A general mechanism for the copper- and silver-catalyzed olefin aziridination reactions: Concomitant involvement of the singlet and triplet pathways

Article abstract of DOI:10.1021/ja307229e

The olefin aziridination reactions catalyzed by copper and silver complexes bearing hydrotris(pyrazolyl)borate (Tp^x) ligands have been investigated from a mechanistic point of view. Several mechanistic probe reactions were carried out, specifically competition experiments with p-substituted styrenes, stereospecificity of olefins, effects of the radical inhibitors, and use of a radical clock. Data from these experiments seem to be contradictory, as they do not fully support the previously reported concerted or stepwise mechanisms. But theoretical calculations have provided the reaction profiles for both the silver and copper systems with different olefins to satisfy all experimental data. A mechanistic proposal has been made on the basis of the information that we collected from experimental and theoretical studies. In all cases, the reaction starts with the formation of a metal-nitrene species that holds some radical character, and therefore the aziridination reaction proceeds through the radical mechanism. The silver-based systems however hold a minimum energy crossing point (MECP) between the triplet and closed-shell singlet surfaces, which induce the direct formation of the aziridines, and stereochemistry of the olefin is retained. In the case of copper, a radical intermediate is formed, and this intermediate constitutes the starting point for competition steps involving ring-closure (through a MECP between the open-shell singlet and triplet surfaces) or carbon-carbon bond rotation, and explains the loss of stereochemistry with a given substrate. Overall, all the initially contradictory experimental data fit in a mechanistic proposal that involves both the singlet and the triplet pathways.

Full text of DOI:10.1021/ja307229e

Article
pubs.acs.org/JACS
A General Mechanism for the Copper- and Silver-Catalyzed Olefin
Aziridination Reactions: Concomitant Involvement of the Singlet and
Triplet Pathways
Lourdes Maestre,† W. M. C. Sameera,‡ M. Mar Díaz-Requejo,*,† Feliu Maseras,*,‡,§ and Pedro J. Perez*,†
́
†Laboratorio de Catal
Investigacion en Química Sostenible (CIQSO), Universidad de Huelva, Campus de El Carmen, 21007 Huelva, Spain
‡Institute of Chemical Research of Catalonia (ICIQ), Av. Països Catalans 16, 43007 Tarragona, Spain
§Departament de Química, Universitat Auton
oma de Barcelona, 08193 Bellaterra, Spain
́ ́
isis Homogenea, Departamento de Química y Ciencia de los Materiales, Unidad Asociada al CSIC, Centro de
́
̀
S
* Supporting Information
ABSTRACT: The olefin aziridination reactions catalyzed by
copper and silver complexes bearing hydrotris(pyrazolyl)-
borate (Tpx) ligands have been investigated from a
mechanistic point of view. Several mechanistic probe reactions
were carried out, specifically competition experiments with p-
substituted styrenes, stereospecificity of olefins, effects of the
radical inhibitors, and use of a radical clock. Data from these
experiments seem to be contradictory, as they do not fully
support the previously reported concerted or stepwise
mechanisms. But theoretical calculations have provided the
reaction profiles for both the silver and copper systems with
different olefins to satisfy all experimental data. A mechanistic
proposal has been made on the basis of the information that
we collected from experimental and theoretical studies. In all cases, the reaction starts with the formation of a metal−nitrene
species that holds some radical character, and therefore the aziridination reaction proceeds through the radical mechanism. The
silver-based systems however hold a minimum energy crossing point (MECP) between the triplet and closed-shell singlet
surfaces, which induce the direct formation of the aziridines, and stereochemistry of the olefin is retained. In the case of copper, a
radical intermediate is formed, and this intermediate constitutes the starting point for competition steps involving ring-closure
(through a MECP between the open-shell singlet and triplet surfaces) or carbon−carbon bond rotation, and explains the loss of
stereochemistry with a given substrate. Overall, all the initially contradictory experimental data fit in a mechanistic proposal that
involves both the singlet and the triplet pathways.
been the most frequently employed, and undergo highly
selective transformations with good yields.1 Isolated or in situ
generated hypervalent iodine reagents are the common nitrene
sources (Scheme 1) with azides being the other preferred
reactant. There are still uncertainties concerning the mecha-
nism that governs this transformation. It has been suggested
that the metal−nitrene intermediate is the active intermediate
for this reaction, which is electrophilic in nature. Further, such
species have been isolated or spectroscopically detected in a few
cases: Che and co-workers developed a ruthenium-based
system capable of promoting the nitrene transfer reaction to
olefins with a diimido complex shown in Scheme 2a.3 In the
case of copper, Warren and co-workers isolated10a a nitrene-
bridged dinuclear compound, and is active in nitrene transfer
reactions (Scheme 2b), whereas the group of Ray has recently
trapped a copper nitrene intermediate with scandium(III)
INTRODUCTION
■
The conversion of olefins into aziridines by the metal-catalyzed
addition of a nitrene NR group to the CC bond constitutes a
useful tool in the preparation of these three-member rings
(Scheme 1).1 Nine out of the fifteen metals from groups 7−11
have been reported to induce such transformation.2−9 Among
them, rhodium-, ruthenium-, and copper-based catalysts have
Scheme 1. The Metal-Catalyzed Olefin Aziridination
Reaction
Received: July 24, 2012
Published: December 31, 2012
© 2012 American Chemical Society
1338
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
a
Scheme 2. Metal-Nitrene Complexes and Reactions Involving Them
a
(a−d) Metal−nitrene complexes relevant to olefin aziridination reaction. (e) The singlet and triplet states of the metal-nitrenes. (f) A simple vision
of the two commonly proposed pathways for this transformation. (g−i) Mechanistic probes.
triflate.10b Zhang, de Bruin, and collaborators have recently
reported the EPR detection of paramagnetic cobalt nitrene
species that are responsible for both olefin aziridination and C−
H amidation (Scheme 2c).11 Also, an iron−imido complex,
active for nitrene transfer reactions, has been isolated by Betley
and co-workers (Scheme 2d).12 The metal−nitrene intermedi-
ate plays a crucial role in the subsequent catalytic steps. Two
electronic states, singlet and triplet (Scheme 2e) are possible
for the “M-NR” species (Scheme 2f shows a deceptively simple
representation of both). The triplet species has been usually
associated to a stepwise mechanism, whereas a concerted
pathway has been frequently assigned as a consequence of the
singlet state (Scheme 2f). The presence of the radical or the
concerted paths were established on the basis of a series of
model experiments, from which a typical behavior is expected
depending on the nature of the nitrene intermediate. Thus,
competition experiments with p-substituted styrenes (Scheme
2g) have provided either linear dependence with the plain
Hammett’s equation (concerted) or with a dual-parameter
equation, involving radical constants (stepwise). The use of
radical clocks as substrates (Scheme 2h) or the addition of
radical inhibitors have also been employed as probes. However,
most frequently performed experiments considered olefins with
a given stereochemistry (Z or E) to study the stereospecificity
of the reaction (Scheme 2i), where retention of the initial
geometry of the olefin supported a concerted mechanism, while
the observance of a certain loss of stereoselectivity indicated a
stepwise mechanism.
of loss of stereochemistry with aryl-substituted ole-
fins.2,3,7−9,13,14 With some of those catalysts, alkyl-substituted
olefins performed aziridination in a stereospecific manner,7a,14
whereas other systems showed a partial loss of stereochemistry
during the catalytic reaction.15,16 On the other hand, rhodium-
based catalysts induced a complete retention of the initial
geometry with both aryl- and alkyl-containing olefins.17 With
this metal, competition experiments with 3- and 4-substituted
styrenes gave a good correlation with the Hammett equation.
The same behavior was described for a series of cinnamate
esters with copper-based catalysts.18 However, Hammett
studies were contradictory with the results obtained from
stereospecificity experiments.3,19 This diversity has led for years
to propose that the mechanism of this reaction is highly system-
dependent and is always under the influence of the concerted or
stepwise routes.
Computational studies of transition metal-catalyzed olefin
aziridination reactions have been reported in the litera-
ture,13,18,20 but the description of a mechanistic picture
compatible with all experimental observations is yet unde-
scribed. The metal−nitrene complexes have been suggested as
the active intermediates for two-electron transfer aziridination,
but questions remain on the spin multiplicity (singlet or triplet)
of the species throughout the reaction. Density functional
approaches (DFT) indicated that the ground state of the
metal−nitrene is a triplet,20a−c while complete active space self-
consistent-field (CASSCF) calculations suggested a singlet
ground state, with only a small energy gap between the singlet
and triplet.20d Previous studies on reactivity seem to favor a
stepwise mechanism through a triplet state. Andersson and
Norrby demonstrated the involvement of singlet and triplet
biradical species in the aziridination of alkenes catalyzed by the
(N−N)CuNSO2Ar (N−N = chelating diamine) system.20a
The aforementioned experimental observations implied that
the concerted or radical pathways could be explained
straightforwardly, which is not always true. For example, it is
common to observe that stereoselection depends on the olefin.
Several systems have been described to induce a certain degree
1339
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
Comba and co-workers performed experimental and theoretical
studies of the aziridination reaction catalyzed by (bispidine)-
copper complexes,13 where formation of metal−nitrene is the
key for efficient catalytic aziridination via the stepwise
mechanism, involving a radical metal−nitrene intermediate.
Zhang and de Bruin reported a cobalt−nitrene intermediate in
the Co-porphyrin-mediated aziridination of alkenes that occurs
through the stepwise mechanism.20c The radical mechanism
would be seemingly at odds with the experimental observation
of retention of configuration in a number of cases. In the
presence of a low-energy gap between the singlet and triplet
states, the aziridination reactions may proceed through the both
singlet and triplet energy profiles, which is a common scenario
in two-state reactivity (TSR) or multistate reactivity (MSR).21
Spin crossover between both surfaces is likely feasible through
spin−orbital coupling at the metal.22
mechanism that explains all of them. This is the first case, to the
best of our knowledge, in which with such an array of
experimental data that seems to be inconsistent, a mechanistic
picture that involves both the singlet and triplet pathways is
proposed.
RESULTS AND DISCUSSION
■
Experimental Studies. To collect a full set of data that
could be used to propose a pathway for the nitrene transfer
reaction, we have performed the following experiments with a
series of TpxM complexes (M = Cu, Ag, Scheme 3): (i) study
the stereospecificity of the reaction with several aryl- and alkyl-
substituted olefins; (ii) competition experiments with p-
substituted styrenes, (iii) effect of the addition of radical
inhibitors, and (iv) use of a radical clock.
Aziridination of Z- or E-Olefins: Study of the Stereo-
specificity of the Reaction. Retention or certain loss of
stereochemistry of olefin substituents in the aziridines (Scheme
2h) have been usually invoked to propose either a concerted or
a stepwise mechanism, respectively. To check this issue, we
have carried out several experiments with a series of TpxM (M
= Cu, Ag) complexes and four different olefins. As shown in
Scheme 3, Z-2-pentene, E-β-methyl-styrene, and E-2-hexene
were converted into the corresponding aziridines with complete
retention of the stereochemistry. This was observed with
several catalysts bearing quite different Tpx ligands. These
results are in contrast with those previously reported with E,E-
hexadien-1-ol when copper- or silver-based catalysts were
employed.24 The latter gave the trans-aziridine in an exclusive
manner, whereas in the copper case, a certain loss of the
stereochemistry was observed (ca. 2:1 favoring the trans-
aziridines). Therefore, the group of olefins now tested, initially
Z or E, with alkyl or aryl-containing groups, indicate that the
retention of stereochemistry was observed with the exception of
that containing an additional hydroxylic functional group.
Further, this group seems to be noninnocent in the nitrene
transfer reaction with the copper-based catalyst.
Our group has some experience in the development of
copper-23 and silver-based24 catalysts containing trispyrazolyl-
borate ligands (Tpx, see Scheme 3) for the olefin aziridination
Scheme 3. Stereospecificity of the Olefin Aziridination
Reaction
In the seminal work of this area,7a Evans and co-workers
observed that alkyl-substituted olefins lead to stereospecific
aziridinations, whereas nonstereospecific reactions were
observed with aryl-containing olefins with simple copper salts
as catalyst precursors. This is the common pattern with copper-
based catalysts, with some exceptions, as that in our system in
which we observed retention with alkyl and aryl-substituted
olefins (excluding the OH case). Therefore, any plausible
mechanism for this transformation should contain an
explanation for the distinct behavior of different olefins in
their aziridination with the same catalyst
reaction. To date, we have mainly focused on the catalytic
activity, with a scarce incursion in the mechanism, and more
specifically, on a comparison between the olefin aziridination
and cyclopropanation reactions catalyzed by the complex
TpMe2Cu(C2H4) (TpMe2 = hydrotris(3,5-dimethyl-1-pyrazolyl-
borate).23b On the basis of the lack of a general mechanistic
perspective of this relevant reaction, we have now performed a
complete study (60 different experiments) in which a series of
copper- and silver complexes of composition TpxM (M = Cu,
Ag) has been exhaustively tested in the probe reactions shown
in Scheme 2. Thus, six different catalysts have been employed
for Hammett studies and stereoretention experiments with four
olefins. A radical clock has been employed with representative
copper and silver catalysts and the effect of a radical inhibitor
such as t-butylhydroxytoluene (BHT) has also been inves-
tigated with both metals. From all these experiments we have
collected contradictory data that, in principle, cannot support
an unambiguous mechanistic pathway, intended as concerted or
stepwise. But DFT studies have shown that all those data, albeit
contradictory at the first sight, are the consequence of a reaction
Competition Experiments with p-Substituted Styrenes:
The Dual-Parameter Hammett Equation. As mentioned
above, the fitting of experimental data obtained upon
competition reactions with substituted styrenes with the
Hammett equation is a common probe in this area. Negative
values of ρ are indicative of the existence of a transition state
with a certain positive charge. As a result of this, the styrenes
bearing electron withdrawing groups in the aromatic ring are
less reactive than those with electron-donating substituents.
The reported data17b,18,19 show small values for ρ within the
range of −0.38 to −0.65.25 We have already reported the
comparison of these competition studies for the styrene
cyclopropanation and aziridination reactions with the complex
TpMe2Cu(C2H4).24 We found that the cyclopropanation
b
reaction data could be nicely fitted to the Hammett equation
1340
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
Table 1. Ratio of Aziridines (X/H) Obtained from Competition Experiments with p-Substituted Styrenes
entry
TpxM
OMe
Me
Ph
F
H
Cl
NO2
1
2
3
4
5
6
TpMe2Cu
Tp*,BrCu
TpMsCu
TpBr3Cu
TpMe2Ag
Tp*,BrAg
2.10
2.31
1.78
2.18
3.20
3.06
1.44
1.46
1.57
1.38
1.48
1.57
1.38
1.41
1.33
1.25
1.73
1.42
1.03
0.89
1.08
1.00
0.84
0.75
1.00
1.00
1.00
1.00
1.00
1.00
1.11
1.06
1.06
0.87
0.90
1.00
1.01
0.90
0.96
0.82
0.75
0.60
Table 2. Values of ρ+, ρ• and ρ+/ρ• Obtained Using the Hammett’s Dual-Parameter Equation in the Aziridination Reaction of p-
Substituted Styrenes
entry
catalyst
ρ+
ρ•
R2
ρ+/ρ•
1
2
3
4
5
6
7
TpMe2Cu
Tp*,BrCu
TpMsCu
TpBr3Cu
TpMe2Ag
Tp*,BrAg
CuI
−0.250 0.024
−0.314 0.043
−0.224 0.021
−0.318 0.030
−0.475 0.065
−0.499 0.074
−0.283 0.014
0.272 0.045
0.309 0.082
0.212 0.040
0.248 0.057
0.416 0.122
0.307 0.140
0.172 0.019
0.9666
0.9310
0.9659
0.9653
0.9308
0.9179
0.9978
0.925
1.016
1.066
1.282
1.141
1.625
1.645
(eq 1). However, the attempts to fit the data from the related
aziridination reaction failed, since the electron-withdrawing
groups, −Cl or −NO2 for instance, led to aziridines in such
amounts that the correlation expected from eq 1 was lost.
Actually, this behavior was previously observed by Evans and
co-workers,7a when they described nearly 1:1 mixtures of the
aziridines derived from styrene and p-nitrostyrene in dichloro-
methane using plain CuClO4 as catalyst precursor, although a
full set of competition experiments was not given. We found
that the experimental data could be fitted to a dual-parameter
equation containing both polar and radical contributions (eq
2). After our report, Che and co-workers applied3 the same idea
to the experimental results obtained in stoichiometric nitrene
transfer reactions from bis(imido)ruthenium complexes to
olefins.
electron donating groups favored the aziridination reaction in
all cases. However, a distinct behavior was observed in the case
of electron-withdrawing groups, particularly with p-NO2.
Different degrees of deviation from the expected linear
correlation (according to Hammett’s equation) were found
for the series of catalysts employed (Table 1). Once the
generality of this behavior was assessed for a group of different
ligands (with distinct electronic and steric properties) and both
metals, we applied the dual parameter equation (eq 2) using the
three scales for σ• proposed by Jiang and Ji,26 Jackson,27 and
Fischer.28 The best fitting was so far obtained with Jackson’s
scale (see Supporting Information for all fitting data and plots),
from which the series of values of ρ• and ρ+ shown in Table 2
were obtained. Figure 1 displays the plot of the experimental
log(KX/KH) versus calculated ρ+σ+ + ρ•σ• for Tp*,BrM (M =
Cu, Ag). These correlations evidence that both polar and
radical effects must be taken into account to explain this
transformation. The negative value of ρ+ indicates an
KX
KH
log
log
= ρσ
(1)
KX
KH
= ρ+σ+ + ρ•σ•
(2)
We have now carried out an extensive study (Tables 1 and 2,
entries 2−6 and 2−7, respectively) with a series of copper and
silver complexes to evaluate the validity of our previous study
(referred to single example), and to check the effect of the Tpx
ligand and the metal in this transformation. In a general
experiment, 1:1 mixtures of styrene and a p-substituted styrene
were reacted with PhINTs, in the presence of catalytic
amounts of TpxM (M = Cu, Ag) as the catalyst, with a 1:
20:100:100 ratio of catalyst, PhINTs, and both styrenes.
After 12 h of stirring at room temperature, the solvent was
removed and the residue was investigated by 1H to provide the
relative ratio of aziridines (overall yields in aziridines were
within the interval 60−80%, and the remaining initial PhINTs
was converted into TsNH2). Table 1 shows the relative ratios of
products, from which it can be concluded that styrenes bearing
Figure 1. Plot of experimental log(KX/KH) vs calculated ρ+σ+ + ρ•σ•
(from Table 2) for Tp*,BrM (M = Cu, Ag).
1341
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
Table 3. Values of ρ+, ρ•, and ρ+/ρ• Obtained Using the Hammett’s Dual-Parameter Equation in the Cyclopropanation Reaction
of p-Substituted Styrenes
entry
catalyst
ρ+
ρ•
R2
ρ+/ρ•
1
2
3
TpBr3Cu
Tp*,BrCu
TpMsCu
−0.406 0.076
−0.623 0.04
−0.333 0.069
−0.048 0.144
0.022 0.075
0.007 0.131
0.9189
0.9891
0.8984
8.45
28.3
47.5
electrophilic attack of the olefin and the development of a
positive charge, but very early in the reaction pathway, as
inferred from the small value (−0.25 to −0.49). The same
reason can be applied for the radical ρ• contribution: radical
species must be located at a certain early stage of the reaction
coordinate.26−28 The positive value indicates that all substituents
contribute to spin delocalization. It is worth mentioning that
the observation of both contributions was also observed with
CuI as the catalyst precursor (Table 2, entry 7), as a model for
the array of simple copper salts described by Evans and co-
workers.7a
shows both correlations for the series of four copper catalysts,
from which it is concluded that the polar contribution is
enhanced by electron-poor metal centers, that are more difficult
to oxidize; that is, the radical contribution decreases when the
redox potential increases.
Effect of the Presence of BHT as the Radical Inhibitor. A
set of experiments has been performed for twin styrene
aziridination reactions catalyzed by Tp*,BrM (M = Cu, Ag) as
the catalysts that only differ in the presence or absence of t-
butylhydroxitoluene (BHT) as the radical inhibitor. Scheme 4
The relative parameter ρ+/ρ• also provides some interesting
information. First, this ratio is close to unity (0.96−1.60) in all
cases. This is at variance with data obtained for a well-known
transformation that occurs in a concerted manner, the copper
−catalyzed cyclopropanation reaction. From similar competi-
tion studies with p-substituted styrenes and ethyl diazoacetate
as the carbene source, we have also fitted the collected data
with the dual parameter eq 2 (Table 3). The ρ+/ρ• values
appear within the range 8.45−47.5, largely favoring the polar
contribution, as expected. Actually, fitting to the classical
Hammett eq 1 gave nearly the same ρ+ values (see Supporting
Information). The appearance of a small radical contribution in
the carbene transfer reaction to olefins could be associated with
the existence of an asynchronous step in the cyclopropanation
reaction, as recently proposed by Echavarren and co-workers.29
We have also found that the relative ratio ρ+/ρ• also
correlates with two magnitudes: (a) the electron density
located at the metal center (evaluated with the aid of FTIR data
obtained from a series of TpxCu(CO) complexes) and (b) the
Scheme 4. Effect of the Presence of a Radical Inhibitor
shows the results obtained: the copper-based catalyst showed a
drastic decrease (by 50%) of the aziridine yield. However,
nearly identical yields were obtained in the case of the silver-
based system (within experimental error) with or without BHT.
This observation has important implications for the mechanistic
proposal. Further, the Ag-based systems may not form the
radical intermediate or if achieved, it would be a short-lived
species that could not be intercepted by the radical inhibitor. In
contrast, radical intermediate formation and interception to a
certain degree seems to be possible in the case of the Cu-based
system.
Cu(I)/Cu(II) oxidation potential, measured as E0
.
30 Figure 2
1/2
Aziridination of Dicyclopropylethylene as Radical Clock.
The use of a radical clock to gain mechanistic insights in the
copper-catalyzed olefin aziridination reaction was described by
Evans and co-workers.7a With CuClO4 as catalyst, the reaction
in Scheme 2h only provided the aziridines, with no evidence of
any product derived from the radical-induced aperture of
cyclopropyl substituent. We have now employed 1,1-
dicyclopropylethylene as the substrate for the aziridination
reaction in the presence of TpxM (M = Cu, Ag) catalysts, an
olefin that has been described as a radical clock to assess the
implication of carbon radical intermediates.31 Interestingly, we
have observed with TpBr3Cu and Tp*,BrAg (the most active
catalysts) the preferential formation of the imine shown in
Scheme 5. With the silver catalyst no other product was
Figure 2. Plot of ρ+/ρ• vs ν(CO) and E0 of a series of TpxCu
complexes.
1/2
1342
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
Scheme 5. Aziridination of the Radical Clock Probe 1,1-
Dicyclopropylethylene
Scheme 6. Mechanistic Picture of the Olefin Aziridination
a
Reaction Catalyzed by TpxMNTs (M = Cu, Ag)
detected, whereas with the copper-based catalyst minor
amounts of other species showing olefinic protons were
detected by NMR. Although the mechanistic interpretation of
these results will be provided later in this contribution, at this
stage we can comment on them in a parallel manner with those
in the above paragraph on the use of BHT: a certain degree of
involvement of radical species is inferred from both sets of
experiments for the copper system (BHT and radical clock),
whereas the silver-based catalyst does not seem to involve those
species on the basis of our experiments.
a
The two Kohn−Sham orbitals, σ* are 1π*, are the singly occupied
Summary of Experimental Results. Table 4 shows in brief
the results collected and discussed in the previous sections.
molecular orbitals (SOMO) of the triplet state of TpMe2MN−Ts.
intermediate, TpMe2MN−Ts (M = Ag, Cu), holds two unpaired
electrons on the 1π* and σ* orbitals (SOMO, Scheme 7). We
Table 4. Summary of Experimental Results
stereochemistry
studies
radical
radical
b
catalysts
TpxCu
Hammett studies
inhibitor
clock
Scheme 7. Free Energy Profiles for the Formation of the
Active Metal-Nitrene Intermediate, TpMe2MN−Ts (M = Ag,
a
retention
polar and radical
contribution
affected
affected
a
Cu)
TpxAg
retention
polar and radical
contribution
unaffected
unaffected
a
b
With the exception of E,E-hexadien-1-ol. No aziridines formed; see
Scheme 5 for products.
With both families of catalysts (Cu- or Ag-based), the results
are somewhat contradictory in terms of unambiguously
pointing toward one pathway, either concerted or stepwise.
Thus, the concerted pathway, such as that similar to the
cyclopropanation reaction in which both C−N bonds are
formed synchronously, would not be in accord with the
observance of a radical contribution in the Hammett studies.
But the observance of the retention of stereochemistry, either
with aryl- or alkyl-substituted olefins with both metals could be
interpreted (as it has been in the past when using this probe
only) in the opposite sense, as the prevalence of a concerted
pathway. The formation of the imine from the radical clock
(Scheme 5) has also important implications, since it could be
the result (as explained below) of a stepwise mechanism in
which the second C−N bond cannot form. Therefore, a first
sight to the overall picture points toward the lack of any general
explanation for all these results. It is worth mentioning at this
stage that most previously proposed pathways have been
supported on the basis of just one of the above experiments,
and that the biunivocal correlations concerted-singlet nitrene
and stepwise-triplet nitrene have also been invoked to propose
corresponding pathways.
Computational Studies. With the aim of providing a
general mechanistic explanation accounting for all the above
experimental data, DFT studies have been carried out. Current
experimental and previous13,20 theoretical studies have
anticipated two mechanistic proposals for transition metal-
catalyzed aziridination reactions, namely, the concerted and
stepwise (or radical) mechanisms (Scheme 6). We have the
TpMe2M (M = Cu, Ag) as the model systems in the present
study. The triplet state of the active metal−nitrene
a
Free energies (kcal mol−1) and spin densities (in red) of the Ag-based
system are in full text and Cu-based systems are in italic.
follow the orbital labeling of Phillips and co-workers20b in their
analogous Ru-based system. In the simplified mechanistic
picture, the triplet state would follow the stepwise mechanism:
the first step consists of the transfer of one electron from an
olefin-π orbital to the metal−nitrene 1π* orbital, giving rise to a
triplet biradical intermediate (RI, Scheme 6). The presence of
this intermediate is consistent with the observation of inversion
of the stereochemistry of the olefin, as a rotation around the
carbon−carbon bond may take place before the second
electron transfer from the olefin−π orbital to the metal−
nitrene σ* orbital. The alternative pathway from the (closed-
shell) singlet electronic state of TpMe2MN−Ts would perform
the two-electron transfer from the olefin π orbital to the σ*
orbital of TpMe2MN−Ts, and this concerted process leads to
complete retention of the stereochemistry of the olefin. An
additional electronic state that could be present is the open-
shell singlet, with an antiparallel occupation of the same open-
1343
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
shell orbitals of the triplet. This open-shell singlet should
behave similarly to the triplet in terms of stereochemistry.
However, experimental results described in the previous
section cannot be directly fitted within this mechanistic picture.
A concerted mechanism, through a singlet spin state, can be
proposed for the retention of configuration in the reaction of Z-
2-pentene, E-β-methyl-styrene, and E-2-hexene with both
TpxAg and TpxCu systems, as well as in the reaction of E,E-
hexadien-1-ol with TpxAg. On the other hand, observed loss of
stereochemistry in the aziridination of E,E-hexadien-1-ol with
the TpxCu system supported for the radical mechanism. Finally,
correlations of the experiments carried out with p-substituted
styrene indicated a contribution of both radical and concerted
mechanisms with both metals. Herein we present a detailed
computational study to rationalize these puzzling observations
of aziridination reactions catalyzed by TpxAg and TpxCu
complexes.
We must admit that the relative energies of states with
different multiplicities depend heavily on the functional used.
We have included in the Supporting Information (see section
S6) a comparison of the performance of the M06 and BP86
functionals for this structure, as well as a justification that our
preferred M06 is more appropriate for this system.
Aziridination of E-2-Hexene. Computed free energy profiles
for aziridination of E-2-hexene catalyzed by TpMe2AgN−Ts are
shown in Scheme 8. The first part of the reaction takes place
Scheme 8. Computed Free Energy Profiles for the
a
Aziridination of E-2-Hexene Catalyzed by TpMe2AgN−Ts
Formation of the Metal−Nitrene Intermediate. We have
first focused on calculated free energy profiles for the active
metal−nitrene intermediate (TpMe2MN−Ts) formation
(Scheme 7). The catalytic cycle starts from TpMe2M (2),
which is initially mixed with C2H4 (as the model olefin in this
step), leading to the reversible formation of TpMe2M(C2H4)
complex (1). Adducts of composition TpMe2M(L) are known in
the literature as catalyst precursors, and dissociation pre-
equilibrium of the ligand (L) is well-established.32 Dissociation
of ethene from TpMe2M(C2H4) needs 3.3 and 10.1 kcal mol−1
for Ag- and Cu-based systems, respectively. The next step
consists of PhINTs coordination on TpMe2M, giving rise to a
TpMe2M−N(PhI)−Ts complex (3) with the calculated nitro-
gen−iodine bond lengths in TpMe2Ag−PhIN−Ts of 3.16 Å and
in TpMe2Cu−N(PhI)−Ts of 2.99 Å. In the presence of weak
nitrogen−iodine bond, PhI dissociation would be easier (we
were unable to locate transition states for this step), leading to
the singlet metal−nitrene (1R). However, the existence of a
minimum energy crossing point (MECP) between the singlet
and the triplet profiles allows the formation of the triplet
a
Free energies are in kcal mol−1 and spin densities are in red.
through the triplet energy surface, where the first nitrogen−
3
carbon bond formation occurs via TS with a barrier of 12.8
kcal mol−1 above the separate reactants. The closed-shell singlet
transition state, TS, has a relative energy of 15.1 kcal mol−1,
1
and therefore is a minor contributor for the overall rate. We
searched for the open-shell singlet electronic transition state,
but we were unable to locate it. On the basis of our DFT
calculations, aziridination of E-2-hexene occurs through the
radical mechanism, where the favored transition state for the
3
metal−nitrene, R.
The optimized triplet state of TpMe2AgN−Ts intermediate
holds spin densities on nitrogen of 1.32 and silver of 0.30,
showing the two unpaired electrons are mainly localized on the
nitrogen atom (see section S5, Supporting Information).
Despite several attempts, we were unable to locate the open-
shell singlet state of this system. The closed-shell singlet state
(1R) of TpMe2AgN−Ts is 7.6 kcal mol−1 higher than the triplet
state. In the analogous TpMe2CuNTs system, triplet state (3R)
is 10.6 kcal mol−1 more stable than the closed-shell singlet state
(1R). Moreover, this triplet state holds the calculated spin
densities on nitrogen of 1.09 and copper of 0.56, which denotes
significant spin density delocalization over the metal center.
This electronic nature helps us to locate an open-shell singlet
3
reaction is TS that leads to the triplet radical intermediate
(3RI), which is −8.2 kcal mol−1 below the reactants. However,
this intermediate may not be observed, because there is a spin
crossing between the triplet and closed-shell singlet energy
profiles (MECP at −8.1 kcal mol−1). After spin crossover to the
singlet surface, the product (1P) is reached with a high stability
of −49.9 kcal mol−1. Therefore, TpMe2Ag system performs
aziridination of E-2-hexene with retention of stereochemistry,
3
because radical intermediate RI, where the scrambling could
occur, is not formed.
Scheme 9 shows the computed free energy profile for the
analogous copper-based system. In a very similar way to the
silver-analogue, the reaction starts from the triplet metal−
nitrene intermediate, and the formation of the radical
1
state, RUKS, which is 5.6 kcal mol−1 above the triplet state. In
this open-shell singlet state, the calculated spin density on the
metal−nitrene unit, ρ(Cu) = −0.65 and ρ(N) = +0.79,
indicated that the spin vectors are antiparallel, and the
calculated spin contamination, ⟨S2⟩, is 1.00. The two unpaired
electrons are located in similar orbitals in the triplet (parallel
spins) and open-shell singlet (antiparallel spin) states. On the
basis of our calculations, the electronic structure of the copper−
nitrene intermediate is rather TpMe2CuIIN•Ts, while the silver−
nitrene intermediate holds the two unpaired electrons mainly
localized on nitrogen.
intermediate, RI, has a barrier of 11.6 kcal mol−1 (3TS). In
3
this case, we were able to locate the open-shell singlet transition
state (1TSUKS), which is 4.4 kcal mol−1 above the 3TS, while the
closed-shell singlet transition state (1TS) is a further 4.0 kcal
mol−1 higher in energy. The open-shell singlet replaces thus the
closed-shell singlet as an alternative to the triplet in the copper
system. This open-shell singlet plays however no role before
intermediate 3RI, as it remains always higher in energy than the
1344
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
transition state 3TSRot, which has a relative energy of −8.4 kcal
Scheme 9. Computed Free Energy Profiles for Aziridination
a
1
3
mol−1. Transition state TSU′ KS is 1.3 kcal mol−1 below TSRot
,
of E-2-Hexene Catalyzed by TpMe2CuN−Ts
3
and thus the triplet intermediate, RI, would mostly evolve to
the products before it has time to scramble. Therefore, both
silver- and copper-based systems lead to retention of stereochemistry,
but there are two different pathways to account for such behavior.
Aziridination of E,E-2,4-Hexadien-1-ol. Computed free
energy profiles for aziridination of E,E-2,4-hexadien-1-ol are
qualitatively similar to those of E-2-hexene as can be seen in
Figure S3 (section S8, Supporting Information). The analysis of
the profile for the reaction with the silver-based catalyst
(Supporting Information, Figure S3a) produces the same
conclusions for E-2-hexene described above; specifically the
reaction starts on the triplet state, but the scrambling cannot
take place because the system moves to the closed-shell singlet
3
state before reaching the radical intermediate, RI.
In the case of the copper-based complex (Figure S3b, section
S8, Supporting Information), a significant qualitative difference
appeared which we highlight in Scheme 10b. The ordering
between the relative energies of the two competing paths from
3
intermediate RI is opposite, and the transition state (1TSU′ KS
)
leading to the ring closure has a relative energy of −15.8, which
is 5.5 kcal mol−1 higher than the transition state leading to
scrambling (3TSRot). As a result, rotation can take place around
the C−C bond before the ring closure takes place, and both cis-
and trans-aziridines can be formed. The proportion between the
two aziridines will depend on the relative energies of the
corresponding ring-closure transition states, 1TSU′ KS (leading to
a
Free energies are in kcal mol−1 and spin densities are in red.
triplet. There is a significant qualitative difference between the
energy profiles of the silver- and copper-based systems. Despite
several attempts, we were unable to locate a MECP between
the triplet and closed-shell singlet (or open-shell singlet) energy
profiles before the triplet energy profile reaches to the radical
intermediate (3RI), which is −13.4 kcal mol−1 below the entry
channel. The open-shell singlet from of the radical intermediate
1
trans-aziridine) and TS1′UKS (leading to cis-aziridine). Accord-
1
ing to our calculations, TSU′ KS is 1.9 kcal mol−1 more stable
1
than TS1U′ KS, and therefore trans-aziridine can be formed as
the major product.
(1RIUKS) is only 0.2 kcal mol−1 above the RI. We have also
3
Therefore, the lack of scrambling cannot be used as
unambiguous evidence for the nonradical mechanism, because
even the reaction with retention of configuration starts from the
triplet state. For scrambling to take place, two conditions are
necessary. The first of them is the formation of triplet radical
intermediate (3RI), where only one N−C bond has to be
formed. This is not achieved for the silver-based systems that
we have tested, due to the presence of a direct crossing to the
singlet surface, leading to the products after the first transition
state (3TS). Therefore, the electronic properties of the silver−
nitrene intermediate prevent scrambling. The second condition
for scrambling is that once the 3RI intermediate is reached, the
rotation around the C−C bond leading to inversion has to be
faster than the product formation. This intermediate is reached
for the copper-based catalysts, but the rotation is only fast
enough in the case of E,E-hexadien-1-ol. This could be
explained as the result of a conjugation between the double
bonds in this species stabilizing the open ring radical
intermediate, and increasing the barrier for the ring closure.
Mechanistic Proposal. At this point, unification of
experimental data and theoretical calculations is required. As
shown in Scheme 11, the reaction starts in all cases with the
formation of a metal−nitrene intermediate in the triplet state.
Further, copper−nitrene intermediate is rather TpxCuIIN•Ts,
whereas the silver−nitrene displays a biradical character on the
nitrogen atom. Their formation takes place by means of a
singlet-to-triplet spin crossover characterized by a MECP. This
metallonitrene species can attack the olefin through a transition
state (3TS) leading to the first carbon−nitrogen bond
formation. From here, two different routes seem to operate.
In the silver case, the triplet route intercepts with that of the
searched the closed-shell singlet form of this species, which is
however not a minimum that ultimately converged to the
product, 1P (see section S7, Supporting Information for
details).
3
Once this radical intermediate RI is formed on the triplet
profile, inversion of the stereochemistry of olefin may take
place, but we need to consider a new competition between ring
closure (toward the product) and bond rotation (toward the
scrambling). The reaction toward the product (Scheme 9 and
Scheme 10a) proceeds through a MECP between triplet and
open-shell singlet profiles (11.8 kcal mol−1 below reactants)
and transition state 1TS′UKS (relative energy, −9.8 kcal mol−1),33
which is the highest energy point before reaching the final
1
product, P (−31.0 kcal mol−1). The path toward scrambling
involves rotation around the carbon−carbon bond through the
Scheme 10. Computed Free Energy Profiles (kcal mol−1) for
the Second Electron Transfer Process of the Radical
Mechanism with (a) E-2-Hexene and (b) E,E-Hexadien-1-ol
Catalyzed by TpMe2CuN−Ts
1345
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
Scheme 11. Mechanistic Proposal Based on Experimental and Theoretical Data from This Work
singlet before a true intermediate (3RI) is formed, leading to
Scheme 12. Plausible Explanation for the Cu- or Ag-
Catalyzed Nitrene Transfer to 1,1-Dicylopropylethylene
the aziridines product. Because of this, retention of stereo-
chemistry is observed with all the olefins, in a process that if not
purely synchronous, can be considered at least as concerted.
Since the radical intermediate is out of the most probable
reaction pathway, there is no effect of the presence of radical
traps. However, the lack of fitting to the plain Hammett
equation and the proposal of a radical character must be related
to the triplet nature of the transition state 3TS previous to ring
closing. This affects both copper and silver, in an early stage of
the reaction as already discussed.
3
For the copper-based system the radical intermediate RI is
reached, explaining the decrease of the reaction outcome with a
radical inhibitor. Now, the crossing of the triplet and singlet
pathways takes place beyond that intermediate, and the relative
value of the corresponding MECP for such crossing and the
barrier for rotation around the C−N bond will control the
stereochemistry.
A piece of information is yet to be explained, namely the lack
of formation of aziridines when 1,1-dicyclopropylethylene was
employed as the substrate. Scheme 12 shows a plausible
explanation for this issue. Assuming that a similar transition
state is formed (3TS), it is very likely that the steric pressure
exerted by the two cyclopropyl groups disfavors the aziridine
ring closure and leads to the imine as an alternative product.
Either with (for copper) or without (for silver) formation of the
radical intermediate, the imine could be formed in one step by a
1,2-hydrogen shift. The observance of small amounts of
cyclopropyl-opening product with the copper catalyst would
these MECPs compared with those of radical intermediates and
transition states involved in ring closing or C−C bond rotation
account for the experimental observations in all cases, allowing
us to propose a general mechanism. As the result of it, the
generally accepted proposal that the retention of the stereo-
chemistry of the olefin is indicative of a concerted pathway
based on a singlet nitrene seems to be no longer valid. Herein
we have demonstrated that a stepwise pathway, involving triplet
copper−nitrene species can also lead to stereospecific
aziridination. Also we have shown that even in the case of a
concerted (albeit asynchronous) step such as in the case of
silver, the initial metallonitrene is a biradical species, not the
usually assumed singlet. We believe our findings can be
extended to the other reported catalytic systems for the
aziridination reaction that could be worth revisiting from a
mechanistic point of view.
3
proceed from the RI intermediate as proposed by Evans and
co-workers,7a this intermediate being only formed in the case of
copper.
CONCLUSIONS
■
On the basis of our experimental data and theoretical
calculations, we have developed a mechanistic proposal for
the olefin aziridination reaction catalyzed by TpxM (M = Cu,
Ag; Tpx = hydrotrispyrazolylborate ligand). With both metals,
the reaction is triggered by the formation of a metal−nitrene
intermediate in the triplet state, holding radical character. The
reaction profile is characterized by the presence of minimum
energy crossing points (MECPs), where a change from the
triplet to the singlet surface takes place. The relative energy of
EXPERIMENTAL SECTION
■
General Methods. All preparations and manipulations were
carried out under a oxygen-free nitrogen atmosphere using conven-
tional Schlenk techniques or inside a drybox. All the olefins were
1346
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
purchased from Aldrich. Olefins and solvents were rigorously dried
before use. The copper and silver complexes employed as catalysts
were prepared according to the literature.23,24,30,34 PhINTs was also
prepared following the reported methods.35 NMR experiments were
run in a Varian Mercury 400 MHz spectrometer. GC data were
collected with a Varian GC-3900.
that used by Norrby and co-workers.20a Real energy of the open-shell
singlet electronic states (∈s) were evaluated by considering the energy
∈0 of the optimized broken-symmetry solution and energy ∈1 from
separate spin-unrestricted ms = 1 calculation at the same geometry,
considering the following formula:41
S12 ∈ 0 −S02∈ s
∈ s≈
General Catalytic Aziridination Reaction. The copper or silver
complex (0.01 mmol) and the olefin (1 mmol) were dissolved in
CH2Cl2 (10 mL). After 5 min of stirring, PhINTs was introduced in
one portion (0.2 mmol), and the mixture was stirred overnight at
room temperature. No solid PhINTs was observed in the reaction
mixture after that time. Volatiles were removed under vacuum, and the
S12 − S02
where S0 and S1 are spin contaminations of the open-shell singlet and
the triplet states, respectively.
1
aziridines were identified by H NMR spectroscopy of the reaction
ASSOCIATED CONTENT
■
crude in CDCl3 and comparison with the reported data (see
Supporting Information).
S
* Supporting Information
NMR spectra of the experiments cited in the text, computed
structures of metallonitrene complexes, comparison of M06 and
BP86 functionals, calculated spin density distribution of open-
shell states, total energies and Cartesian coordinates of the
computed structures. This material is available free of charge via
the Internet at http://pubs.acs.org.
Catalytic Competition Experiments. These experiments were
done by following the above procedure, but using mixtures of styrene
and the corresponding p-substituted styrene (0.5 mmol of each). Solid
PhINTs was not observed in the reaction mixture after stirring
overnight. Volatiles were removed under vacuum and the relative ratio
of aziridines was estimated by H NMR spectroscopy of the reaction
crude in CDCl3 (see Supporting Information).
1
Effect of Radical Inhibitor. This experiment was set up in an
identical manner to that described above, with the addition of 5 equiv
(with respect to the catalyst) of t-butylhydroxytoluene (BHT) to the
reaction mixture, and with the use of Tp*,BrM (M = Cu, Ag) as the
catalyst. Yields shown in Scheme 4 were calculated by 1H NMR
spectroscopy in CDCl3 (see Supporting Information).
AUTHOR INFORMATION
■
Corresponding Author
mmdiaz@dqcm.uhu.es; fmaseras@iciq.es; perez@dqcm.uhu.es
Notes
1,1-Dicyclopropylethylene as Radical-Clock Probe. 1,1-
Dicyclopropylethylene was prepared by means of a typical Wittig
reaction, from dicyclopropylketone and Ph3PCH3Br. Once isolated,
0.5 mmol of the olefin was reacted with 0.5 mmol of PhINTs in 10 mL
of dried CH2Cl2 in the presence of 0.025 mmol of either TpBr3Cu or
Tp*,BrAg as the catalyst. After the mixture was stirred overnight, the
volatiles were removed under vacuum and the residue was investigated
by NMR. In both cases the major product was identified as N-(2,2-
dicyclopropylethylidene)-4-methylbenzenesulfonamide. 1H NMR
(400 MHz, CD2Cl2): δ 8.57 (d, 1H, J = 6.2 Hz), 7.68 (d, 2H, J =
8.2 Hz), 7.37 (d, 2H, J = 8.2 Hz), 2.41 (s, 3H), 1.24 (dt, 1H, J = 6.2,
8.5 Hz), 0.88 (m, 2H), 0.58 (m, 2H), 0.45 (m, 2H), 0.28 (m, 2H),
0.16 (m, 2H). 13C{1H} NMR (400 MHz, CD2Cl2): 179.8 (iminic
CH), 129.8, 127.8 (aromatic CH), 53.5 (CH3), 21.4 (CH), 11.6 (CH),
3.3 (CH2), 2.0 (CH2).
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank MINECO (CTQ2011-28942-CO2-01, CTQ2011-
27033 and Consolider Ingenio 2010 CSD2006-0003), Junta de
́
Andalucıa (Proyecto P10-FQM-06292), and the ICIQ
Foundation. L.M. thanks Plan Propio de Investigacio
(Universidad de Huelva) for a research fellowship
́
n
REFERENCES
■
(1) (a) Chang, J. W. W.; Ton, T. M. U.; Chan, P. W. H. Chem. Rec.
2011, 11, 331−357. (b) Halfen, J. A. Curr. Org. Chem. 2005, 9, 657−
669. (c) Muller, P.; Fruit, C. Chem. Rev. 2003, 103, 2905−2919.
̈
(2) (a) Mansuy, D.; Mahy, J.-P.; Dureault, A.; Bedi, G.; Battioni, P. J.
Chem. Soc., Chem. Commun. 1984, 1161−1163. (b) Mahy, J.-P.; Bedi,
G.; Battioni, P.; Mansuy, D. J. Chem. Soc., Perkin Trans. 2 1988, 1517−
1524.
Computational Methods. Gas phase structure optimizations
were carried out using DFT as implemented in the Gaussian 09
package36 with the M06 functional.37 The BP86 functional was
considered as a nonhybrid alternative, but the results collected (in the
Supporting Information) indicate it performs worse for this particular
problem. The SDD basis set was used for silver, copper, and iodine.38
The 6-31G* basis sets were considered for the remaining atoms.39
Final energies were obtained by doing single-point calculations on the
optimized structures, where silver, copper, and iodine were still
described with SDD, and the remaining atoms were treated with the 6-
311++G** basis sets. All geometry optimizations were full with no
restrictions, and vibrational frequency calculations performed in order
to establish the stationary points were minima or transition states.
Further, the connectivity of the transition states was confirmed by
relaxing the transition state geometry toward both the reactant and the
product. Free-energy corrections at 298.15 K and 105 Pa pressure were
used, including zero-point energy corrections. Minimum energy
crossing points (MECP) were calculated with the MECP program
of J. H. Harvey and co-workers.22 Open-shell electronic configurations
were obtained by using the broken symmetry methodology,40 and
calculated total spin density distributions of selected open-shell
stationary points are summarized in the Supporting Information (see
section S9). This method struggles in the description of the transition
from open-shell to closed-shell singlets, as happens in the ring closure
(3) Sze-Man, A.; Huang, J.-S.; Yu, W.-Y.; Fung, W.-H.; Che, C.-M. J.
Am. Chem. Soc. 1999, 121, 9120−9132.
(4) Ruppel, J. V.; Jones, J. E.; Huff, C. A.; Kamble, R. M.; Chen, Y.;
Zhang, X. P. Org. Lett. 2008, 10, 1995−1998.
(5) Muller, P.; Baud, C.; Jacquier, Y. Tetrahedron 1996, 52, 1543−
̈
1548.
(6) Antunes, A. M. M.; Marto, S. J. L.; Branco, P. S.; Prabbakar, S.;
Lobo, A. M. Chem. Commun. 2001, 405−406.
(7) (a) Evans, D. A.; Faul, M. M.; Bilodeau, M. T. J. Am. Chem. Soc.
1994, 116, 2742−2753. (b) Li, Z.; Conser, K. R.; Jacobsen, E. N. J.
Am. Chem. Soc. 1993, 115, 5326−5317. (c) Evans, D. A.; Faul, M. M.;
Bilodeau, M. T.; Anderson, B. A.; Barnes, D. M. J. Am. Chem. Soc.
1993, 115, 5328−5329. (d) Evans, D. A.; Faul, M. M.; Bilodeau, M. T.
J. Org. Chem. 1991, 56, 6744−6746.
(8) Cui, Y.; He, C. J. Am. Chem. Soc. 2003, 125, 16202−16203.
(9) Li, Z.; Ding, X.; He, C. J. Org. Chem. 2006, 71, 5876−5880.
(10) (a) Baidei, Y. M.; Dinescu, A.; Dai, X.; Palomino, R. M.;
Heinemann, F. W.; Cundari, T. R.; Warren, T. H. Angew. Chem. Int.
Ed. 2008, 47, 9961−9964. (b) Kundu, S.; Miceli, E.; Pfaff, F. F.;
Kulhlmann, U.; Hildebrant, P.; Braun, B.; Greco, C.; Ray, K. J. Am.
Chem. Soc. 2012, 134, 14710−14713.
1
1
transition states TSU′ KS and TS1U′ KS Because of that, the structure
and energy of these transition states had to be estimated through a
relaxed scan of the N−N−C bond angle, in a procedure analogous to
(11) Lyaskovskyy, V.; Suarez, A. I. O.; Lu, H.; Jiang, H.; Zhang, X. P.;
de Bruin, B. J. Am. Chem. Soc. 2011, 133, 12264−12273.
1347
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348

Journal of the American Chemical Society
Article
(12) Ling, E. R.; Hennesy, E. T.; Betley, T. A. J. Am. Chem. Soc. 2011,
133, 4917−4923.
(36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
(13) (a) Comba, P.; Haaf, C.; Lienke, A.; Muruganantham, A.;
Wadepohl, H. Chem.Eur. J. 2009, 15, 10880−10887. (b) Comba, P.;
Lang, C.; Lopez de Laorden, C.; Muruganantham, A.; Rajaraman, G.;
Wadepohl, H.; Zajaczkowski, M. Chem.Eur. J. 2008, 14, 5313−5328.
(14) Li, Y.; He, J.; Khankhoje, V.; Herdtweck, E.; Kholer, K.;
̈
Storcheva, O.; Cokoja, M.; Kuhn, F. E. Dalton Trans. 2011, 40, 5746−
̈
5754.
(15) Xu, Q.; Apella, D. J. Org. Lett. 2008, 10, 1497−1500.
(16) Ma, L.; Du, D. M.; Xu, J. Chirality 2006, 18, 575−580.
(17) (a) Muller, P.; Baud, C.; Jacquier, Y.; Moran, M.; Nageli, I. J.
̈
̈
Phys. Org. Chem. 1996, 8, 341−347. (b) Muller, P.; Baud, C.; Jacquier,
̈
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Y. Can. J. Chem. 1998, 76, 738−750.
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.
(37) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101−
194118.
(18) Gillespie, K. M.; Crust, E. J.; Deeth, R. J.; Scott, P. Chem.
Commun. 2001, 785−786.
(19) Han, H.; Park, S. B.; Kim, S. K.; Chang, S. J. Org. Chem. 2008,
73, 2862−2870.
(38) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
̈
(20) (a) Brandt, P.; Sodergren, M. J.; Andersson, P. G.; Norrby, P.-O.
̈
Theor. Chim. Acta 1990, 77, 123−141.
J. Am. Chem. Soc. 2000, 122, 8013−8020. (b) Lin, X.; Che, C.-M.;
Phillips, D. L. J. Org. Chem. 2008, 73, 529−537. (c) Suarez, A. I. O.;
Jiang, H.; Zhang, X. P.; de Bruin, B. Dalton Trans. 2011, 40, 5697−
5705. (d) Cundari, T. R.; Dinescu, A.; Kazi, A. B. Inorg. Chem. 2008,
47, 10067−10072. (e) Hopmann, K. H.; Ghosh, A. ACS Catal. 2011,
1, 597−600. (f) Kundu, S.; Miceli, E.; Farquhar, E.; Pfaff, F. F.;
Kuhlmann, U.; Hildebrandt, P.; Braun, B.; Braun, B.; Greco, C.; Ray,
C. J. Am. Chem. Soc. 2012, 134, 14710−14713.
(39) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,
56, 2257−2261. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta
1973, 28, 213−222. (c) Gordon, M. S. Chem. Phys. 1980, 76, 163−
168.
(40) (a) Noodleman, L.; Lovell, T.; Han, W.-G.; Li., J.; Himo, F.
Chem. Rev. 2004, 104, 459−508. (b) Noodleman, L. J. Chem. Phys.
1981, 74, 5737−5743. (c) Noodleman, L.; Case, D. A. Adv. Inorg.
Chem. 1992, 38, 423−470.
(41) Chan, K. S.; Li, X. Z.; Dzik, W. I.; de Bruin, B. J. Am. Chem. Soc.
2008, 130, 2051−2061.
(21) (a) Meunier, P.; de Visser, S. P.; Shaik, S. Chem. Rev. 2004, 104,
3947−3980. (b) Shaik, S.; Kumar, D.; de Visser, S. P.; Altun, A.; Thiel,
W. Chem. Rev. 2005, 105, 2279−2328.
(22) Harvey, J. N.; Aschi, M.; Schwarz, H.; Koch, W. Theor. Chem.
Acc. 1998, 99, 95−99.
(23) (a) Mairena, M. A.; Díaz-Requejo, M. M.; Belderraín, T.;
Nicasio, M. C.; Trofimenko, S.; Per
253−256. (b) Díaz-Requejo, M. M.; Per
Templeton, J. L. Organometallics 1997, 16, 4399−4402. (c) Per
J.; Brookhart, M.; Templeton, J. L. Organometallics 1993, 12, 261−262.
(24) Llavería, J.; Beltran, A.; Díaz-Requejo, M. M.; Mathieu, M. I.;
Castillon, S.; Perez, P. J. Angew. Chem., Int. Ed. 2010, 49, 7092−7095.
́
ez, P. J. Organometallics 2004, 23,
ez, P. J.; Brookhart, M.;
ez, P.
́
́
́
́
́
(25) Data in references 17b and 18 have been obtained using 3- and
4-aryl substituted styrenes in the same fitting, assuming there is no
steric effect with the 3-substituted substrates.
(26) Jiang, X.-K.; Ji, G.-Z. J. Org. Chem. 1992, 57, 6051−6056.
(27) Dincţ urk, S.; Jackson, R. A. J. Chem. Soc., Perkins 2 1981, 1127−
̈
1131.
(28) Fisher, T. H.; Meierhoefer, A. W. J. Org. Chem. 1978, 43, 224−
228.
(29) Per
A.; Maseras, F.; Echavarren, A. M. Chem. Sci. 2011, 2, 141−149.
́
́ ́
ez-Galan, P.; Herrero-Gomez, E.; Hog, D. T.; Martin, N. J.
́
(30) Mairena, M. A.; Urbano, J.; Carbajo, J.; Maraver, J. J.; Alvarez,
E.; Díaz-Requejo, M. M.; Perez, P. J. Inorg. Chem. 2007, 46, 7428−
7435.
(31) Gozdz, A. S.; Maslak, P. J. Org. Chem. 1991, 56, 2179−2189.
́
́
(32) (a) Munoz-Molina, J. M.; Sameera, W. M. C.; Alvarez, E.;
̃
Maseras, F.; Belderrain, T. R.; Per
2458−2467. (b) Díaz-Requejo, M. M.; Caballero, A.; Belderraín, T.;
Trofimenko, S.; Per
́
ez, P. J. Inorg. Chem. 2011, 50,
́
ez, P. J. J. Am. Chem. Soc. 2002, 124, 978−983.
(c) Mealli, C.; Arcus, C. S.; Wilkinson, J. L.; Marks, T. J.; Ibers, J. J.
Am. Chem. Soc. 1976, 98, 711−718.
(33) In this region of the potential energy surface, the singlet
electronic state has an open-shell biradical character, which ultimately
converts to the closed-shell singlet as the system evolves to the
1
product. The transition state for this process, TS′UKS, could not be
located, and its structure and geometry were estimated through a
relaxed scan on the N−N−C bond angle, following the approach
described in ref 20a.
(34) Gom
J. Organometallics 2008, 27, 4126−4130.
(35) Yamada, Y.; Yamamoto, T.; Okawara, M. Chem. Lett. 1975, 361.
́ ́
ez-Emeterio, B. P.; Urbano, J.; Díaz-Requejo, M.; Perez, P.
1348
dx.doi.org/10.1021/ja307229e | J. Am. Chem. Soc. 2013, 135, 1338−1348