Oxidative activation in aromatic substitutions. Reactions of N,N- dimethylanilines with secondary anilines promoted by thallium triacetate

Article abstract of DOI:10.1021/jo982276j

The reactions of N,N-dimethyl-p-anisidine (1a), N,N-dimethylaniline (1b), and N,N-dimethyl-p-fluoroaniline (1c) toward secondary anilines 2(a- d)-H in the presence of thallium triacetate sesquihydrate have been studied as representative of a novel oxidatively activated aromatic substitution affording 1,4-benzenediamine derivatives 3a-d. All of the substrates considered gave substitution with diphenylamine (2d-H). However, with anilines 2b,c-H, only 1a underwent substitution, and substrates 1b,c were practically unreactive. The observed differences in reactivity are well accounted for within a mechanistic framework wherein oxidative activation of both the substrate and the secondary aniline is regarded as alternatively (or simultaneously) possible, depending on the redox characteristics of the reactants. For instance, it can be stated, beyond any reasonable doubt, that reactions of 1a with 2b,c-H proceed via nucleophilic attack of the latter on 1a^+., and that the reaction of 1a-c with 2d-H must involve the diphenylamino radical 2d·.

Full text of DOI:10.1021/jo982276j

J . Org. Chem. 1999, 64, 2459-2464
2459
Oxid a tive Activa tion in Ar om a tic Su bstitu tion s. Rea ction s of
N,N-Dim eth yla n ilin es w ith Secon d a r y An ilin es P r om oted by
Th a lliu m Tr ia ceta te
Francesco Ciminale,* Antonella Ciardo,† Salvatore Francioso,† and Angelo Nacci
Centro CNR di Studio sulle Metodologie Innovative di Sintesi Organiche, Dipartimento di Chimica,
Universita`, via Amendola 173, 70126 Bari, Italy
Received November 16, 1998
The reactions of N,N-dimethyl-p-anisidine (1a ), N,N-dimethylaniline (1b), and N,N-dimethyl-p-
fluoroaniline (1c) toward secondary anilines 2(a -d )-H in the presence of thallium triacetate
sesquihydrate have been studied as representative of a novel oxidatively activated aromatic
substitution affording 1,4-benzenediamine derivatives 3a -d . All of the substrates considered gave
substitution with diphenylamine (2d -H). However, with anilines 2b,c-H, only 1a underwent
substitution, and substrates 1b,c were practically unreactive. The observed differences in reactivity
are well accounted for within a mechanistic framework wherein oxidative activation of both the
substrate and the secondary aniline is regarded as alternatively (or simultaneously) possible,
depending on the redox characteristics of the reactants. For instance, it can be stated, beyond any
reasonable doubt, that reactions of 1a with 2b,c-H proceed via nucleophilic attack of the latter on
1a +•, and that the reaction of 1a -c with 2d -H must involve the diphenylamino radical 2d •.
Sch em e 1
In tr od u ction
Radical cations formed by single electron removal from
neutral molecules are common reactive intermediates in
a variety of thermal, photochemical, and electrochemical
reactions occurring under oxidative conditions.1-3 Among
other possible reactions, these electron-poor species are
known to undergo addition of nucleophiles. In fact,
contrary to some theoretical considerations4 leading to
the classification of direct nucleophilic attack on a radical
cation as “forbidden”, kinetic measurements have dem-
onstrated that this process can be very rapid, even
approaching diffusion control.5
adduct I as a radical (X•) giving the substitution product
ArNu (path i).
Diverse aromatic substitutions affording either hydro-
gen or ipso substitution products have been described as
involving the addition of nucleophiles to unsubstituted
or substituted positions of aromatic radical cations,
respectively. Unlike conventional aromatic substitutions
taking place on neutral substrates, in which the replaced
group invariably keeps the electronic identity of the
attacking species, a characteristic feature of these reac-
tions is the possibility of three different electronic identi-
ties for the leaving group.
As shown in Scheme 1, the addition of an anionic
nucleophile (Nu-) to a radical cation (ArX+•) produces an
uncharged nucleophilic adduct (I), which is none other
than the homolytic adduct obtainable by addition of the
radical Nu• to the neutral substrate ArX. Thus, a first
possibility for the leaving group is to dissociate from
Because of the oxidizing conditions operating in this
type of reaction, the second most likely possibility for the
leaving group is to detach as an electrophile (X+) from
the Wheland intermediate II formed by oxidation of
adduct I (path ii). In comparison, H-substitution reac-
tions such as anodic cyanation and acetoxylation of
anisole and other aromatic compounds,6 pyridination of
mesitylene by photoactivation of its electron donor-
acceptor complex with N-nitropyridinium,7 chlorination
of methyl-substituted methoxybenzenes with iodine
monochloride,8 and azidation and acetoxylation of 1,4-
dimethoxybenzene induced by hypervalent iodine9 have
been interpreted as nucleophilic displacements of H+. An
analogous mechanistic formulation in terms of nucleo-
philic deprotonation was advanced by Eberson10 to ac-
count for the formation of H-substitution products from
the direct reaction of certain nucleophiles (AcO-, Cl-,
* To whom correspondence should be addressed. Fax: +39-080-
5442924. E-mail: ciminale@chimica.uniba.it.
† Taken in part from the respective graduation theses.
(1) Eberson, L. Electron-Transfer Reactions in Organic Chemistry;
Springer: Berlin, 1987.
(2) Schmittel, M.; Burghart, A. Angew. Chem., Int. Ed. Engl. 1997,
36, 2550.
(3) Kyriacou, D. Modern Electroorganic Chemistry; Springer: Berlin,
1994; Chapter 2.
(4) (a) Pross, A. J . Am. Chem. Soc. 1986, 108, 3537. (b) Shaik, S. S.;
Pross, A. J . Am. Chem. Soc. 1989, 111, 4306.
(5) Parker, V. D.; Tilset, M. J . Am. Chem. Soc. 1987, 109, 2521.
(6) (a) Andreades, S.; Zahnov, E. W. J . Am. Chem. Soc. 1969, 91,
4181. (b) Eberson, L.; Nyberg, K. Acc. Chem. Res. 1973, 6, 106.
(7) Kim, E. K.; Bochman, T. M.; Kochi, J . K. J . Am. Chem. Soc. 1993,
115, 3091.
(8) Hubig, S. M.; J ung, W.; Kochi, J . K. J . Org. Chem. 1994, 59,
6233.
(9) Kita, Y.; Tohma, H.; Hatanaka, K.; Takada, T.; Fujita, S.; Mitoh,
S.; Sakurai, H.; Oka, S. J . Am. Chem. Soc. 1994, 116, 3684.
(10) (a) Eberson, L.; Larsson, B. Acta Chem. Scand. B 1986, 40, 210.
(b) Eberson, L.; Larsson, B. Acta Chem. Scand. B 1987, 41, 367.
10.1021/jo982276j CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/18/1999

2460 J . Org. Chem., Vol. 64, No. 7, 1999
Ciminale et al.
CN-) with the tris(4-bromophenyl)aminium radical cat-
ion. Moreover, a relevant example of ipso Nu-/X+ sub-
stitution is the anodic cyanation of 1,4-dimethoxybenzene
to give 4-methoxybenzonitrile; it was proposed in this
case that the methoxyl group is lost from a cationic
adduct (II in Scheme 1) as formaldehyde, the stable
deprotonated form of the electrophile MeO+.6a
Sch em e 2
The third possibility (path iii) is that the leaving group
is expelled as a nucleophile (X-) from the primary adduct
I, thus generating the radical cation of the substitution
product ArNu+•. Actually, this fragmentation is better
known as part of a chain reaction mechanism, (i.e., the
SON2 mechanism).11,12 In this mechanism, it is followed
by a single electron transfer from a new substrate
molecule to the radical cation ArNu+•, thus completing
the propagation sequence by giving the substitution
product and regenerating the active species ArX+•. How-
ever, as pointed out by Eberson et al.,12 simple thermo-
chemical considerations would seem to limit the scope
of this mechanism to those cases involving barely oxidiz-
able nucleophiles and leaving groups; otherwise suspi-
cions should arise that both Nu- and X- might partici-
pate in the substitution process in their oxidized forms,
namely, Nu• and X• or X+. Therefore, under the specific
oxidative conditions commonly employed, only nucleo-
philes such as AcO-, CF3CO2-, OH-, and CN- and
leaving groups such as F- and Cl- have so far been
judged to be firmly compatible with the SON2 mechanism.
Moreover, because of a common situation in oxidative
substitution chemistry, namely, that the product ArNu
is often more easily oxidized than the substrate ArX, clear
evidence for the chain reaction attributes of this mech-
anism could seldom be achieved and only at a low degree
of conversion12 when ArX/ArNu ratios are high enough
to sustain somehow the catalytic efficiency despite the
Ta ble 1. Rea ction betw een N,N-Dim eth yl-p-a n isid in e
(1a ) a n d N-Meth yla n ilin e (2b-H) in th e P r esen ce of
Th a lliu m Tr ia ceta te Sesqu ih yd r a te (TTA)
reactn time conva 3b yielda
run solvent TTA/1a
(min)
(%)
(%)
3a /3b
1
MeNO2
1
20
60
20
5
20
60
20
5
40
48
47
35
0.26
0.23
1
2
3
MeCN
MeNO2
1
2
44
45
47
68
44
38
0.1
4
5
MeCN
MeNO2
2
3
0.15
<0.05
40
64
75
39
20
a
As determined by GC; yelds are based on the amount of
substrate consumed.
of TTA in oxidizing N-methylated anisidines to the
corresponding radical cations without any reference to
its hydrate form,13,14 we are now forced to specify that
we had always been using sesquihydrate TTA. Fortuitous
circumstances in the present research showed that
anhydrous TTA was quite ineffective as an oxidant but
performed better in reactions when small quantities of
water were added to the solvent.15 The oxidative activa-
tion was relatively effective in both nitromethane and
acetonitrile, whereas it was substantially absent in
dichloromethane. The efficiency of the MeO-substitution
leading to the 1,4-benzenediamine derivative 3b was
contrasted with the concurrent demethylation of 1, as
inferred from the detection in the reaction mixtures not
so much of small amounts of N-methyl-p-anisidine (2a -
H) but primarily of 3a , which is the product of subse-
quent MeO-substitution by this secondary anisidine
(Scheme 2).
Inspection of 3a /3b ratios reported in Table 1 shows
that more competing demethylation took place in aceto-
nitrile and at lower oxidant-to-substrate molar ratios
(TTA/1). Use of a 3-fold excess of TTA in MeNO2 (entry
5) inhibited this undesired side reaction to the adequately
negligible extent of less than 5%. The yields of 3b and
substrate conversions at various reaction times are also
included in Table 1 and reveal that the substitution
process takes place with maximum efficiency within a
few minutes when conversion is less than 50%. In the
case of highest oxidant excess (TTA/1 ) 3), the reaction
could be forced to reach a 64% conversion after a longer
time (20 min), but this occurred at the expense of yield,
which fell from 75% to 39%.
endoergonic electron transfer from ArX to ArNu+•
.
In 1988, a product study13 aimed at identifying one of
the two radical cations observed during single electron
oxidation of N,N-dimethyl-p-anisidine (1a ) provided us
with a new example of aromatic substitution promoted
by oxidation, the formation of a tetrasubstituted 1,4-
benzenediamine 3 by reaction of 1a with a secondary
aniline 2-H in the presence of thallium triacetate (TTA)
as the oxidant. Relevant to the chemistry described
above, we have now investigated this reaction (eq 1) in
consideration of the fact that it constitutes a particularly
interesting novel situation for evaluating the mechanistic
variety summarized in Scheme 1.
(1)
Resu lts a n d Discu ssion
The reaction of 1a with different secondary amines was
thus carried out on a larger scale in MeNO2 in the
Seeking optimum reaction conditions, we performed
control experiments for the reaction of N,N-dimethyl-p-
anisidine (1a ) with N-methylaniline (2b-H) and TTA.
Although we have already reported on the effectiveness
(13) Ciminale, F.; Curci, R.; Portacci, M.; Troisi, L. Tetrahedron Lett.
1988, 29, 2463.
(14) Ciminale, F. Tetrahedron Lett. 1994, 35, 3375.
(15) In view of the two mecanisms for the oxidation of a tertiary
amine with TTA reported in the next reference, it is conceivable that
water might have a certain influence on the oxidative power of TTA
because it would enter as a component of the metal coordination shell.
(11) Alder, R. W. Chem. Commun. 1980, 1184.
(12) Eberson, L.; J o¨nsson, L.; Wistrand, L.-G. Tetrahedron 1982, 38,
1087 and references therein.

Oxidative Activation in Aromatic Substitutions
J . Org. Chem., Vol. 64, No. 7, 1999 2461
Ta ble 2. Su bstitu tion P r od u cts fr om Rea ction s of
N,N-Dim eth yla n ilin es (p-XC6H4NMe2) 1a -c w ith
Secon d a r y Am in es 2(a -f)-H, P r om oted by Th a lliu m
Tr ia ceta te Sesqu ih yd r a te
Sch em e 3
p-XC6H4NMe2
1a (X ) MeO)
amine
products (yield, %)a
convb (%)
2a -H
2b-H
2c-H
2d -H
2e-H
2f-H
2a -H
2b-H
2c-H
2d -H
2a -H
2b-H
2c-H
2d -H
3a (40)
3b (44)
3c (53)
3d (35), 4 (8)
3a (na)
3a (na)
3a (ca. 20)c
3bd
45
50
53
55
1b (X ) H)
1c (X ) F)
3cd
3d (39)e
3bd
41e,f
31e,f
3cd
3c (ca. 10)c
3d (63)e
reaction pathway for this amino-demethoxylation is
delineated in Scheme 3 as a proper adaptation of the
general Scheme 1 on the basis of the one proton difference
between related nucleophilic species. In fact, III should
easily lose a proton to give the homolytic adduct IV that
might get involved in the product-forming step either
directly (path a or b) or by undergoing oxidative trans-
formation into the Wheland adduct V (path c). However,
it is worth considering that path a is not experimentally
discernible from paths b and c on the basis of the
competent oxidation level of the reaction product, which
because of its very low oxidation potential18 is obtained
as radical cation 3+•, even though the fragmentation of
the proper precursor would imply its primary formation
at the lower oxidation level, that is, as 3.
In line with the working hypothesis of nucleophilic 2-H/
1a +• interaction, the competition between demethylation
and substitution referred to above seems to be governed
by a balance between the basicity of the medium and the
nucleophilicity of the attacking amine. In fact, the larger
proportion of demethylation observed in acetonitrile at
TTA/1 ) 1 (entry 2 of Table 1) can be imputed to the
relatively higher basicity of this solvent20 and to the fact
that a lower amount of oxidant would cause the aminium
radical 1+• to be generated in the presence of a greater
quantity of the basic parent compound. On the other
hand, the substitution process is compatible with those
secondary aromatic amines, 2a -c-H, that evidently may
be considered too weak as bases and good enough as
nucleophiles. Moreover, with aliphatic amine 2f-H and
nitro-substituted secondary aniline 2e-H, deprotonation
is the dominant process owing to a considerable basicity
of the former and a scarce nucleophilicity of the latter
amine. In this case, the observed reactivity is akin to that
displayed in the absence of any added secondary amine.
However, nucleophilic addition of the secondary amine
2d -H to 1a +• is no longer applicable to the formation of
the direct substitution product 3d in the case of diphe-
nylamine. In fact, judging from pKa values21 of the
corresponding ammonium ions, 2d -H (pKa ) 0.79) should
be considered at least as feeble a nucleophile as 2e-H
(pKa > 1)22 and likewise incapable of adding to 1+•. The
a
Unless otherwise noted, yields refer to isolated products and
are based on the amount of substrate reacted; na ) not analyzed.
b Unless otherwise noted, conversions refer to 1a -c. c As estimated
by GC-MS. Negligible amounts. e As determined by 1H NMR
d
(see Experimental Section). f Conversion referred to 2d -H.
presence of a 3-fold excess of TTA over ca. 10 min. As
shown in Table 2, with N-methylanilines 2(a -c)-H, the
direct amino-demethoxylation products 3a -c could be
isolated for characterization in 40-53% yields following
a reductive workup of the reaction mixtures (see Experi-
mental Section) in which they were present in the
oxidized form of radical cation. The yield of recovered 1a
was 47-55%. Similarly, reaction with diphenylamine
(2d -H) afforded the MeO-substitution product 3d in 35%
yield at 55% conversion. This, however, could not be fully
isolated but was eventually obtained in mixture with a
minor (8%) nuclear H-substitution product (most likely
4) and unreacted 2d -H. Thus, product identification and
yields were based only on GC-MS and/or 1H NMR
analyses of this mixture, as specified in the Experimental
Section. It is important to note that in the other cases
only comparatively very small amounts of H-substitution
products were recognized on GC-MS spectra of the
reaction mixtures.
In reactions with N-methyl-p-nitroaniline (2e-H) or
diethylamine (2f-H), only relevant undetermined quanti-
ties of the same 1,4-benzenediamine 3a were detected,
indicating that 2e,f-H are incompatible with substitution.
The reasons for such an incompatibility with direct MeO-
substitution should be different in the two cases and will
be discussed later on.
With the support of ESR detection, we have already
verified that radical cation 1a +• is the primary reactive
intermediate involved in demethylation of 1 with TTA.13
According to the commonly accepted mechanism for
oxidative dealkylation of tertiary amines,16 the decom-
position of 1a +• would be triggered by its rapid deproto-
nation.17 Concerning the formation of substitution prod-
ucts of type 3, we have also advanced that this reaction
might proceed through a nucleophilic attack of a second-
ary aniline on the same aminium intermediate 1a +•
.
(18) As a typical oxidation potential of compounds 3, the E1/2 value
of N,N,N′,N′-tetramethyl-p-phenylenediamine, reported in ref 19, may
be considered: -0.10 V (CH3CN, vs Ag/0.01 N Ag+); a comparatively
higher value is reported for 1a : 0.33 V (CH3CN, vs Ag/0.01 N Ag+).
(19) Weinberg, N. L.; Weinberg, H. R. Chem. Rev. 1968, 68, 449.
(20) Bagno, A.; Scorrano, G. J . Am. Chem. Soc. 1988, 110, 4577.
(21) Perrin, D. D. Dissociation Constants of Organic Bases; Plenum
Press: New York, 1965.
Starting from the resulting adduct III, a conceivable
(16) Butler, R. N. Chem. Rev. 1984, 84, 249.
(17) Large deprotonation rates (k ≈ 104-105 M-1 s-1) have been
determined for radical cations of N,N-dimethyl-p-anisidine and of
related di-p-anisylmethylamine: (a) Zhang, X.; Yeh, S.-R.; Hong, S.;
Freccero, M.; Albini, A.; Falvey, D. E.; Mariano, P. S. J . Am. Chem.
Soc. 1994, 116, 4211. (b) Dinnocenzo, J . P.; Banach, T. E. J . Am. Chem.
Soc. 1989, 111, 8646.
(22) The pKa of N-methyl-p-nitroaniline was estimated by referring
to the nonmethylated analogue: pKa ) 1.

2462 J . Org. Chem., Vol. 64, No. 7, 1999
Ciminale et al.
fact that 2d -H, in contrast to 2e-H, is nevertheless
involved in a substitution reaction might be due to its
relatively lower oxidation potential [Eox(2d ) - Eox(2e) =
-0.3 V],23 suggesting that diphenylamine can actually
participate in the associative step in an oxidized form.
Moreover, a comparison of oxidation potentials based on
E1/2 values indicates that 2d -H (E1/2 ) 0.39 V vs Ag/0.01
N Ag+)24 has an oxidation potential close to that of
oxidizable 1a (E1/2 ) 0.33 V vs Ag/0.01 N Ag+)19 and,
consequently, its oxidation by TTA is very likely.
As expected, substitution was found to occur with
substrates 1b,c, but only in reactions with diphenylamine
(Table 2). H-substitution at the para position of 1b and
F-substitution on 1c afforded the same product, 3d ,
obtained by MeO-substitution from 1a , in comparable
amounts. In reactions with N-methylaniline or N-methyl-
p-fluoroaniline, however, only negligible amounts of the
corresponding substitution product 3b,c were detected
by GC-MS analysis. Other significant products that
could testify for the occurrence of starting compounds
oxidation were not observed except for small amounts of
substrate monodemethylation products 2b,c-H28 and
trace amounts of a p,p′-homocoupling product, tetram-
ethylbenzidine, in the reactions of 1b. More appreciable
amounts of substitution product 3a were detected in
reactions with methylated p-anisidine 2a -H, mainly in
the case of substrate 1b. Significantly, as revealed by
GC-MS data, these reactions were also characterized by
a markedly uneven consumption of reactants, namely,
an almost complete disappearance of 2a -H and relatively
low conversions of substrates 1b,c. This suggests that
2a -H undergoes oxidation more easily than 1b,c, and the
aminyl radical 2a • thus formed, similarly to but to a
lesser extent than 2d • in the case of diphenylamine,
might be responsible for H- or F-substitution on 1b,c.
For the most part, 2a •, according also to our previous ESR
investigation,14 seems to become involved in homocou-
pling reactions, as pointed out by the formation of small
amounts of N,N′-bis(4-methoxyphenyl)dimethylhydrazine
(5), N-(4-methoxyphenyl)-N,N′-dimethyl-1,4-benzenedi-
amine (6), and other unidentified products of probable
further oxidation.
Reaction pathways based on other plausible combina-
tions of active species in addition to 2-H/1a +• are also
traced in Scheme 3. Thus, for oxidizable secondary
amines such as 2d -H, a suitable sequence might involve
an amino radical 2d •, most likely engendered by rapid
deprotonation of the primary product of one-electron
oxidation, 2d -H+•. The homolytic addition of 2d • to 1a
would directly produce adduct IV, which in the first
formulation follows the initial formation of adduct III,
deriving from the 2-H/1a +• interaction. Another conceiv-
able alternative for 2d • should be the coupling with 1a +•
to give directly adduct V as a likely precursor of the
substitution product. It is altogether reasonable to believe
that a radical rather than a nucleophilic involvement of
the secondary amine might be the cause of the observed
different regioselectivity of 2d -H giving relatively more
conspicuous H-substitution than 2(a -c)-H.
Taking advantage of the fact that, according to Scheme
3, the combination 1/2• is the only viable reaction
pathway for substrates that, contrary to 1a , are not
amenable to oxidation, we sought to substantiate the
supposed radical reactivity of 2d -H as distinct from the
nucleophilic reactivity of 2(a -c)-H with recourse to
analogous reactions with N,N-dimethylaniline (1b) and
N,N-dimethyl-p-fluoroaniline (1c). These analogues of 1a ,
bearing potential leaving groups (X ) H, F) compatible
with oxidative substitution, were chosen as appropriate
substrates because they can be reasonably considered as
relatively more difficult to oxidize. In fact, compared to
25
1a (Ep/2 ) 0.49 V vs SCE),26 1b is reported to have a
higher redox potential (Ep/2 ) 0.71 V vs SCE), and its
radical cation 1b+• is reported to be highly unstable.26
For 1c, we could roughly estimate an equally high redox
potential (Ep/2 ≈ 0.74 V) by applying the electronic effect
correlation of para substituents (σ+ values)27 with related
Ep/2 values reported in the same electrochemical study.26
In addition, direct evidence that 1c is rather resistant
to oxidation under our reaction conditions (i.e., TTA as
oxidant) became available after several ESR attempts to
detect its radical cation. Contrary to what had been
A closing remark on the mechanistic features of our
substitution reaction has to be made concerning the
product-forming step. Fragmentation modes (a and b
from adduct IV and c from adduct V) are reported in
Scheme 3, each bearing reference to the leaving group-
(s) that may be considered as likely candidate(s) for that
specific cleavage of the C-X bond, consistent with the
oxidation level of the resulting X. For the fragmentation
of IV, we suggest that F and H are the para substituents
that should be ejected as X- (path a) and X• (path b),
respectively, whereas MeO might be compatible with
both of the possibilities. This proposal agrees with
Eberson’s criterion12 as applied when considering the
redox potentials for both of the fragments of C-X
cleavage, namely, for 3+•/329 and X•/X- 30 couples. For X
leaving as X+ from V (path c), the proposed compatibility
with MeO and H and the exclusion of F are simply in
observed for 1a +• 13 1c+• turned out to be highly unstable
,
inasmuch as to obtain it in detectable amounts it was
necessary to effect the oxidation with TTA in the presence
of methansulfonic acid.
(23) This difference in oxidation potentials was roughly estimated
on the basis of E1/2 or Ep/2 values reported in ref 19; in default of data
referring to N-methyl-p-nitroaniline, the Ep/2 of the N,N-dimethyl
analogue was considered.
(24) This E1/2 value vs Ag/0.01 N Ag+ was derived from that given
vs SCE in ref 19 taking into account the difference of 0.44 V between
the oxidation potentials of the two reference electrodes.
(28) This otherwise banal claim about detection of demethylation
products must, obviously, be understood as referring to reactions in
which substrate 1 and secondary aniline 2-H are differently substituted
at the para position (X * Y).
(29) For a homogeneous comparison with values of ref 27, as typical
redox potential of the 3+•/3 couple, the value of E° ) 0.25 V vs NHE
for N,N,N′,N′-tetramethyl-p-phenylenediamine, reported in ref 1, may
be considered.
(25) Used, instead of the formerly quoted E1/2 value, to allow a more
homogeneous comparison with the available oxidation potential of 1b
given as Ep/2
.
(26) Seo, E. T.; Nelson, R. F.; Fritsch, J . M.; Marcoux, L. S.; Leedy,
D. W.; Adams, R. N. J . Am. Chem. Soc. 1966, 88, 3498.
(27) Brown, H. C.; Okamoto, Y. J . Am. Chem. Soc. 1958, 80, 4979.
(30) Eberson, L. Acta Chem. Scand. B 1984, 38, 439.

Oxidative Activation in Aromatic Substitutions
J . Org. Chem., Vol. 64, No. 7, 1999 2463
Op tim iza tion Stu d ies. Gas chromatographic analyses on
the proceeding reaction to determine percent substrate conver-
sions and product yields reported in Table 1 were performed
according to the following general procedure. An equimolar
solution (typically 6.6 × 10-2 M) of anilines 1a and 2b-H in 5
mL of the selected solvent was added dropwise to a magneti-
cally stirred suspension of an appropriate quantity of TTA in
15 mL of the same solvent. The reaction solution immediately
turned blue and became paramagnetic (ESR tested) as a result
of the formation of the radical cation of the substitution
product(s). Aliquots (4 mL) of this solution were then with-
drawn at timed intervals and subjected to reducing treatment
with a 10% Na2S2O3 solution (ca. 30 mL). After the blue color
faded, each aliquot was extracted with diethyl ether. The
extracts were dried (Na2SO4) and filtered, and the solvent was
removed in vacuo. The residue was then taken up in 0.5 mL
of a 0.1 M solution of p-anisidine, which served as an internal
standard in the subsequent quantitative GC analysis.
Isola tion a n d Ch a r a cter iza tion of Su bstitu tion P r od -
u cts. The general procedure adopted to isolate substitution
products 3a -d is as follows. A solution of 1a (0.25 g, 1.6 mmol)
and an equimolar quantity of the appropriate secondary
aniline from among 2(a -d )-H in MeNO2 (15 mL) was added
dropwise to a stirred suspension of TTA (1.94 g, 4.7 mmol) in
60 mL of the same solvent. After the addition, the resultant
blue mixture was allowed to react for 10 min, treated with a
solution of Na2S2O3 (10%, 250 mL), and extracted with diethyl
ether. The organic phase was dried (Na2SO4) and concentrated
in vacuo. Flash column chromatography (silica gel, petroleum
ether/Et2O 5:2) of the residue gave substitution products 3a -c
separated from unreacted 1a (minor materials were left
unrecovered in column). Similar chromatographic treatment
(eluant, petroleum ether/Et2O 4:1), when applied to the crude
of the reaction with 2d -H, afforded 3d mixed with unreacted
2d -H and another substitution product (see below). All NMR
spectra of substitution products were recorded in (CD3)2CO.
The use of this solvent rather than CDCl3 was prescribed by
the occurrence in the latter solvent of spontaneous oxidation
of 1,4-benzenedianmines 3a -d to the corresponding radical
cations (solutions took on a shade of blue), causing disturbing
paramagnetic broadening of NMR resonances. Such a broad-
ening effect was also observed in a limited fashion in uncolored
(CD3)2CO solutions but could be completely eliminated on
treating the samples with powdered zinc prior to recording
the spectrum.
N-(4-Meth oxyp h en yl)-N,N′,N′-tr im eth yl-1,4-ben zen ed i-
a m in e (3a ): 74 mg (40%) [133 mg of recovered 1a , conv. 45%];
mp 104-105 °C (recrystallized from hexane/dichloromethane);
IR (KBr) 3037, 2988-2805, 1612, 1505, 1333 (C-N), 1240 (Ar-
O), 1033 (O-CH3), 830 and 820 (substitution bands) cm-1; 1H
NMR (200 MHz, (CD3)2CO) δ 2.88 (s, 6H), 3.16 (s, 3H), 3.73
(s, 3H), 6.72-6.96 (m, 8H); 13C NMR (125.76 MHz, (CD3)2CO)
δ 41.2 (two partially overlapped signals are distinguishable
on the enlarged spectrum, 41.18 and 41.22), 55.7, 114.7, 115.1,
119.8, 124.2, 141.2, 145.2, 147.8, 154.2; MS m/z (rel abundance)
256 (M+, 100), 241 (68), 226 (41), 120 (12).
N ,N ,N ′-T r im e t h y l-N ′-p h e n y l-1,4-b e n ze n e d ia m in e
(3b): 80 mg (44%) [121 mg of recovered 1a , conv. 50%]; mp
44-45 °C; IR (KBr) 3034, 2878-2803, 1600, 1519, 1338 (C-
N), 822 and 750 (substitution bands); 1H NMR (200 MHz,
(CD3)2CO) δ 2.91 (s, 6H), 3.19(s, 3H), 6.65-6.78 (m, 5H), 6.99-
7.15 (m, 4H); 13C NMR (125.76 MHz, (CD3)2CO) δ 40.6, 40.9,
114.4, 115.2, 117.9, 127.5, 129.4, 139.4, 149.2, 151.2; MS m/z
(rel abundance) 226 (M+, 100), 211 (63), 196 (17), 167 (37), 77
(33).
accord with the well-known reactivity in electrophilic
substitutions.
Con clu sion s
Three different combinations of active species are
considered to account for TTA-promoted oxidative sub-
stitutions occurring at the para position of N,N-dimethy-
lanilines 1a -c by several secondary anilines 2(a -d )-H
(eq 1): a radical cation-nucleophile combination (1+•/2-
H), a substrate-aminyl radical combination (1/2•), and
a radical cation-aminyl radical combination (1+•/2•).
Which one is in fact involved depends on whether the
oxidative activation is effected on the substrate, on the
secondary aniline, or on both of the reactants, respec-
tively. The results reported herein suggest that, in the
case of easily oxidizable substrate 1a , the MeO-substitu-
tion may proceed either via nucleophilic attack at the
substrate radical cation (1+•/2-H interaction) with barely
oxidizable secondary anilines such as 2(b,c)-H or via
attack by the aminyl radical 2d • on the original (1/2•
interaction) and/or oxidized (1+•/2• interaction) substrate
with easily oxidizable and non nucleophilic 2d -H. Obvi-
ously, all three reaction modes would be compatible with
oxidizable and nucleophilic 2a -H. In line with such a
reactivity pattern, it was found that dimethylanilines
1b,c, despite being relatively difficult to convert into
reactive radical cations, nevertheless undergo substitu-
tion (H- and F-substitution, respectively) but only with
oxidizable secondary anilines, 2a ,d -H; a 1/2• interaction
is evidently the unique viable pathway for these sub-
strates.
The 1+•/2-H interaction is currently being investigated
in more detail by a more direct approach based on isolable
1+• analogues.
Exp er im en ta l Section
Ma ter ia ls. Commercially available MeNO2 (Aldrich HPLC)
and MeCN (Fluka purum) were used as solvents without
further purification. Thallium triacetate sesquihydrate (TTA)
(Aldrich) was also used as received. N,N-dimethyl-p-anisidine
(1a ) was prepared by methylation of commercial (Aldrich)
p-anisidine by Me2SO4 following a described procedure.31 N,N-
dimethyl-p-fluoroaniline (1c) and N-methyl-p-fluoroaniline (2c-
H) were obtained in a single preparation by methylation of
neat p-fluoroaniline (Lancaster Synthesis) (7.2 g, 65 mmol)
with CH3I (9.2 g, 65 mmol) and flash column chromatography
(hexanes/Et2O 5:2) of the crude of reaction. 1c (1 g): mp 33-
35 °C [lit.32 33 °C]; IR (KBr) 3051, 2925-2805, 1611, 1515,
1347 (C-N), 1228 (C-F), 816 (substitution band) cm-1 1H
;
NMR (200 MHz, CDCl3) δ 2.90 (s, 6H), 6.64-6.72 (m, 2H),
6.90-7.02 (m, 2H); MS m/z (rel abundance) 139 (M+, 73), 138
(100), 123 (20), 122 (21), 95 (20), 75 (14), 42 (12). 2c-H (1.5 g):
oil [lit.33 bp 136 °C (120 mmHg)]; IR (neat) 3420 (N-H), 3060,
3000-2800, 1612, 1472, 1512, 1320 (C-N), 1220 (C-F), 821
1
(substitution band) cm-1; H NMR (200 MHz, CDCl3) δ 2.80
(s, 3H), 3.33 (bs, 1H), 6.51-6.62 (m, 2H), 6.86-6.98 (m, 2H);
MS m/z (rel abundance) 125 (M+, 78), 124 (100), 97 (19), 96
(16), 95 (15), 83 (19), 75 (16), 57 (12). N,N-dimethylaniline (1b)
(Aldrich), N-methylaniline (2b-H) (J anssen), N-methyl-p-ni-
troaniline (2e-H) (Aldrich), and diphenylamine (2d -H) (Ald-
rich) were used without further purification. N-methyl-p-
anisidine (2a -H) (Aldrich) was recrystallized (ligroin, 80-120
°C) before use.
N-(4-F lu or op h en yl)-N,N′,N′-t r im et h yl-1,4-b en zen ed i-
a m in e (3c): 114 mg (53%) [117 mg of recovered 1a , conv.
53%]; mp 92-93 °C; IR (KBr) 3039, 2982-2804, 1609, 1504,
1340 (C-N), 1212 (C-F), 822 and 807 (substitution bands)
cm-1 1H NMR (200 MHz, (CD3)2CO) δ 2.9O (s, 6H), 3.16(s,
;
3H), 6.65-6.77 (m, 4H), 6.86-7.01 (m, 4H); 13C NMR (125.76
2
MHz, (CD3)2CO) δ 40.9, 41.1, 114.5, 115.7 (d, J CF ) 22.1 Ηz),
3
117.0 (d, J CF ) 7.4 Hz), 126.6, 139.8, 148.0, 149.0, 156.7 (d,
(31) Bordwell, F. G.; Boutan, P. J . J . Am. Chem. Soc. 1956, 78, 87.
(32) Sellers, C.; Suschitzky, H. J . Chem. Soc. 1965, 6186.
(33) Fox, B. L.; Doll, R. J . J . Org. Chem. 1973, 38, 1136.
1J CF ) 233.9 Hz); MS m/z (rel abundance) 244 (M+, 100), 229
(73), 185 (49), 122 (20).

2464 J . Org. Chem., Vol. 64, No. 7, 1999
Ciminale et al.
N ,N -Dim e t h yl-N ′,N ′-d ip h e n yl-1,4-b e n ze n e d ia m in e
(3d ): MS m/z (rel abundance) 288 (M+, 100), 273 (35), 167
(43), 144 (25), 77 (46), 51 (27); 1H NMR (200 MHz, (CD3)2CO)
δ 2.87 (s, 6H). These spectroscopic characteristics were reliably
extracted from proton NMR and GC-MS spectra of a chro-
matographic fraction (250 mg) that, in addition to 3d , was
composed of unreacted 2d -H and a product that was identified
as the H-substitution product 4 on the basis of the following
spectroscopic data: MS m/z (rel abundance) 318 (M+, 100), 303
(13), 286 (24), 77 (82), 51 (42); 1H NMR (200 MHz, (CD3)2CO)
δ 2.47 (s, 6H), 3.65 (s, 3H). 2d -H conversion (55%) and related
product yields (3d , 35%; 4, 8%) were easily estimated on
quantitative 1H NMR analysis of the chromatographic mixture.
The OMe and NMe2 protons of 4 were clearly distinguishable
in the spectrum of the original (CD3)2CO solution by the correct
intensity ratio (1:2) of their signals. The broad singlet (δ, 7.48)
of the exchangeable NH proton of 2d -H was advantageously
eliminated from the aromatic chemical shift region on treat-
ment with D2O. Final integrals measurement for quantitative
analysis was performed after debroadening powdered Zn
treatment of this solution.
Ack n ow led gm en t. This work was supported in part
by Ministero della Ricerca Scientifica e Tecnologica,
Rome - Progetto Cofinanziato “Radicali Liberi e Radi-
cali Ioni nei Processi Chimici e Biologici”.
J O982276J