Mechanistic investigation of oxidative Mannich reaction with tert-butyl hydroperoxide. the role of transition metal salt

Article abstract of DOI:10.1021/ja3113559

A general mechanism is proposed for transition metal-catalyzed oxidative Mannich reactions of N,N-dialkylanilines with tert-butyl hydroperoxide (TBHP) as the oxidant. The mechanism consists of a rate-determining single electron transfer (SET) that is uniform from 4-methoxy- to 4-cyano-N,N-dimethylanilines. The tert-butylperoxy radical is the major oxidant in the rate-determining SET step that is followed by competing backward SET and irreversible heterolytic cleavage of the carbon-hydrogen bond at the α-position to nitrogen. A second SET completes the conversion of N,N-dimethylaniline to an iminium ion that is subsequently trapped by the nucleophilic solvent or the oxidant prior to formation of the Mannich adduct. The general role of Rh₂(cap) ₄, RuCl₂(PPh₃)₃, CuBr, FeCl ₃, and Co(OAc)₂ in N,N-dialkylaniline oxidations by T-HYDRO is to initiate the conversion of TBHP to tert-butylperoxy radicals. A second pathway, involving O₂ as the oxidant, exists for copper, iron, and cobalt salts. Results from linear free-energy relationship (LFER) analyses, kinetic and product isotope effects (KIE and PIE), and radical trap experiments of N,N-dimethylaniline oxidation by T-HYDRO in the presence of transition metal catalysts are discussed. Kinetic studies of the oxidative Mannich reaction in methanol and toluene are also reported.

Full text of DOI:10.1021/ja3113559

Article
pubs.acs.org/JACS
Mechanistic Investigation of Oxidative Mannich Reaction with tert-
Butyl Hydroperoxide. The Role of Transition Metal Salt
Maxim O. Ratnikov and Michael P. Doyle*
Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742, United States
S
* Supporting Information
ABSTRACT: A general mechanism is proposed for transition metal-
catalyzed oxidative Mannich reactions of N,N-dialkylanilines with tert-butyl
hydroperoxide (TBHP) as the oxidant. The mechanism consists of a rate-
determining single electron transfer (SET) that is uniform from 4-methoxy-
to 4-cyano-N,N-dimethylanilines. The tert-butylperoxy radical is the major
oxidant in the rate-determining SET step that is followed by competing
backward SET and irreversible heterolytic cleavage of the carbon−hydrogen
bond at the α-position to nitrogen. A second SET completes the conversion
of N,N-dimethylaniline to an iminium ion that is subsequently trapped by
the nucleophilic solvent or the oxidant prior to formation of the Mannich adduct. The general role of Rh2(cap)4, RuCl2(PPh3)3,
CuBr, FeCl3, and Co(OAc)2 in N,N-dialkylaniline oxidations by T-HYDRO is to initiate the conversion of TBHP to tert-
butylperoxy radicals. A second pathway, involving O2 as the oxidant, exists for copper, iron, and cobalt salts. Results from linear
free-energy relationship (LFER) analyses, kinetic and product isotope effects (KIE and PIE), and radical trap experiments of
N,N-dimethylaniline oxidation by T-HYDRO in the presence of transition metal catalysts are discussed. Kinetic studies of the
oxidative Mannich reaction in methanol and toluene are also reported.
In the past decade several catalytic systems were developed
for the conversion of N,N-dialkylanilines to their corresponding
iminium ions. The scope of catalysts for this transformation
include Ru,4a,5 Rh,3 Cu,2b,c,6 Fe,7 Mo,8 and V9 complexes.
Recent advances in the oxidative Mannich reaction of N,N-
dialkylanilines enabled the use of dioxygen4a,6a,b or di-tert-
butylperoxide (DTBP) as oxidants.10 However, mild con-
ditions3 and often higher yields2b,6a,f are achieved with tert-butyl
hydroperoxide (TBHP) that prevails as the oxidant of choice.
The mild conditions of transition metal-catalyzed oxidation
of N,N-dialkylanilines to iminium ions with TBHP enable
utilization of selected nucleophiles such as siloxyfurans,3 enol
silanes,6b,11 allyl silanes,7b indoles,5b,6c,7c,12 cyanides,4a,5c,7a,8,9b,13
alkynes,2b,c ketones,5a,6c,e,9a,14 nitroalkanes,5d,8,15 1,3-dicarbonyl
compounds,15b,16 electron-rich arenes,17 heterarenes,6a,7c,17
boronic acids,18 and heteroatom nucleophiles19 in the oxidative
Mannich reaction. The majority of these nucleophiles were
employed in the copper(I)-catalyzed oxidative reactions of N-
aryl-tetrahydroisoquinolines.2b,c,6a The use of Rh3 and Ru5b,19a
complexes, on the other hand, provides a broader substrate
scope that includes N,N-dilakylanilines, but this substrate class
has a limited scope of nucleophiles that can be employed. The
question remains open whether the scope of N,N-dialkylani-
lines and reactive nucleophiles in the oxidative Mannich
reactions is catalyst-specific or could be generalized among
different catalysts and conditions. An understanding of the role
of the transition metal in the N,N-dialkylaniline oxidation step
INTRODUCTION
■
The Mannich reaction is a powerful transformation for the
syntheses of nitrogen-containing natural products and active
pharmaceutical ingredients.1 Iminium ions, key intermediates in
the Mannich reaction, react with a large variety of
nucleophiles.2 Despite the synthetic value of iminium ions,
only simple N,N-dimethylmethylidene-ammonium halides
(Eschenmoser’s salts) are commercially available. Hence,
iminium ions are commonly prepared in situ by methods that
limit the scope of the nucleophiles with which they are
compatible. Strategies allowing access to Mannich reaction
intermediates include condensation of secondary amines with
aldehydes,1a α-fragmentation of iminium ion precursors,1b and
the oxidation of tertiary amines (Scheme 1).2a−c,3,4 The latter
method directly and under mild conditions functionalizes
otherwise inert C−H bonds that are adjacent to nitrogen.
Scheme 1. Strategies for Preparation the Key Intermediate of
the Mannich Reaction
Received: November 19, 2012
Published: January 8, 2013
© 2013 American Chemical Society
1549
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
and in the subsequent Mannich reaction would provide an
answer to this question. The detailed mechanism of each step
could also lead to a reduction in the excess of N,N-
dialkylanilines required to achieve high yields3,5b and facilitate
the development of new catalytic systems that utilize
inexpensive catalysts and more atom economical oxidants,
such as O2.
Oxidative Mannich reactions consist of two independent
steps − oxidation of N,N-dialkylaniline to the iminium ion and
reaction of the iminium ion or the associated intermediate with
a nucleophile (Scheme 2). Investigations of possible
aqueous solution of tert-butyl hydroperoxide (T-HYDRO)
not only allows comparison with other mechanistic models that
have been proposed for other metal-catalyzed oxida-
tions2c,5b,6a,19a but the results obtained can be extrapolated to
the more reactive tetra-isoquinolines.
Iminium Ions in the Mannich Reaction. From
investigations of possible intermediates leading to the Mannich
adduct (Scheme 2)3,5b,6a relating to the mechanism of catalytic
N,N-dialkylaniline oxidations by TBHP,5b,19a Doyle postulated
a competitive capture of an iminium ion by the solvent
methanol and siloxyfuran 1 in the oxidative Mannich reaction
(Scheme 4a).3 Che found that peroxy hemiaminal 7 formed in
Scheme 2. General Scheme of Oxidative Mannich Reaction
with TBHP
Scheme 4. Proposed Intermediates of in the Oxidative
Mannich Reaction with TBHP
intermediates leading to the Mannich adduct have received
considerable attention3,5b,6a while only a few mechanistic
studies have been reported on the transition metal-catalyzed
N,N-dialkylaniline oxidations by TBHP (discussed below).5b,19a
Dirhodium caprolactamate [Rh2(cap)4] is a useful catalytic
platform for mechanistic comparison of oxidations by TBHP
that proceed via transition metal−oxo complexes or tert-
butylperoxy radicals.20 Prior publications have shown unequiv-
ocally that the sole role of Rh2(cap)4 in reactions with TBHP is
to generate a flux of tert-butylperoxy radicals in both allylic
oxidation and the peroxidation of phenols (Scheme 3).20
Comparison of the characteristics of Rh2(cap)4-catalyzed
TBHP oxidations with those of with other catalysts can
overcome uncertainty in their mechanistic descriptions.
the Ru-catalyzed oxidative Mannich reaction performed in
toluene is converted to 8 under the reaction conditions
(Scheme 4b).5b However, no kinetic data were provided to
answer whether 7 is an intermediate or a side-product of a
competing pathway that can be converted to 8 under the
reaction conditions. Klussmann showed that the rate of
formation of 10 in the oxidative Mannich reaction has a
sigmoidal shape which indicates that peroxy hemiaminal 9 is a
direct precursor of 10 (Scheme 4c).6a Klussmann coined the
term “iminium ion reservoir” to describe the role of peroxy
hemiaminal 10. Li, on the other hand, postulated2c the
intermediacy of copper-bound iminium ion 11 (Scheme 4d)
although no experimental evidence for formation of 11 was
provided. The latter proposal was subsequently adopted by
several groups for the iron(III)7b and copper(I)6f-catalyzed
oxidative Mannich reaction.
We began investigations of the oxidative Mannich reaction
by focusing on understanding the role of intermediates in the
Mannich reaction (Scheme 2). The conditions previously
developed in our lab3 were adopted for the oxidative Mannich
reaction. We have reported the exclusive formation of 13 in the
N,N-dimethylaniline oxidation by T-HYDRO in methanol in
the absence of nucleophile and that 13 exists in an equilibrium
between 14 (eq 1),3 and we have selected the same conditions
to obtain data on the rate of exchange between 13 and 14.
Scheme 3. Mechanism of Rh2(cap)4-Catalyzed Conversion of
tert-Butyl Hydroperoxide to the tert-Butylperoxy Radicals
In this article we present a comprehensive mechanistic
investigations of the Rh2(cap)4-catalyzed oxidative Mannich
reaction with TBHP. The role of the transition metal−oxo
complexes in N,N-dialkylaniline oxidation and formation of
possible intermediates in nucleophilic and non-nucleophilic
media is assessed. The generality of the proposed mechanism is
evaluated in a comparative study of Rh, Ru, Cu, Fe, and Co
complexes.
RESULTS AND DISCUSSION
■
N,N-Dimethylanilines have higher oxidation potentials21 than
similar tetrahydroisoquinolines.22 The selection of N,N-
dimethylanilines5b,19a,23 for mechanistic investigations of the
Rh2(cap)4-catalyzed oxidative Mannich reaction with 70%
1550
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
Oxidation of N,N-dimethylaniline by T-HYDRO in methanol
catalyzed by Rh2(cap)4 furnishes 13 and 14.3 As previously
reported,3 13 is formed as a major product after 5 min, and the
ratio of 13 to 14 is 12.0 (Figure 1). It is of note that the ratio of
hemiaminal 13 is formed under kinetic control at the initial
stage and serves as a resting state for the iminium ion 15 that is
an immediate result of oxidation of 12.
Determination of the reactive intermediates for the
Rh2(cap)4-catalyzed N,N-dimethylaniline oxidation by T-
HYDRO raised the question of whether 13 and 14 are side
products2b,c,3,4 or iminium ion reservoirs.6a To distinguish the
two roles of 13 and 14, kinetic data were obtained for the
Rh2(cap)4-catalyzed oxidative Mannich reaction in methanol by
T-HYDRO of 12 and siloxyfuran 1 at 40 °C. The mechanistic
model of Li2c suggested that under these conditions 1 should
outcompete the less nucleophilic methanol and TBHP to yield
16 without the formation of peroxy or methoxy hemiaminals.
However, this hypothesis was not consistent with our
experimental data. The initial rate of consumption of 12 was
greater than the initial rate of siloxyfuran 1 consumption
(Figure 3a). Furthermore, methoxy hemiaminal 13 was the
Figure 1. Change in the ratio of 13-to-14 with time in the oxidation of
N,N-dimethylaniline by 2 equiv of T-HYDRO catalyzed by 0.5 mol %
of Rh2(cap)4.
methanol and TBHP in this reaction mixture is 12.4. Thus,
nearly identical ratios of 13-to-14 and methanol-to-TBHP
suggest that iminium ion 15 preferentially reacts with the
nucleophilic solvent, methanol (eq 2). Additionally, the larger
amount of 13 at the initial stages of 12 oxidation by TBHP is
not consistent with reaction between the α-amino radical and
the tert-butylperoxy radical leading to 14 in the last step of the
Rh2(cap)4-catalyzed N,N-dialkylaniline oxidation by T-
HYDRO.
The ratio of 13-to-14 decreased from 12.0 to 4.9 over 80 min
(Figure 1), suggesting that an equilibrium between 13 and 14
favoring 143 could play a role in a variable ratio. To evaluate
this hypothesis, peroxy hemiaminal 14 was prepared,3 and its
conversion to 13 in methanol at 0.14 M concentration occurred
on the time scale of the Rh2(cap)4-catalyzed N,N-dimethylani-
line oxidation by T-HYDRO (Figure 2). Identical ratios of 13
and 14 measured after 8 and 24 h gave the equilibrium constant
of 11.7 that favors 14 (eq 3). Thus, the data confirm the
proposal3 that the thermodynamically less stable methoxy
Figure 3. Kinetics of the Rh2(cap)4-catalyzed oxidative Mannich
reaction of N,N-dimethylaniline and siloxyfuran 1 by T-HYDRO in
methanol at 40 °C.
major product at low conversion (Figure 3b). The accumu-
lation of 13 in the reaction mixture at low conversion was
followed by a decrease in its concentration. A noticeable rate of
formation of 16 is evident only after the concentration of 13
reaches its maximum (Figure 3b). Interestingly, the concen-
tration of 14 remained lower than the concentration of 16,
indicating the preferential trapping of the iminium ion formed
from 13 by siloxyfuran 1 and not by TBHP.
The role of 13 and 14 as intermediates was evaluated with an
isotope-scrambling experiment. The Rh2(cap)4-catalyzed oxi-
dative Mannich reaction of d6-17 and 1 was conducted in the
presence of an excess of 37% aqueous solution of formaldehyde
(eq 4). The isolated product lacked any deuterium labeling in
Figure 2. Kinetics of equilibration of 14 to 13 in 0.14 M solution of
methanol at room temperature.
1551
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
Scheme 6. Summary of Conversion of an Iminium Ion
Formed in the Oxidation of N,N-Dialkylanilines to the
Mannich Adduct
its methylene group, and this result unequivocally showed that
hemiaminal equilibria are faster than iminium ion trapping by 1.
In other words, formation of the Mannich adduct of iminium
ion and 1 is a thermodynamic sink for the oxidative Mannich
reaction.
Having proposed a mechanistic model of the oxidative
Mannich reaction with T-HYDRO in a nucleophilic solvent, the
oxidative Mannich reaction by T-HYDRO in non-nucleophilic
toluene was investigated. Che showed the viability of these
conditions for the Ru-catalyzed oxidative Mannich reaction of
N,N-dialkylanilines with indoles.5b However, according to the
Che proposal, non-nucleophilic media should facilitate the
capture of an iminium ion by indoles in the oxidation of N,N-
dialkylanilines. When the oxidative Mannich reaction of 12 and
N-methylindole (19) using Rh2(cap)4 as a catalyst was
performed in toluene intermediates 14 and 20 were
preferentially formed initially (Scheme 5), but continuation of
oxidations by TBHP have been reported.5b,19a Murahashi19a
used primary isotope effects values of 1.64 and 3.53 and the
−0.84 ρ value in the Hammett linear free-energy relationship
(LFER) obtained for demethylation of N,N-dimethylanilines by
TBHP catalyzed by RuCl2(PPh3)3 to adopt a mechanism
postulated25 for cytochrome P450-catalyzed demethylation of
N,N-dimethylanilines by O2. According to this proposal the
ruthenium−oxo complex formed upon reaction of
RuCl2(PPh3)3 and TBHP abstracts two electrons and a proton
in a series of single electron transfers (SET) and hydrogen
atom transfer (HAT) steps (Scheme 7).19a Recently, Che
Scheme 7. Proposed Mechanism of RuCl2(PPh3)3-Catalyzed
N,N-Dialkylaniline Oxidation by TBHP
Scheme 5. Rh2(cap)4-Catalyzed Oxidative Mannich Reaction
of 12 and 19 in Toluene
obtained a Hammett ρ value of −1.09 and primary isotope
effect values of 1.2−3.8 in the oxidative Mannich reaction of
N,N-dimethylanilines and indoles catalyzed by 6.5b Further-
more, the ruthenium porphyrin oxo complex of 6 was found to
be a competent stoichiometric oxidant, and 6 catalyzed the
oxidation of N,N-dimethylaniline by TBHP and by 2,6-
dichloropyridine N-oxide. Li proposed two possible nonradical
mechanisms of copper(I)-catalyzed N,N-dialkylaniline oxida-
tions by TBHP involving copper−oxo complexes (Scheme 8a
and 8b). A decrease in the isolated yield of 23 from 84% to 70%
Scheme 8. Experiment with Radical Scavenger and Two
Mechanisms Proposed by Li for the CuBr-Catalyzed
Oxidative Mannich Reaction with TBHP As the Oxidant
this reaction for 3 h quantitatively converted 14 and 20 into the
final adduct 21 (Scheme 5). Thus, hemiaminals are the
precursors to the Mannich adducts even in non-nucleophilic
media.
In summary, the transition metal-catalyzed oxidation of N,N-
dimethylanilines by T-HYDRO forms an iminium ion that is
initially trapped by either nucleophilic solvent or TBHP. The
resulting methoxy and peroxy hemiaminals exist in the thermal
equilibrium with iminium ions that irreversibly form Mannich
adducts with a carbon nucleophile present in the reaction
(Scheme 6). It is unlikely that Rh2(cap)4, a weak Lewis acid,24
can catalyze the formation of iminium ions from oxidation
intermediates in the presence of large amounts of water.
Oxidative Mechanism of the Transition Metal-Cata-
lyzed Oxidation of N,N-Dialkylanilines by T-HYDRO.
Systematic mechanistic studies of catalyzed N,N-dialkylaniline
1552
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
in the copper(I)-catalyzed oxidative Mannich reaction
performed in the presence of the radical trap BHT (Scheme
8c) was used by Li to support nonradical pathways. However,
Huang6c and Klussmann6a stated that, although BHT has a
detrimental effect on the oxidative Mannich reaction, the
copper(I)-catalyzed oxidation of 22 has radical character.
Referencing the work of Minisci26 and Doyle,20b Klussmann
proposed hydrogen atom abstraction from the benzylic position
of 22 by the tert-butoxy radical followed by tert-butylperoxy
group transfer from 24 to α-amino radical 25 (Scheme 9) to
account for the primary isotope effect value of 3.4 and
deceleration of the Mannich reaction of 22 and nitromethane in
the presence of BHT.6a
Figure 4. Hammett LFER analysis of the Rh2(cap)4-catalyzed
oxidations by T-HYDRO of 4-substituted N,N-dimethylanilines at
room temperature in methanol. Each reaction rate was measured at
least three times, and the error margins were calculated as a standard
deviation of the reaction rates.
Scheme 9. Mechanism proposed by Klussmann for CuBr-
catalyzed oxidation of tetrahydroisoquinoline by TBHP
nature of the single electron acceptor. The higher oxidation
potential of the tert-butylperoxy radical (0.52 V vs SCE)27
relative to the tert-butoxy radical (−0.30 V vs SCE)28 and
[Rh2(cap)4]+ (11 mV vs SCE)29 defines the peroxy radical as
the thermodynamically preferred oxidant.
The nature of the C−H cleavage step was investigated by
kinetic and product isotope effects (KIE and PIE, respectively).
Kinetic isotope effects for Rh2(cap)4-catalyzed oxidation by T-
HYDRO were determined from the deuteration level of the
methylene and methyl groups of the corresponding methoxy
hemiaminals in oxidation of 1:1 mixtures of N,N-dimethylani-
lines and N,N-bis(trideuteromethyl)anilines (Scheme 10a).
Previous studies of oxidative Mannich reactions catalyzed by
transition metal compounds have led to divergent mechanistic
proposals: (1) rate-limiting steps SET from N,N-dialkylaniline
to a ruthenium−oxo complex in ruthenium-catalyzed reactions
(Scheme 7),5b,19a (2) hydrogen atom transfer (HAT) from the
1-position of 4-aryl-tetrahydroisoquinoline by the tert-butoxy
radical in copper(I)-catalyzed reactions (Scheme 9),6a and (3)
oxidation of copper followed by elimination of water in
copper(I)-catalyzed reactions (a and b of Scheme 8).2c,6f
Hence, the initial focus of our mechanistic investigation was to
discern the electron demand of the rate-determining step. The
ρ value of the Hammett analysis close to zero would suggest
HAT or oxidation of a metal center by TBHP as the rate-
determining step; a positive ρ value would indicate elimination
of water as the rate-determining step; finally, a negative ρ would
be consistent with a SET step.23
Scheme 10. KIE and PIE of the Rh2(cap)4-Catalyzed
Oxidations of 4-Substituted N,N-Dimethylanilines by T-
HYDRO
The Hammett LFER analysis was conducted for a series of
competition experiments of the Rh2(cap)4-catalyzed oxidations
by T-HYDRO of 4-methoxy- (27), 4-methyl- (17), 4-fluoro-
(28), 4-bromo- (29), 4-ethylcarboxylato- (30), 4-trifluorometh-
yl- (31), and 4-cyano-N,N-dimethylanilines (32) against 12.
The initial competition experiments were performed under the
conditions used for monitoring of oxidation of 12 by T-
HYDRO in methanol (0.5 mol % of Rh2(cap)4, 2 equiv of T-
HYDRO, 40 °C, methanol as a solvent).
The correlation of natural logarithms of the relative initial
rates was higher with σ+ Hammett parameters (R2 = 0.975)
than with σpara Hammett parameters (R2 = 0.909). The linear
correlation from 4-methoxy- to 4-cyano- groups was observed,
indicating that there was no change in the nature of the rate-
determining step for the studied range of 4-substituents (Figure
4). Additionally, the measured ρ value of −2.638 was consistent
with SET in the rate-determining step.19a,22,23
Product isotope effects were determined from the ratio of
integrated absorptions in H NMR spectra that corresponded
1
to methylene and methyl groups of methoxy hemiaminals in
oxidation of N-methyl-N-(trideuteromethyl)anilines (Scheme
10b). Each measurement of KIE and PIE was repeated at least
three times, and the average value along with error margins are
reported. The measured values of KIE and PIE values remained
on the level of 1.61 0.08 and 2.39 0.24, respectively, from
17 to 29 (Figure 5). The KIE values start to increase for the 30
and approach PIE values for 32.
The characteristic gap between KIE and PIE values,
previously reported by Watanabe23 for the enzymatic
demethylations of N,N-dimethylanilines catalyzed by cyto-
chrome P450, was also observed for Rh2(cap)4-catalyzed
oxidations of 4-substituted N,N-dimethylanilines by T-
HYDRO. According to the steady-state kinetic model of
The Hammett analysis was consistent with the SET
mechanism but did not provide information regarding the
1553
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
−0.311), 12 (σ+ = 0.00), and 29 (σ+ = 0.150) were chosen to
represent substrates of different electronic properties. The KIE
and PIE values were determined in the same fashion as KIE and
PIE of the Rh2(cap)4-catalyzed oxidations (Scheme 10). Thus,
the measurements of KIE and PIE values for three different
substrates totals up to six independent measurements of
transition metal salt-catalyzed oxidation of N,N-dimethylani-
lines by T-HYDRO.
The comparison of KIE (Table 1) and PIE (Table 2) values
shows identical trends as a function of N,N-dimethylaniline
substituent. The identical isotope effects in each table are above
the dotted line (Tables 1 and 2). The KIE and PIE numbers
determined for oxidations of three N,N-dimethylanilines by T-
HYDRO catalyzed by Rh2(cap)4 and RuCl3 show identical
values within experimental error that suggests the same
operating mechanism for RuCl3 as for Rh2(cap)4. Thus, the
role of Rh2(cap)4 and RuCl3 is to convert TBHP into tert-
butylperoxy radicals according to the mechanism described in
Scheme 3. The copper(I) bromide-catalyzed oxidation shows
values identical to those determined for Rh2(cap)4-catalyzed
oxidation of 12 and 17 but not that of 29, which shows slightly
higher values. Iron(III) and cobalt(II) salts-catalyzed oxidations
have consistently higher KIE and PIE values than those
measured for the Rh2(cap)4-catalyzed oxidation.
Figure 5. KIE and PIE measured for the Rh2(cap)4-catalyzed oxidative
Mannich reaction by T-HYDRO. Each measurement was repeated at
least three times, and the error margins were calculated as a standard
deviation of the measured isotope effects.
Watanabe,23 a rate-determining forward SET forms a constant
concentration of cation radicals that are consumed either by a
competing backward SET or by an irreversible C−H bond
cleavage (Scheme 11a). The relationship between KIE and PIE
Scheme 11. Kinetic Model of the Rh2(cap)4-Catalyzed
Oxidations of 4-Substituted N,N-Dimethylanilines by T-
HYDRO
The storage of methanol solutions opened to air of 12 and
d6-12 in the presence of Co(OAc)2 and 17 and d6-17 in the
presence of RuCl3 (Table 3, entries 6−7) offered a clue to the
explanation of the observed deviations of KIE and PIE for
CuBr, FeCl3, and Co(OAc)2. In these reactions N,N-
dimethylanilines underwent partial oxidation by dioxygen in
air; in the presence of RuCl3 and Co(OAc)2, KIE values were
2.09
0.11 and 3.36
0.33, respectively, from reactions
was derived by Watanabe23 from this kinetic model and is
presented in Scheme 11b. The PIE value depends only on the
C−H bond cleavage step and is determined by its late
transition state. On the other hand, KIE values additionally
depend on the backward SET (k−SET). Thus, KIE values
approach 1.0 if the rate of backward SET is slower than the rate
of the irreversible C−H cleavage (kPIE > k−SET). Alternatively, a
higher KIE value is a result of a higher rate of backward SET
relative to the rate of the irreversible step (kPIE < k−SET). In the
data from Figure 5, oxidation of 32 is the limiting case of k−SET
≫ kPIE where the KIE value (3.20) approached PIE value
(3.26).
KIE and PIE values measured for the Rh2(cap)4-catalyzed
oxidation of 4-substituted N,N-dimethylanilines by T-HYDRO
are two independent experimental measurements that offer an
answer to the question of whether RuCl2(PPh3)3, CuBr, FeCl3,
and Co(OAc)2 form oxo complexes that act as sole oxidants
under the same conditions. N,N-Dimethylanilines 17 (σ+ =
performed overnight at room temperature. Noteworthy, the
value measured for the Co(OAc)2-catalyzed N,N-dimethylani-
line oxidation by T-HYDRO was between the KIE values of
tert-butylperoxy radical oxidative pathway (entry 1) and the
dioxygen oxidative pathway (entry 6). This fact indicated that
the values measured for CuBr, FeCl3, and Co(OAc)2-catalyzed
oxidations of 4-substituted N,N-dimethyl-anilines could be a
combination of the competing tert-butylperoxy radical oxidative
pathway and a dioxygen oxidative pathway. If this is the case,
bubbling N2 through the reaction mixture would decrease the
contribution from the of dioxygen oxidative pathway by
removal of most of the dioxygen originating from air and
from dimerization of the tert-butylperoxy radicals (eq 5).
Indeed, N2 flow decreased KIE values for Co-catalyzed
oxidation of 12 to 2.17 0.05 and CuBr-catalyzed oxidation
of 29 to 1.58 0.06. Consequently, dioxygen interference was
confirmed for the Cu(I), Fe(III), and Co(II) salts-catalyzed
oxidations of 4-substituted N,N-dimethylanilines by T-
Table 1. Kinetic Isotope Effects of the Transition Metal Salt-Catalyzed Oxidations of 4-Substituted N,N-Dimethylanilines by T-
HYDRO
KIE
entry
catalyst
catalyst loading, %
4-Me (17)
4-H (12)
4-Br (29)
a
1
2
3
4
5
Rh2(cap)4
RuCl3
1.0
1.0
1.63 0.03
1.69 0.16
1.58 0.07
1.91 0.01
1.94 0.03
1.71 0.08
1.66 0.18
1.79 0.01
1.95 0.02
2.42 0.16
1.54 0.02
1.55 0.06
1.71 0.15
1.67 0.02
1.69 0.02
CuBr
5.0
FeCl3
2.0
Co(OAc)2
10.0
a
Error margins are calculated as a standard deviation of measured kinetic isotope effects.
1554
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
Table 2. Product Isotope Effects of the Transition Metal Salt-Catalyzed Oxidations of 4-Substituted N,N-Dimethylanilines by T-
HYDRO
PIE
entry
catalyst
catalyst loading, %
4-Me (17)
4-H (12)
4-Br (29)
a
1
2
3
4
5
Rh2(cap)4
RuCl3
1.0
1.0
2.15 0.02
2.20 0.06
2.21 0.10
2.46 0.07
4.84 0.16
2.46 0.22
2.37 0.06
2.36 0.04
3.16 0.20
4.42 0.23
2.63 0.01
2.55 0.08
3.04 0.14
2.62 0.07
2.93 0.12
CuBr
5.0
FeCl3
2.0
Co(OAc)2
10.0
a
Error margins are calculated as a standard deviation of measured kinetic isotope effects.
Table 3. Kinetic Isotope Effects of the Transition Metal Salt-Catalyzed Oxidations of 4-Substituted N,N-Dimethylanilines by T-
HYDRO with and without N2 Flow and by O2
KIE
entry
catalyst
RuCl3
catalyst loading, %
oxidant
N2 flow
4-Me (17)
4-H (12)
4-Br (29)
a
1
2
3
4
5
6
7
1.0
5.0
T-HYDRO
T-HYDRO
T-HYDRO
T-HYDRO
T-HYDRO
O2
no
no
no
yes
yes
no
no
1.69 0.16
1.58 0.07
1.94 0.03
1.66 0.18
1.79 0.01
2.42 0.16
nd
1.55 0.06
1.71 0.15
1.69 0.02
1.58 0.06
nd
CuBr
Co(OAc)2
CuBr
10.0
5.0
b
nd
Co(OAc)2
RuCl3
10.0
1.0
nd
2.17 0.05
2.09 0.11
nd
nd
nd
Co(OAc)2
10.0
O2
3.36 0.33
nd
a
b
Error margins are calculated as a standard deviation of measured kinetic isotope effects. nd - not determined.
HYDRO. Furthermore, the KIE value of 1.58 0.06 is identical
within experimental error to the KIE values obtained for
Rh2(cap)4, RuCl3, and CuBr, thus indicating the same operating
mechanism for these transition metal catalysts.
Scheme 13. Rate of Cyclopropane-Opening Rates of Radical
Trap Substituents
t‐BuOO· + ·OOt‐Bu → t‐BuOOt‐Bu + O2
(5)
The two possible pathways for the irreversible C−H cleavage
step include proton transfer (PT) forming an α-amino radical
35 (Scheme 12a) and hydrogen atom transfer (HAT) leading
Scheme 14. Rh2(cap)4-Catalyzed Oxidations by T-HYDRO
of N,N-Dialkylanilines Bearing Radical Trap Substituents
Scheme 12. Possible Mechanisms of the Irreversible C−H
Cleavage in Transition Metal Salt-Catalyzed Oxidation of
N,N-Dimethylanilines
directly to the iminium ion 34 (Scheme 12b). Radical trap
substituents were employed to test the possible formation of a
transient α-amino radical. These substituents do not undergo
ring-opening in the presence of adjacent onium cations,30 but
they show high rates of intramolecular ring-opening when they
are adjacent to a carbon radical center (Scheme 13).31
oxidation of 40 that had a more sensitive radical trap
substituent (Scheme 14b) led to decomposition of N,N-
dialkylaniline with the formation of unknown products with
high-molecular weight. p-Cymene was added as a hydrogen
atom donor to minimize possible radical side reactions of the
resulting benzyl radical (Scheme 13b). Indeed, the oxidation of
40 in 1:1 mixture of methanol and p-cymene furnished a
number of absorptions corresponding to aldehydes according
Conditions developed in our lab3 for the oxidative Mannich
reaction were chosen for the radical trap experiments (Scheme
14). The Rh2(cap)4-catalyzed oxidation by T-HYDRO of 36
formed its methoxy hemiaminal in 65% yield without
cyclopropane ring-opening (Scheme 14a) that was seemingly
consistent with either the HAT step (Scheme 12b) or with a
SET that was much faster than cyclopropane ring-opening. The
1555
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
1
to the H NMR spectroscopy of the reaction mixture. Free 4-
cally preferred oxidant, in the rate-determining step that is
uniform between 4-methoxy- and 4-cyano-substituted N,N-
dimethylanilines. The difference between KIE and PIE of N,N-
dimethylaniline oxidations by THBP is explained by the
competition between backward SET and irreversible proton
transfer after the rate-determining SET. The formation of a
transient α-amino radical is consistent with the cyclopropane
ring-opening in the radical trap experiments. No experimental
evidence of kinetic competence of ruthenium−oxo complex
was found in the RuCl3-catalyzed N,N-dialkylaniline oxidation
by TBHP. Unlike earlier models involving ruthenium−5b,19a
and copper−oxo complexes2c, the general role of Rh2(cap)4,
RuCl3, CuBr, FeCl3, and Co(OAc)2 in converting TBHP to the
tert-butylperoxy radical is consistent with their identical KIE
and PIE values for oxidations of 12, 17, and 29 by T-HYDRO
in methanol and the decrease of KIE and PIE values of CuBr
and Co(OAc)2-catalyzed oxidations with N2 flow. The role of
hemiaminals as precursors to Mannich adducts and not side
products3 was determined by monitoring the initial stage of the
Rh2(cap)4-catalyzed oxidative Mannich reaction of 12 and 1 in
methanol and 12 and N-methylindole in toluene.
phenylbutanal was trapped with 2,4-dinitrophenylhydrazine
(38) and was isolated in 15% yield as 41 along 12% of
hydrazone 42, suggesting that the pathway described in Scheme
12a is operative with a very fast SET that follows the PT step
which occurs in competition with cyclopropane ring-opening of
the cyclopropylcarbinyl radical generated from 40. However, an
alternate interpretation that the two pathways that were
outlined in a and b of Scheme 12 are competitive cannot be
excluded.
Opening of the cyclopropane ring is consistent with the
formation of a transient α-amino radical (Scheme 12a) in the
transition metal salt-catalyzed N,N-dialkylaniline oxidation by
T-HYDRO. Proton transfer should occur at the most acidic
carbon of unsymmetrical N,N-dialkylaniline cation radicals.
According to the report of Albini, Falvey, and Mariano,32 the
acidity of carbon adjacent to nitrogen decreases in order
Me,Me > Me,H > H,H > H,CHCH2 > H,Ph > H, CCH. Thus,
irreversible PT offers the first explanation of the observed
regioselectivity of the oxidation of unsymmetrical N,N-dialkylani-
lines by T-HYDRO.3 Steric factors may also play a role in
regioselectivity because the oxidative Mannich reaction of N-
methyl-N-ethylaniline and siloxyfuran 1 occurs only at the N-
methyl group (eq 6) although the acidity ratio suggests 3:1
regioselectivity.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures and product characterization are
included in Supporting Information. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
mdoyle3@umd.edu
■
Notes
The authors declare no competing financial interest.
CONCLUSION
■
In conclusion, the mechanistic models of Murahashi,19a Che,5b
Li,2c and Klussmann6a of the transition metal salt-catalyzed
oxidative Mannich reaction of N,N-dialkylaniline with TBHP
were evaluated, and a new mechanistic scheme is proposed
(Scheme 15).
ACKNOWLEDGMENTS
■
Support for this research from the National Science Foundation
(CHE- 0748121) is gratefully acknowledged. We thank Jason
Nichols, Caitlyn Edgley, James Myslinski, and Dmitrijs
Sabasovs for their efforts in obtaining preliminary data for
many of these studies.
A combination of LFER analysis and the oxidation potentials
of [Rh2(cap)4] and the tert-butylperoxy radical suggests that,
unlike the Li2c and Klussmann6a proposals, N,N-dialkylanilines
undergo SET to tert-butylperoxy radicals, the thermodynami-
REFERENCES
■
(1) (a) Casiraghi, G.; Battistini, L.; Curti, C.; Rassu, G.; Zanardi, F.
Chem. Rev. 2011, 111, 3076. (b) Toure, B. B.; Hall, D. G. Chem. Rev.
2009, 109, 4439.
Scheme 15. Summary of Mechanistic Investigations of
Transition Metal-Catalyzed Oxidative Mannich Reaction
with TBHP as an Oxidant
(2) (a) Murahashi, S.-i.; Komiya, N. Ruthenium-catalyzed oxidation
for organic synthesis. In Modern Oxidation Methods, 2nd ed.; Baeckvall,
J.-E., Ed. Wiley-VCH: Weinheim, Germany, 2010; pp 241; (b) Li, C.-J.
Acc. Chem. Res. 2009, 42, 335. (c) Li, Z.; Bohle, D. S.; Li, C.-J. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 8928. (d) Arend, M.; Westermann,
B.; Risch, N. Angew. Chem., Int. Ed. 1998, 37, 1045.
(3) Catino, A. J.; Nichols, J. M.; Nettles, B. J.; Doyle, M. P. J. Am.
Chem. Soc. 2006, 128, 5648.
(4) (a) Murahashi, S.-I.; Nakae, T.; Terai, H.; Komiya, N. J. Am.
Chem. Soc. 2008, 130, 11005. (b) Jones, K. M.; Klussmann, M. Synlett
2012, 23, 159.
(5) (a) Rueping, M.; Vila, C.; Koenigs, R. M.; Poscharny, K.; Fabry,
D. C. Chem. Commun. 2011, 47, 2360. (b) Wang, M.-Z.; Zhou, C.-Y.;
Wong, M.-K.; Che, C.-M. Chem.Eur. J. 2010, 16, 5723.
(c) Murahashi, S.; Komiya, N.; Terai, H.; Nakae, T. J. Am. Chem.
Soc. 2003, 125, 15312. (d) Condie, A. G.; Gonzalez-Gomez, J. C.;
Stephenson, C. R. J. J. Am. Chem. Soc. 2010, 132, 1464.
(6) (a) Boess, E.; Schmitz, C.; Klussmann, M. J. Am. Chem. Soc. 2012,
134, 5317. (b) Boess, E.; Sureshkumar, D.; Sud, A.; Wirtz, C.; Fares,
1556
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557

Journal of the American Chemical Society
Article
C.; Klussmann, M. J. Am. Chem. Soc. 2011, 133, 8106. (c) Yang, F.; Li,
J.; Xie, J.; Huang, Z.-Z. Org. Lett. 2010, 12, 5214. (d) Zhao, L.; Li, C.-J.
Angew. Chem., Int. Ed. 2008, 47, 7075. (e) Xie, J.; Huang, Z.-Z. Angew.
Chem., Int. Ed. 2010, 49, 10181. (f) Ghobrial, M.; Schnuerch, M.;
Mihovilovic, M. D. J. Org. Chem. 2011, 76, 8781.
(7) (a) Han, W.; Ofial, A. R. Chem. Commun. 2009, 5024.
(b) Kumaraswamy, G.; Murthy, A. N.; Pitchaiah, A. J. Org. Chem.
2010, 75, 3916. (c) Liu, P.; Zhou, C.-Y.; Xiang, S.; Che, C.-M. Chem.
Commun. 2010, 46, 2739.
(8) Alagiri, K.; Prabhu, K. R. Org. Biomol. Chem. 2012, 10, 835.
(9) (a) Sud, A.; Sureshkumar, D.; Klussmann, M. Chem. Commun.
2009, 3169. (b) Singhal, S.; Jain, S. L.; Sain, B. Chem. Commun. 2009,
2371.
(10) Volla, C. M. R.; Vogel, P. Org. Lett. 2009, 11, 1701.
(11) (a) Sureshkumar, D.; Sud, A.; Klussmann, M. Synlett 2009,
1558. (b) Chu, L.; Zhang, X.; Qing, F.-L. Org. Lett. 2009, 11, 2197.
(12) Li, Z.; Li, C.-J. J. Am. Chem. Soc. 2005, 127, 6968.
(13) Murata, S.; Teramoto, K.; Miura, M.; Nomura, M. Bull. Chem.
Soc. Jpn. 1993, 66, 1297.
(14) Shen, Y.; Li, M.; Wang, S.; Zhan, T.; Tan, Z.; Guo, C.-C. Chem.
Commun. 2009, 953.
(15) (a) Li, Z.; Li, C.-J. J. Am. Chem. Soc. 2005, 127, 3672. (b) Basle,
O.; Li, C.-J. Green Chem. 2007, 9, 1047. (c) Zeng, T.; Song, G.;
Moores, A.; Li, C.-J. Synlett 2010, 2002.
(16) Li, Z.; Li, C.-J. Eur. J. Org. Chem. 2005, 3173.
(17) Ohta, M.; Quick, M. P.; Yamaguchi, J.; Wuensch, B.; Itami, K.
Chem.Asian J. 2009, 4, 1416.
(18) Basle, O.; Li, C.-J. Org. Lett. 2008, 10, 3661.
(19) (a) Murahashi, S.; Naota, T.; Yonemura, K. J. Am. Chem. Soc.
1988, 110, 8256. (b) Zhang, Y.; Fu, H.; Jiang, Y.; Zhao, Y. Org. Lett.
2007, 9, 3813. (c) Basle, O.; Li, C.-J. Chem. Commun. 2009, 4124.
(d) Han, W.; Ofial, A. R. Chem. Commun. 2009, 6023.
(20) (a) Ratnikov, M. O.; Farkas, L. E.; McLaughlin, E. C.; Chiou,
G.; Choi, H.; El-Khalafy, S. H.; Doyle, M. P. J. Org. Chem. 2011, 76,
2585. (b) McLaughlin, E. C.; Choi, H.; Wang, K.; Chiou, G.; Doyle,
M. P. J. Org. Chem. 2009, 74, 730.
(21) Dombrowski, G. W.; Dinnocenzo, J. P.; Zielinski, P. A.; Farid,
S.; Wosinska, Z. M.; Gould, I. R. J. Org. Chem. 2005, 70, 3791.
(22) Basle, O.; Borduas, N.; Dubois, P.; Chapuzet, J. M.; Chan, T.-
H.; Lessard, J.; Li, C.-J. Chem.Eur. J. 2010, 16, 8162.
(23) Goto, Y.; Watanabe, Y.; Fukuzumi, S.; Jones, J. P.; Dinnocenzo,
J. P. J. Am. Chem. Soc. 1998, 120, 10762.
(24) Xu, X.; Doyle, M. P. Inorg. Chem. 2011, 50, 7610.
(25) (a) Heimbrook, D. C.; Murray, R. I.; Egeberg, K. D.; Sligar, S.
G.; Nee, M. W.; Bruice, T. C. J. Am. Chem. Soc. 1984, 106, 1514.
(b) Burka, L. T.; Guengerich, F. P.; Willard, R. J.; Macdonald, T. L. J.
Am. Chem. Soc. 1985, 107, 2549.
(26) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Banfi, S.;
Quici, S. J. Am. Chem. Soc. 1995, 117, 226.
(27) Das, T. N.; Dhanasekaran, T.; Alfassi, Z. B.; Neta, P. J. Phys.
Chem. A 1998, 102, 280.
(28) Bietti, M.; Di Labio, G. A.; Lanzalunga, O.; Salamone, M. J. Org.
Chem. 2010, 75, 5875.
(29) Das, K.; Kadish, K. M.; Bear, J. L. Inorg. Chem. 1978, 17, 930.
(30) Barrett, A. G. M.; Doubleday, W. W.; Tustin, G. J.; White, A. J.;
Williams, D. J. J. Chem. Soc., Chem. Commun. 1994, 1783.
(31) (a) Beckwith, A. L. J.; Bowry, V. W. J. Org. Chem. 1989, 54,
2681. (b) Newcomb, M.; Manek, M. B. J. Am. Chem. Soc. 1990, 112,
9662.
(32) Zhang, X.; Yeh, S.-R.; Hong, S.; Freccero, M.; Albini, A.; Falvey,
D. E.; Mariano, P. S. J. Am. Chem. Soc. 1994, 116, 4211.
1557
dx.doi.org/10.1021/ja3113559 | J. Am. Chem. Soc. 2013, 135, 1549−1557