztyk, J . J . Am. Chem. Soc. 2000, 122, 2355-2360.
(24) Mallard, W. G.; Westley, F.; Herron, J . T.; Hampson, R. F.;
Frizzell, D. H. NIST Chemical Kinetics Database, version 2Q98; NIST
Standard Reference Data, National Institute of Standards and Tech-
nology: Gaithersburg, MD, 1998.
4254 J . Org. Chem., Vol. 68, No. 11, 2003