Control of chemoselectivity by coordinated water and relative size of ligands to metal cations of Lewis acid catalysts for cycloaddition of an oxirane derivative to an aldehyde: Theoretical and experimental study

Article abstract of DOI:10.1021/om500068m

The role played by Lewis acid catalysts in the selective cleavages of C-O and C-C bonds of oxirane derivatives with aldehydes is investigated both theoretically and experimentally. According to the different chemoselectivities, various catalysts are divided into four series: C-O selectivity, both, C-C selectivity, and none, respectively. The involvement of coordinated water molecules is crucial to rationalize the experimental observation of C-C selectivity for the Ni(ClO₄)₂·6H ₂O-catalyzed reaction, which is supported by experiment on changing originally unreactive Ni(OTf)₂ to be an effective catalyst by mixing with water. Furthermore, the steric hindrance from the anion in Lewis acid and the water molecule have significant influence on the efficiency of catalysts. A steric parameter, α, defined as the relative ratio of ligand size to radius of the center metal cation, gives a general picture to understand the selectivities of various Lewis acid catalysts. The ineffective M(OTf) ₂ type catalysts have remarkable steric hindrance with α > 4.5. Large cations (R_M > 74 pm) relative to their surrounding ligands with α < 4 prefer the C-O bond cleavage path, while small cations (R_M < 70 pm) with α < 4.5 lead to C-C bond breaking. An understanding of the relationship between selectivity and Lewis acid catalysts may guide the design of more selective and versatile Lewis acid catalysts for organic synthesis.

Full text of DOI:10.1021/om500068m

Article
pubs.acs.org/Organometallics
Control of Chemoselectivity by Coordinated Water and Relative Size
of Ligands to Metal Cations of Lewis Acid Catalysts for Cycloaddition
of an Oxirane Derivative to an Aldehyde: Theoretical and
Experimental Study
Ziqi Tian,† Yuanjing Xiao,‡ Xiangai Yuan,† Zuliang Chen,‡ Junliang Zhang,*,‡ and Jing Ma*,†
†Key Laboratory of Mesoscopic Chemistry of MOE School of Chemistry & Chemical Engineering, Nanjing University 22 Hankou
Road, Nanjing, 210093, People’s Republic of China
‡Shanghai Key Laboratory of Green Chemistry and Chemical Processes Department of Chemistry, East China Normal University
3663 North Zhongshan Road, Shanghai 200062, People’s Republic of China
S
* Supporting Information
ABSTRACT: The role played by Lewis acid catalysts in the
selective cleavages of C−O and C−C bonds of oxirane
derivatives with aldehydes is investigated both theoretically
and experimentally. According to the different chemo-
selectivities, various catalysts are divided into four series:
C−O selectivity, both, C−C selectivity, and none, respectively.
The involvement of coordinated water molecules is crucial to
rationalize the experimental observation of C−C selectivity for
the Ni(ClO4)2·6H2O-catalyzed reaction, which is supported by
experiment on changing originally unreactive Ni(OTf)2 to be an effective catalyst by mixing with water. Furthermore, the steric
hindrance from the anion in Lewis acid and the water molecule have significant influence on the efficiency of catalysts. A steric
parameter, α, defined as the relative ratio of ligand size to radius of the center metal cation, gives a general picture to understand
the selectivities of various Lewis acid catalysts. The ineffective M(OTf)2 type catalysts have remarkable steric hindrance with
α > 4.5. Large cations (RM > 74 pm) relative to their surrounding ligands with α < 4 prefer the C−O bond cleavage path, while
small cations (RM < 70 pm) with α < 4.5 lead to C−C bond breaking. An understanding of the relationship between selectivity
and Lewis acid catalysts may guide the design of more selective and versatile Lewis acid catalysts for organic synthesis.
Scheme 1. C−C vs. C−O Bond Cleavage
1. INTRODUCTION
Dipolar [3 + 2] cycloaddition reactions1 are hot topics in
organic chemistry due to their wide application in the synthesis
of various compounds containing five-membered rings, the
common skeletons in natural and pharmaceutical compounds.2
In cycloaddition with 1,3-dipolar compounds, including oxirane
derivatives containing three-membered rings,3 many species
with polarized double bonds are considered as dipolarophiles.4
Oxiranes (for instance, aryloxiranyldicarboxylate, R1) have a
propensity of strain-induced ring opening and two asymmetric
centers; thus, they are very important reagents in numerous
industrial processes and syntheses of natural products.5
Nucleophilic ring opening by an SN2 mechanism has been
well recognized. However, without other reactants and
catalysts, the calculated energy barrier (24.1 kcal/mol at the
B3LYP/6-31+G(d) level) of C−C bond cleavage in R1 is
smaller than that (37.1 kcal/mol) of C−O cleavage, due to the
evident polarity of the C−O bond relative to the C−C bond.
As a result, selective C−O or C−C bond cleavage from the
same substrates is an exciting challenge. In additions with
aldehydes R2, C−O and C−C bonds of R1 could be selectively
cleaved by using different metal-centered Lewis acid catalysts,6
as shown in Scheme 1.
Received: January 22, 2014
Published: March 28, 2014
© 2014 American Chemical Society
1715
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
According to the different chemoselectivities of catalyzed
reactions for opening of oxirane by aldehyde,6c ML2 (·nH2O)
type catalysts (L= triflate or perchlorate, n = 1−6) could be
classified into four series: (I) catalysts which lead to C−O bond
cleavage only, called C−O selectivity, (II) catalysts with no
selectivity for breaking C−O or C−C bonds, named as both,
(III) catalysts which just favor C−C bond cleavage, i.e., C−C
selectivity, and (IV) ineffective catalysts for both C−C and C−O
bond breaking, labeled as none. The collection of the
aforementioned experimental results, as shown in Figure 1,
Figure 1. Classification of the ML2 (·nH2O) type catalysts with respect
to the selectivity. The experimental results in the plot are mainly taken
from ref 6c.
gives us an impetus to systematically study the role of Lewis
acid catalysts in the control of selectivity.
Figure 2. Possible C−O and C−C cleavage pathways. Ligands
coordinated to metal ions are omitted for clarity.
Theoretical rationalizations of [3 + 2] dipolar cycloaddition
reactions have been devoted to drawing possible reaction paths
and explaining chemo- and stereoselectivities.7 The qualitative
description of addition processes by frontier molecular orbital
(FMO) theory has been widely accepted. The dipolar [3 + 2]
cycloaddition reactions have been further classified into three
different types by Sustmann,8 according to the character of the
highest occupied molecular orbital (HOMO) and lowest
unoccupied molecular orbital (LUMO) interactions in
reactants. In this work, calculations will show that the studied
of Sn(OTf)2-catalyzed reactions of R1 with R2, as shown in
Figure 2.6c It was supposed that C−O bond cleavage may
proceed through an SN2 pathway, in which nucleophilic
aldehyde attacked oxirane from the back. As another possibility,
the C−C bond may be weakened by the coordination of a
Lewis acid.13 This leads to thermal ring opening to generate a
carbonyl ylide intermediate, which is an emblematic 1,3-dipole
to react with dipolarophile. C−O and C−C bond cleavage
pathways are always concomitant and competitive.
Our recent calculations on Sn(OTf)2-catalyzed reactions
show that the activation energy barrier, Ea, for C−O cleavage is
lower than that for C−C bond breaking, which is in good
agreement with the experiment that Sn(OTf)2-catalyzed
reaction mainly produced the C−O bond cleaved products
with about 90% yield.6c However, the nearly identical energy
barriers of C−O vs C−C bond cleavage catalyzed by using ideal
water-free Ni(ClO4)2 model catalysts, as depicted by the dots in
Figure 1, cannot explain the evident selectivity in C−C bond
breaking for the Ni(ClO4)2·6H2O-catalyzed experiment. In the
present work, we will show the important role of trace water in
controlling the chemoselectivity brought by Ni(ClO4)2·nH2O.
In fact, water molecules could participate in many organic
processes as a reagent to accelerate reactions, such as the water-
catalyzed [1,2]-hydrogen shift.14 We will demonstrate that the
Ea value of C−O bond cleavage increases when the coordinated
water is involved in a Ni(ClO4)2·nH2O-catalyzed reaction,
which is denoted by the star symbols in Figure 1. However, the
energy barrier of C−C cleavage decreases, enhancing the
catalytic reactions are mainly controlled by LUMOdipole
HOMOdipolarophile interactions, which are different from those
−
without catalysts, because the coordinated catalysts narrow the
energy gaps between LUMOdipole and HOMOdipolarophile
.
However, these calculations will show that the sole indicator
of the HOMO−LUMO gap cannot account for the difference
in chemoselectivities caused by different Lewis acid catalysts,
especially for C−O selectivity. For those complicated reactions
beyond the control of a simple FMO picture, density functional
theory (DFT) was widely used to gain a more quantitative
picture of reaction paths and corresponding activation energy
barriers.9 The catalytic effects of noble metals (for instance, the
complexes of Rh, Pt, et al.)10 and Lewis acids (such as the
metallic salts and BF3)11 have been addressed theoretically.
Despite these advances in the rationalization of chemo- and
stereoselectivities, there are still some open questions for
understanding all of the experiments. For example, whether the
dipolar addition is stepwise or concerted is hotly debated.12
Recently, in cooperation with experiment of a series Lewis
acid catalyzed reactions, we explored the possible reaction paths
1716
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
Afterward, R1 (125 mg, 0.5 mmol) and 4 mL of DCM containing
some H2O were added. The reaction mixture was stirred at room
temperature for 24 h. The reaction mixture was then passed over a
plug of silica with 30 mL of Et2O. The solvent was removed under
reduced pressure, and the ratio of stereoisomers was determined by 1H
NMR spectroscopy using dibromomethane as the internal reference.
For more details, see the Supporting Information and ref 6c.
preference for C−C bond cleavage. To give more evidence for
our hypothesis of water-improved selectivity, some supple-
mentary experiments have then been carried out by mixing
additional trace water into the originally ineffective catalyst
Ni(OTf)2, from which C−C bond cleaved products are mainly
obtained. Furthermore, a simple geometry indicator, defined as
the relative size of the anion (including coordinated water
molecules) with respect to that of the cation for the metal-
centered Lewis acid catalysts, correlates well with the
chemoselectivity for a broad range of Lewis acids, covering all
four series of catalysts shown in Figure 1. Using the radii of
cations and this indicator, we predict that the Mn(ClO4)2-
catalyzed reaction may favor the C−O cleavage path, which is
further supported by experiment.
3. RESULTS AND DISCUSSION
Without Lewis acid catalyst, both C−O and C−C bond
cleavages are impossible at room temperature.23 It will be
demonstrated in section 3.1 that employment of many Lewis
acid reagents does promote C−O and C−C bond cleavages.
Through the consideration of coordinated water in Ni(ClO4)2·
6H2O on selectivity, we will also show in section 3.2 that trace
water molecules tend to coordinate to metal cations,
subsequently altering the height of the energy barriers and
affecting the activity and selectivity of the catalyst. From the
perspective of steric hindrance and electrostatic effects, we
attempt to propose a geometry indicator to generalize the
selectivity by using different Lewis acid catalysts.
2. THEORETICAL MODELS AND COMPUTATIONAL
AND EXPERIMENTAL DETAILS
To understand the chemoselectivity of Lewis acid catalyzed cyclo-
addition reactions of R1 with R2, possible reaction pathways (Figure 2)
for the C−C and C−O cleavages were theoretically studied by using
DFT, respectively. Some related experiments were also performed to
test the theoretical predictions on selectivities of other catalysts.
2.1. Computational Details. All DFT calculations were carried
out with the Gaussian09 program package.15 The B3LYP functional16
was employed for optimizations, with the standard 6-31+G(d) basis
sets for nonmetal elements. The 6-311+G(d) full-electron basis set
was used for Mg. For the other heavier metals, LANL2DZ effective
core potentials were used.17 An alternative functional, M06-2X,18 was
used to study the Sn(OTf)2-catalyzed path to test the influence
of different functionals (Figure S1, Supporting Information). The
M06-2X functional result could give the same conclusion, that the C−O
bond cleavage is favorable, as the B3LYP result and experiment. C−O
bond cleavage was correctly predicted as the time-limiting step in the
C−O path. While M06-2X requires more computational cost than
B3LYP and gives some unreasonable structures (such as IM2-Sn,
shown in Figure S1), the B3LYP functional is used in all the following
computations. The stabilities of wave functions were also tested. The
restricted methods were applicable for most species, except for those
in manganic and nickelous salt catalyzed pathways. Taking Ni(ClO4)2
as an example, the geometries of its singlet and triplet were optimized
by UB3LYP. High-spin Ni(ClO4)2 takes a tetrahedral geometry, while
the low-spin complex tends to be a square-planar structure. The Gibbs
free energy of the triplet is 7.5 kcal/mol energetically lower than that
of the singlet. These intermediates containing Ni2+ were found to have
strong interference in the NMR experiments, supporting that these
nickel species preferred the high spin ground states. It can also be
found from Figure S2 (Supporting Information) that the spin densities
of the high-spin Lewis acid catalysts and reactive species are mainly
localized on the metallic center. Frequency calculations were
performed to test obtained stationary points, intermediates (IM), or
transition states (TS) and to get their zero-point energies and other
thermochemical corrections at 298 K. Only vibrational contributions
to the thermal corrections were considered, because rotation and
transition movements were highly suppressed in solution under the
experimental conditions.19 Single-point energies were computed at the
B3LYP/6-311++G(d,p) level (SDD for metals20) with the SMD
solvation model (dichloromethane solvent, ε = 8.9) at optimized
geometries.21 In discussions, Gibbs free energies of reaction pathways
in solution will be employed in the rationalization of chemoselectivity
of the [3 + 2] cycloaddition. In addition, natural bond orbital (NBO)
calculations22 were performed to show the partial atomic charges and
localized molecular orbitals.
3.1. Overview of Sn(OTf)2-Catalyzed Pathway: C−O
Cleavage Selectivity. Although some preliminary calculated
results of Sn(OTf)2-catalyzed reactions have been reported,6c
some details of the energy profiles as well as the geometries of
some important stationary points are crucial to gain a
comprehensive picture of these reactions. In R1, it is possible
for three oxygen atoms, represented as O1 (from oxirane ring)
and Oa and Ob (from carbonyls) in Scheme 1, respectively, to
coordinate with metal cations of Lewis acid catalysts. In IM1,
the cation binds to the sp3 oxygen (O1) of oxirane and one of
the sp2 oxygens of carbonyl groups (Oa or Ob), activating the
C−O bond of oxirane. Alternatively, if the cation coordinates to
two carbonyl groups to form IM6, C−C cleavage would occur.
As shown in Scheme 2, C−O bond cleavage follows an
SN2-like mechanism. R2 attacks the benzyl carbon atom of
IM1-Sn, leading to C−O bond cleavage and formation of a new
C−O bond. There are two directions of attack for R2, the front
and back of the cleaving C−O bond, as shown in Figure 3. As
expected, an SN2 reaction with the nucleophile group attacking
the reaction center from the back of the leaving group is more
favorable, with an energy barrier of about 6.5 kcal/mol. The
dihedral angle O1−C2−C3−O4 of TS1-Sn is −167.5°, implying
that it is going to form the gauche conformation intermediate
IM2-Sn. It is very difficult to locate the rotation transition state
before ring closure, due to the flat rotation potential energy
surface. A scan of potential energy curves for rotation angle was
performed to find the transition structure TS2-Sn with an eclipsed
conformation (dihedral angles O1−C2−C3−O4 are nearly 120°
and energy barriers are estimated to be 8.5 and 9.4 kcal/mol,
respectively). Taking into account the structure relaxation, the
rotation energy barrier in the real reaction may be even lower than
calculated barriers. In comparison with facile processes such as
single-bond rotation and new bond formation between atoms with
opposite charges, C−O bond cleavage from back attack is
considered to be the rate-limiting step.
In IM3-Sn, C5 possesses a positive charge (0.47 au) and O1
(−0.98 au) has a negative charge; thus, the formation of a new
bond between these two atoms is energetically favored with
relatively low energy barriers of 3.7 and 2.4 kcal/mol for trans
and cis paths, respectively. The cyclization would occur via
transition states TS3-Sn (for trans) or TS4-Sn (for cis) with
eclipsed conformations (dihedral angles O1−C2−C3−O4 are
2.2. Experimental Details. To test the influence of trace water, a
typical procedure was run as follows: in an inert-atmosphere glovebox,
a flame-dried vial was charged with 5 mol % of Ni(OTf)2 catalyst,
100 mg of activated 4 Å molecular sieves (MS), and a magnetic stir bar.
Outside of the glovebox, the vial was placed under an N2 atmosphere
and charged with 1 mL of DCM, followed by R2 (56 μL, 0.55 mmol).
1717
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
a
Scheme 2. B3LYP Free Energy Profile for [3 + 2] Cycloaddition Reaction Paths Catalyzed by Sn(OTf)2
a
The results calculated at the M06-2X/6-311++G(d,p)/SDD //M06-2x/6-31+G(d)/LAN2DZ level are also given in brackets.
slower than the forward reaction. Our models and calculations
agree well with the experimental fact that the trans product P1
could transform into the cis product P2 when Sn(OTf)2 is
present.
C−C bond cleavage resembles a [3 + 2] addition
mechanism. The C3−C2 bond of oxirane can be activated by
catalyst coordination with two carbonyl groups, forming IM6-Sn.
The energy barrier of C−C bond cleavage is 11.1 kcal/mol,
higher than that (6.5 kcal/mol) for C−O cleavage. The broken-
bond intermediate state IM7-Sn would be stabilized by charge
transfer from O1δ− to C2 and C3 atoms (IM1-Sn → IM7-Sn:
−0.49 au → −0.02 au). In IM7-Sn, the plane determined by
C3, O1, C2 and the plane determined by C2 and the two
carbonyl groups are not strictly coplanar, with a dihedral angle
of about 63.5°, implying a weak conjugation effect. After C−C
bond cleavage, the activated site at C3 (with a charge of 0.03 au)
of IM7-Sn is then attacked by the nucleophilic oxygen atom O4
(−0.54 au) of R2. Transition states TS6-Sn and TS7-Sn are
going to generate trans and cis products, respectively, through
formation of IM8-Sn and IM9-Sn intermediates (Figure 4). The
activation energy barrier is 15.0 kcal/mol for trans addition and
12.4 kcal/mol for cis addition, greater than the barrier of C−C
bond breaking pf 11.1 kcal/mol; hence, the addition step is also
predicted to determine the reaction rate.
3.2. Ni(ClO4)2·nH2O-Catalyzed Pathway: C−C Cleavage
Selectivity. 3.2.1. Catalysis without Involvement of Coordi-
nation Water. The energy profile is shown in Figure 5a, and
optimized structures of intermediates and transition states on
reaction paths are displayed in Figure 6. In all of these nickelous
species, Ni cations are at the coordinated octahedral centers of
the complex. For the C−O bond cleavage step, the energy
barrier is 11.2 kcal/mol, obviously higher than the C−O
cleavage barrier catalyzed by Sn(OTf)2 (6.5 kcal/mol). Thus,
under similar experimental conditions, Ni(ClO4)2-catalyzed
C−O cleavage is difficult to carry out. The rotation energy
barrier from TS2-Ni to IM3-Ni is 6.8 kcal/mol. The potential
energy surface near IM3-Ni is extremely flat. Although the
structures of IM3-Ni, TS3-Ni, and TS4-Ni have very significant
differences in the bond lengths of C3−C5 and dihedral angles,
φ(C5−O4−C3−O1), their internal energies are almost the
same. For the C−C bond cleavage path, C−C bond is cleaved
from IM6-Ni to TS5-Ni with an energy barrier of 13.7 kcal/mol.
In IM7-Ni, C3 and C2 both bear sp2 hybrid orbitals. The two
orbital planes of C3 and C2 are almost coplanar with an almost
Figure 3. Optimized stationary point structures on the C−O bond
cleavage path catalyzed by Sn(OTf)2.
49.2 and 11.0°, respectively), shown in Figure 3. Going through
IM4-Sn (IM5-Sn), trans (cis) product is obtained. The
population ratio of cis and trans products is estimated to be
3.3 from Boltzmann statistic calculations. The possibility of
reverse reactions is indicated by relatively low gaps of 9.9 and
14.8 kcal/mol for TS3-Sn → IM4-Sn and TS4-Sn → IM5-Sn,
respectively. While these gaps are larger than the activation
energy barrier of the time-limiting step of the forward reactions
(no more than 10 kcal/mol), the reverse reactions from
IM4-Sn and IM5-Sn to IM3-Sn are thus possible but are much
1718
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
water molecules (n = 2) should have six isomers, denoted 2H2Oab,
2H2Oac, 2H2Oad, 2H2Obc, 2H2Obd, and 2H2Ocd (Table S1,
Supporting Information). The case with more substituted water
molecules may be rare because of the following issues. First, if the
third water molecule coordinates with Ni2+ and replaces one ClO4−
group, the reactant complex cannot keep a nonzero total charge.
Obviously the charge-separated system is difficult to form in a weak
polar solvent, such as CH2Cl2 or toluene in our experiments.
Second, the redundant water is actually more favorable to
replace one coordination site of R1 instead of ClO4−; thus,
Ni(ClO4)2(OH2)3-(η1-R1), an unreactive species, would appear
instead of Ni(ClO4)(OH2)3-R1. The third and perhaps the most
important reason is that water is a typical dipolarophile; thus, the
excess water may tend to attack the reactive site, leading to side
reactions. (In experiments, 4 Å molecular sieves are used to remove
excess water. Even a little more water may cause vicinal diol
generation.)
Water in the reaction system tends to coordinate with
Ni2+. To compare the binding ability of water with Ni2+, we
IM1
IM6
define ΔE
(ΔE ) as the energy differences generated by
H2O
H2O
= EIM1(with nH2O) −
IM1
water coordination: i.e., ΔE
H2O
IM1
H2O
E
IM1(without H2O) − n[E(H2O)]. From the values of ΔE
Figure 4. Species of C−C bond-cleavage reactions catalyzed by
Sn(OTf)2.
IM6
H2O
and ΔΔE , we found that water replacing is extremely
energetically favored, since a larger six-membered-ring structure
−
having hydrogen bonds between the inserted water and ClO4
0° dihedral angle (0.90°) of two esters, which is much smaller
than that (63.5°) in the Sn(OTf)2-catalyzed path. Energy barriers
of trans (from IM7-Ni to TS6-Ni) and cis (from IM7-Ni to
TS7-Ni) addition steps are 15.8 and 8.8 kcal/mol, respectively.
The energy difference between TS6-Ni (for trans) and TS7-Ni
(for cis) is 7.0 kcal/mol, while for Sn(OTf)2-catalyzed reaction,
the energy difference between TS3-Sn (for trans) and TS4-Sn
(for cis) is just 1.3 kcal/mol. This shows that Ni(ClO4)2
processes have stereoselectivity higher than those of Sn(OTf)2.
For Ni(ClO4)2-catalyzed paths, the barrier of the rate-limiting
step for the C−O cleavage path is 11.2 kcal/mol, which is very
close to that (13.7 kcal/mol) of the C−C path. Unfortunately,
this does not agree with the experiment that only C−C bond
cleavage products were obtained.
is formed (shown in the inset of Figure 5b and in Figure 6)
instead of the chelating four-membered ring. The ideal
coordination octahedron has three 4-fold symmetry axes. The
average value, θav, of the three axial angles(such as ∠a−Ni−b in
Figure 5b) can reflect distortion of the geometry from an ideal
octahedron. Obviously, for an ideal octahedron, θav is 180°. The
values of θav for IM1 and IM6 without coordinated water are
165.1 and 165.4°, respectively. Water coordination keeps the
octahedral configuration. When n = 1, the θav values for IM1
and IM6 increase to 167.1 and 168.8° and further to 172.2 and
175.5° if n = 2, implying reinforced orbital overlap between
nickel and ligand orbitals and a weakened stretch.
For C−O cleavage, if n = 1, 1H2Oc leads to the largest value
IM1
H2O
We may recall that in experiments nickel(II) perchlorate hexa-
hydrate is used as a catalyst instead of anhydrous nickel(II)
perchlorate, since the anhydrous salt is always dangerous in organic
reactions. Although 4 Å molecular sieves are mixed in reaction
systems for removing water, there may be still trace water in the
system. The following calculations will reveal that trace water may
coordinate to Ni2+ and change reaction energy barriers sensitively.
3.2.2. Effect of Trace Water. Ni2+ is at the center of an
octahedron. R1 occupies two coordination sites, and the
remaining four sites can be occupied by other ligands: for
instance, ClO4− and H2O. ClO4− could be a chelating ligand by
of ΔE
(14.8 kcal/mol), indicating the largest population in
the single-water-coordinated complexes. For the 1H2Oc-catalyzed
pathway, the activation energy barrier for C−O bond cleavage
(13.7 kcal/mol) is higher than that without water (11.2 kcal/mol)
and any other single-water-coordinated paths. Similarly, for n = 2,
complexes 2H2Oab and 2H2Ocd have relatively large value of
ΔE (26.8 and 26.7 kcal/mol, respectively), and C−O cleavage
IM1
H2O
energy barriers are 13.9 and 13.3 kcal/mol, also higher than the
non-water-coordinated energy barrier (11.2 kcal/mol). Thus,
water coordination may weaken the tendency for C−O bond
cleavage. On the other hand, for the C−C bond cleavage path, we
considered the water-coordinated influence on both the C−C
bond cleavage and [3 + 2] addition steps. Complexes 1H2Ob and
2H2Oab on the C−C cleavage path are the dominant species with
using two oxygen atoms (represented as η2-ClO4 ) or just a
−
−
monodentate ligand (represented as ClO4 ) if a water molecule
occupies a coordination site. In other words, water of
crystallization can replace one of the chelating oxygen atoms
ΔEIHMO6 values of 13.2 and 28.6 kcal/mol, respectively. When n = 1,
−
of the ClO4 group. As shown in Figure 5b, the other four
2
for 1H2Ob, the energy barrier of C−C bond cleavage decreases
from 13.7 to 10.0 kcal/mol. In addition, when n = 2 as 2H2Oab,
this energy barrier is 11.6 kcal/mol. Simultaneously, energy
barriers of cycloaddition leading to cis product change slightly,
from 8.8 to 8.9 kcal/mol for 1H2Ob or 9.3 kcal/mol for 2H2Oab.
In contrast, energy barriers leading to trans product rise relatively
obviously by about 4 kcal/mol (19.1 kcal/mol for n = 1,
and 16.8 kcal/mol for n = 2), enhancing stereoselectivity.
R1-unoccupied coordination sites are marked a−d, respectively.
For a water-free catalyst, two η2-ClO4− ligands occupy the a−b
and c−d sites, respectively. Ni(η2-ClO4)2-R1, which has been
calculated as IM1 or IM6, has no geometric isomer but its own.
However, for the single-water-coordinated complex (represent
as n = 1 in the following), Ni(η2-ClO4)(ClO4)OH2-R1, there
are four isomers denoted 1H2Oa, 1H2Ob, 1H2Oc, and 1H2Od,
respectively. Similarly, Ni(ClO4)2(OH2)2-R1 with two coordination
1719
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
Figure 5. Free energy profile for [3 + 2] cycloaddition reaction paths catalyzed by Ni(ClO4)2·nH2O (a) without and (b) with the consideration of
water coordination. All energies (in kcal/mol) are computed at the UB3LYP/6-311++G(d,p)/SDD//UB3LYP/6-31+G(d)/LANL2DZ level with
TSi
IMi
SMD model for CH2Cl2 solvent. ΔΔG is the change in energy barrier from IMi to TSi.
With consideration of water, energy barriers heights of C−C bond
cleavage (10.0 kcal/mol) and [3 + 2] cis addition (8.9 kcal/mol)
are comparable to each other. It can be conceived that lowering of
the activation barrier in either the C−C cleavage or [3 + 2]
addition step may promote reaction activity. In experiments, it was
also found that electron-withdrawing substituents on oxiranes, e.g.,
two carboxyl groups in R1, are necessary for the reactant to
facilitate C−C bond breaking and, alternatively, electron-donating
groups on benzaldehyde, which increase the nucleophilicity of R2
and decrease the energy barrier of addition, can also accelerate the
reaction.6c In summary, for water-coordinated nickel(II) com-
plexes, energy barriers of C−O bond cleavage rise, while barriers
of C−C cleavage and [3 + 2] cis addition decrease or change
slightly, which indicates that water coordination may change the
chemoselectivity and stereoselectivity of the Lewis acid catalyst.
Furthermore, we calculated the activation energy barriers of
the rate-limiting step for Ni(OTf)2-catalyzed pathways to
investigate the role of trace water further. Ni(OTf)2 is an
ineffective catalyst for both C−O and C−C cleavage in
experiments (Figure 1). Without the consideration of water,
Ni2+-catalyzed reactions have similar energy barriers whether
catalytic effect in DCM in the presence of molecular sieves. The
control experiments further support our hypothesis that trace
water can enhance the catalytic activity and lower the activation
energy of C−C bond cleavage.
It is also possible that water may be the source of protons
that promote the reaction. The control experiments with
5 mol % TfOH also rule out the possibility that the proton
(which may come from water) can catalyze the [3 + 2] cyclo-
addition reaction involving both C−O and C−C bond cleavage
(eq 1 in Scheme 3). Interestingly, when 2,6-di-tert-butylpyridine
was added to the standard reaction conditions, neither C−O nor
C−C cleavage cycloaddition products could be obtained. This
indicated that 2,6-di-tert-butylpyridine may coordinate to the
nickel cation, resulting in ineffective catalysis (eq 2 in Scheme 3).
3.3. General Trends in Chemoselectivity of Lewis Acid
Catalyzed Reactions. Versatile Lewis acid catalyzed reactions
have been applied to many important processes in organic
synthesis.24 In most cases, various types of Lewis acid catalysts
have been screened after trial and error, due to limited
knowledge of the electronic character of the Lewis acid. Thus,
an in-depth understanding about the role of Lewis acid catalysts
is necessary for designing various chemoselective reactions.
There are some general classifications of Lewis acids according
−
the ligand is OTf− or ClO4 . When an appropriate amount of
water was added (Table 1), Ni(OTf)2 showed a moderate
1720
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
cations (Ni2+, Mg2+, Cu2+, Zn2+, Hg2+, Sn2+) can be classified
into four types according to chemoselectivity. A relationship
between the effective cationic radius, RM,28 and chemoselectivity
is presented in Table 2. The cations are hexacoordinated and are
Table 2. Relationship between Chemoselectivity and
Average Spatial Hindrance from the Ligands, Including Acid
Anion (αligand) and Water Molecule (αwater), Relative to the
Size of the Active Metal Center (RM) in Lewis Acid Catalysts
RM
C−O:C−C
a
catalyst
Sn(OTf)2
(pm) αligand αwater
α
selectivity
yield (%)
̅
112
102
102
84
3.5
3.3
3.8
4.0
4.5
4.7
4.9
4.7
5.3
5.4
5.7
3.5
2.6 3.0
3.8
C−O
C−O
C−O
C−O
C−O
both
90:0
Hg(ClO4)2·6H2O
Hg(OTf)2
89:0
90:0
Mn(ClO4)2·6H2O
Zn(ClO4)2·6H2O
Mg(ClO4)2·6H2O
Ni(ClO4)2·6H2O
Mn(OTf)2
3.2 3.6
3.6 4.0
3.7 4.2
3.9 4.4
4.7
78:21
80:18
48:50
0:99
74
72
69
C−C
none
none
none
none
Figure 6. Optimized structures on the Ni(ClO4)2·xH2O (x = 0−2)
pathways. Selected bond lengths are given in angstroms, and the
lengths of hydrogen bonds (HB) are also shown.
84
Zn(OTf)2
74
5.3
Mg(OTf)2
72
5.4
Ni(OTf)2
69
5.7
Table 1. Effect of Trace Water on Ni(OTf)2-Catalyzed
Reaction
a
a
Most of experimental yields have been reported in ref 6c except that
using Mn(ClO4)2*6H2O and Mn(OTf)2, which are added in this
work.
at the centers of octahedra with ∠O−M--O angles of about 90°.
For the C−O bond cleavage path, the metal cation coordinates
to R1 to form a distorted five-membered ring with relatively
longer M−O bonds. For example, in IM1-Sn, the lengths of two
Sn−O bonds are 2.43 and 2.62 Å, respectively. However, for
C−C bond cleavage, the cation is a part of a planar six-
membered ring to make the redundant charge delocalized on the
b
yield (%)
entry
x (mol % H2O)
P1 and P2
P3 and P4
1
2
3
4
0
10
20
0
0
0
trace
28
trace
carbon atom of oxirane. Thus, a cation with small radius (RM
<
30
trace
80
a
b
70 pm, such as Ni2+) may prefer C−C bond cleavage; in contrast,
a large cation (RM > 74 pm, such as Sn2+ and Hg2+) prefers C−O
bond cleavage.
[R1] = 0.1 M. Estimated by 1H NMR spectroscopy using
dibromomethane as the internal reference.
From Table 2, it can be seen that the selectivity not only
relates to RM but also to the size of surrounding ligands relative
to the cation. Intuitively, the spatial hindrance from ligands,
including acid anion and water molecule, is also responsible for
the catalytic reactivity. The large ligand may bury the cation and
inhibit the coordination between cation and R1, leading to
weak reactivity. Trace water in the reaction system may also
coordinate to the cation to disturb the surroundings of the
anion ligand. The sizes of water and other anion ligands can be
estimated with volume keyword by using Gaussian 09 software.
The radius, R, of water (Rwater = 266 pm) is smaller than those
Scheme 3. Control Experiment To Validate Water Effect
to the classical HSAB (hard and soft acids and bases) theory,25
the catalytic effect on Friedel−Crafts or other reactions,26 σ and
π electrophilic properties,27 etc. Here, we will address general
trends in Lewis acid catalyzed cycloaddition reactions from
both geometry considerations and FMO.
−
−
(ROTf = 391 pm, RClO = 335 pm) of employed anions in this
4
work. As a result, water replacement of the anions may weaken
hindrance from the anion ligand. We set αligand = Rligand/RM as a
parameter to measure the steric hindrance from the anion
3.3.1. Geometry Indicators of Catalysts: Relative Size of
Ligands vs Metal Cation. ML2 type catalysts with bivalent
1721
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
Scheme 4. Experimental Result of Mn(ClO4)2·6H2O- and Mn(OTf)2-Catalyzed Reactions
ligand. For Hg2+, Zn2+, and Mg2+ cations, the inner atom
orbitals are fully filled with electrons; thus, they are all
tetracoordinated with two anions in the centers of tetrahedra
without coordination water or hexacoordinated with anions and
coordination water molecules in the centers of octahedrons. In
spite of the fact that Ni2+ has a d8 configuration, the high-spin
nickel complexes also have the structures similar to those
our guess the geometry is dominant, rather than spin
preference, in controlling chemoselectivity, we implemented
an experiment on Mn(ClO4)2·6H2O- and Mn(OTf)2-catalyzed
reactions. Mn2+ is a typical high-spin cation with a ground-state
spin multiplicity of 6 in triflate or perchlorate salts. RMn is 84
pm, larger than that of Zn2+. According to the geometry
indicator, Mn(ClO4)2·6H2O is predicted to have C−O
selectivity (Table 2), which is supported by experiment that
the yield of C−O cleavage products is about 78% (Scheme 4).
However, for Mn(OTf)2, α is 4.7 and this salt is an ineffective
catalyst in experiment.
−
mentioned above. Since ClO4 and OTf− are both weak-field
ligands, coordinate bonds are mainly formed by the outer
orbitals of Ni2+ and orbitals from the ligands. In order to further
consider the effect of water coordination, αwater (Rwater/RM) was
also set. The average ligand hindrance, α, from anion and water
3.3.2. Shifts of Energy Levels of R1 by Catalysts. FMO
theory has been widely used to describe the catalyst effects on
reaction paths.8 Catalyst effects are usually rationalized from the
variations of FMO of R1, especially LUMOR1. The LUMOs
that localized at C−C and C−O bonds are mainly assigned as
antibonding σ* type. For the C−O cleavage path, HOMOR2
interacts with LUMOR1 (or LUMOIM1). A narrower
HOMOR2−LUMOR1 energy gap (HOMOR2−LUMOIM1 after
catalyst coordination) predicts a lower activation energy for
C−O cleavage. In comparison, the first step of the C−C cleavage
path is bond breaking of R1. If the HOMOR1−LUMOR1 energy
gap (HOMOIM6−LUMOIM6 after catalyst coordination) is
lowered by the catalyst, the thermal C−C bond cleavage may
proceed, and IM6 turns out to be IM7/8 via TS6. Subsequently,
the addition of IM7 with R2 occurs, and the HOMOR2−
LUMOIM7 energy gap determines the reactivity of cycloaddition.
We labeled energy gaps corresponding to C−C and C−O bond-
breaking and [3 + 2] addition steps in Table 3. Apparently,
Sn(OTf)2 decreases the HOMOR2−LUMOIM1 energy gap for
C−O cleavage (from 6.16 eV without catalyst to 2.62 eV).
However, Ni(ClO4)2, especially coordinated with one water
molecule, improves C−C cleavage, which is in accord with
experiment and theoretical calculations. For other Lewis acid
catalysts, FMO results can also give qualitative predictions for
selectivity. The reaction tendency for C−O bond cleavage may
be underestimated from the HOMOR2−LUMOR1 energy gap,
partially because the SN2 mechanism is not solely dependent on
the HOMO and LUMO.
̅
coordination is evaluated as 0.5(αligand + αwater), because in the
tetrahedral coordination, water molecules and anions occupy
two coordinated sites, respectively. The catalyst with a small
cation (RM < 70 pm) and relatively small value of α (α < 4.5)
̅
̅
tend to favor C−C cleavage. However, a α value that is too
̅
small always corresponds to large RM (such as Sn(OTf)2, α =
3.5 and RM = 112 pm), leading to C−O cleavage. The larger the
α value is, the weaker the catalyzed effect and chemoselectivity.
If α > 4.5, the catalyst may be ineffective for cycloaddition.
Thus, the catalysts with ClO4 anions often have reactivity
̅
̅
−
higher than that with the more space-demanding OTf− group.
Cu2+ is a special case due to its d9 configuration. The
tetracoordinated Cu2+ is always at the center of a square plane.
For these planar complexes, such as Cu(OTf)2, hindrance along
the axis is weak. Thus, although Cu2+ has a size (73 pm) similar
to that of Mg2+ (72 pm), the steric hindrance from ligands may
be weaker than for the corresponding magnesium salt.
Cu(OTf)2 is a catalyst with no selectivity, the same as
Mg(ClO4)2·6H2O (α = 4.2) but different from the ineffective
̅
Mg(OTf)2 (α = 5.4) catalyst. However, for Cu(ClO4)2·6H2O,
̅
crystalline water molecules may coordinate to Cu2+ from the
axial direction, generating the hexa-oordinated Cu2+. In
consideration of various possible coordination modes ranging
from 4ClO4/2H2O to 2ClO4/4H2O, the average ligand
hindrance indictor, α, is estimated to be 4.3−3.9. Actually,
̅
Cu(ClO4)2·6H2O has a catalytic selectivity for C−C bond
breaking similar to that of Ni(ClO4)2·6H2O (α = 4.4).
̅
It is also interesting to test whether the cations with non-
zero-spin preference, such as Ni2+ and Cu2+, complicate the
reaction mechanism and hence challenge the application of
such a simple geometry indicator, αligand. The spin density
distributions of the bond-breaking transition state (TS5-Ni)
and ylide intermediate (IM7-Ni) on the Ni(ClO4)2-catalyzed
path (Figure S2, Supporting Information) have been analyzed,
showing that spin densities are mainly located on metal centers
without significant spin polarization and delocalization. To test
3.3.3. Stepwise or Concerted? A great number of studies
have focused on whether [3 + 2] addition is concerted or
stepwise.12 Mongin et al.12a investigated the cycloaddition of
oxirane derivatives with aldehydes by experiment and
calculation. They proposed that electrophilically activated
oxiranes reacted with aldehydes concertedly because of the
reverse charge transfer from oxiranes to aldehydes. In our work,
there are also electrophilic substitutions (such as CO) on
oxiranes to activate the C−C bond. By thermal excitation, the
1722
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
mode of C−O formation was eliminated, the other vibration
mode representing ring closure arose immediately. For these
Lewis acid catalyzed reactions, [3 + 2] addition steps may still
exhibit stepwise character, although ring-opening transition
states have hardly been located.
Table 3. Some Important Energy Gaps of Different Lewis
Acid Catalysts for the C−O and C−C Cleavages
4. CONCLUSION
Chemoselectivity of the Lewis acid catalyzed cycloadditions of
aryloxiranedicarboxylates with aldehydes have been systematically
investigated through theoretical calculations. C−O and C−C
bonds of oxirane may be cleaved using different catalysts. We
divided ML2·nH2O type catalysts into four series based on
chemoselectivity: C−O selectivity, both, C−C selectivity, and none.
C−O bond cleavage occurs via an anti attack as an SN2
reaction, while the C−C bond is cleaved to form a polar ylide
intermediate and then cycloaddition proceeds. For the C−C
cleavage path, the energy difference between transition states
leading to the cis and trans products are often larger than that
in the C−O cleavage, indicating better steroselectivity. The
coordinated water plays an important role in tuning the
selectivity of the Ni(ClO4)2-catalyzed reaction. It was found
that water coordination increases the energy barrier of C−O
bond cleavage but decreases that of C−C bond cleavage. In our
experiment, Ni(OTf)2 can change from an inactive species to a
C−C bond selective catalyst with the addition of some trace
water, showing the unexpected effect of water.
The general trends of the Lewis acid catalyzed cycloaddition
has been discussed from the perspective of both geometries of
catalysts and FMO. It is suggested that a divalent cation with a
large radius (RM > 74 pm, such as Sn2+ and Hg2+) in Lewis acids
tend to catalyze C−O bond cleavage. If a cation (such as Ni2+) is
small, its surrounding anion in Lewis acid may prevent the
coordination of the cation with the reactant, weakening the
catalytic effect. The size of the coordinated anion will play a critical
role in selectivity and efficiency. A geometry parameter, α, is
defined to measure the steric hindrance from ligands. A large value
of α indicates significant hindrance and vice versa. A catalyst with a
small cation (RM < 70 pm) and little steric hindrance (α < 4.5)
may be effective for C−C cleavage, while those catalysts with large
values of α (α > 4.5) are often inactive. A coordinated water
molecule may lead to improved catalyst efficiency and chemo-
selectivity by reducing the ligand hindrance relative to the
surrounding large anion ligand.
C−C path
C−O path
ΔEC−C ΔER−IM ΔEIM−R
b c d
a
catalyst
ΔEC−O (eV)
(eV)
(eV)
(eV)
selectivity
none
6.16
2.62
3.75
4.20
3.72
4.73
5.04
3.66
4.86
3.69
4.73
5.59
2.82
3.09
3.10
4.75
2.92
3.27
2.93
3.55
4.74
3.25
4.32
2.65
3.56
3.71
3.59
2.90
3.92
2.90
3.48
2.97
3.53
3.59
4.68
4.67
4.31
4.44
5.04
4.46
5.06
4.64
4.97
4.55
C−C
C−O
C−O
C−O
C−O
both
Sn(OTf)2
Hg(ClO4)2
Hg(OTf)2
Zn(ClO4)2
Mg(ClO4)2
Cu(OTf)2
Ni(ClO4)2
Cu(ClO4)2
Ni(OTf)2
Zn(OTf)2
Ni(ClO4)2·nH2O
n = 1
both
C−C
C−C
none
none
e
4.81
4.73
2.81
2.85
2.95
3.96
4.89
4.44
C−C
C−C
n = 2
a
b
ΔEC−O = E(LUMOIM1) − E(HOMOR2). ΔEC−C = E(LUMOIM6) −
c
d
E(HOMOIM6). ΔER−IM = E(LUMOIM7) − E(HOMOR2). ΔEIM−R
=
e
E(LUMOR2) − E(HOMOIM7). Calculation with water coordination.
Using FMO theory, a qualitative picture can also be obtained
to describe the selectivities of different catalysts. The energy
gaps between LUMOIM6 and HOMOIM6 determine the energy
barriers of C−C bond cleavage, while gaps between LUMOR2
and HOMOIM1 may partially correspond to the energy barriers
of C−O bond cleavage. From the orbital phases of IM7 and R2,
it is suggested that cycloaddition is not concerted but stepwise.
An understanding of the relationship between catalyst selectivity
and cations, ligands, and coordinated water of Lewis acids may
have interesting implications in the development of versatile
Lewis acid catalyzed reactions.
C−C bond of R1 may cleave to form an ylide, represented as
R1-ylide. This ylide is a typical 1,3-dipole to add to aldehydes.
The HOMOR1‑ylide−LUMOR2 energy gap (3.59 eV) is close to
that (3.73 eV) of HOMOR2−LUMOR1‑ylide, which indicates that
charge transfer between the dipole and aldehyde is possibly
bidirectional and cycloaddition may be concerted. A previous
study showed that there exists reverse net charge transfer in a
reaction similar to cycloaddition of thermally opened ylide with
aldehyde.12a However, if R1 coordinates with the catalyst
(taking Ni(ClO4)2 as an example), the energies of the HOMO
and LUMO of the complexes both decrease; thus the
HOMOR2−LUMOIM7 energy gap (2.89 eV) is smaller than
that (5.06 eV) of HOMOIM7−LUMOR2, leading to unidirec-
tional charge transfer from R2 to IM7. On the other hand, the
orbital phase of R2 does not match that of IM7, weakening the
tendency for concerted cycloaddition. During the calculation of
the reaction step of addition, we found that it was hard to locate
a local minimum without any imaginary frequency after C−O
bond formation. Once the imaginary frequency with vibration
ASSOCIATED CONTENT
* Supporting Information
■
S
Figures S1−S5, giving optimized structures on Sn(OTf)2-
catalyzed paths at the M06-2X/6-31+G(d)/LANL2DZ level,
spin density distributions of several nickelous complexes of
high-spin states, energy profile of the front and back attacks for
the Sn(OTf)2-catalyzed C−O cleavage path, structures of
various Lewis acid catalysts, and structures of species on the
1723
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
(5) For reviews of the ring opening of oxiranes, see: (a) Parker, R. E.;
Isaacs, N. S. Chem. Rev. 1959, 59, 737−799. (b) Jacobsen, E. N. Acc.
Chem. Res. 2000, 33, 421−431. (c) Vilotijevic, I.; Jamison, T. F. Angew.
Chem., Int. Ed. 2009, 48, 5250−5281.
(6) For some of our recent examples of chemoselective cyclo-
additions of aryloxiranyldicarboxylate with aldehyde see: (a) Chen, Z.
L.; Wei, L.; Zhang, J. L. Org. Lett. 2011, 13, 1170−1173. (b) Wang, T.;
Wang, C.-H.; Zhang, J. Chem. Commun. 2011, 47, 5578−5580.
(c) Chen, Z.; Tian, Z.; Zhang, J.; Ma, J.; Zhang, J. Chem. Eur. J. 2012,
18, 8591−8595.
Ni(ClO4)2-catalyzed path, Tables S1−S3, giving detailed
information on the energy barriers of the water coordinated
and various Lewis acid catalyzed paths, text giving experimental
1
details for Ni(OTf)2-catalyzed cycloaddition reactions and H
NMR spectra, and tables giving Cartesian coordinates for all the
structures in this paper. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
(7) For a general review and discussion, see: (a) Kuznetsov, M. L.
Russ. Chem. Rev. 2006, 75, 1045−1073. (b) Ess, D. H.; Houk, K. N. J.
Am. Chem. Soc. 2008, 130, 10187−10198.
Corresponding Authors
*E-mail for J.Z.: jlzhang@chem.ecnu.edu.cn.
*E-mail for J.M.: majing@nju.edu.cn.
(8) (a) Sustmann, R. Tetrahedron Lett. 1971, 12, 2721−2724.
(b) Sustmann, R. Tetrahedron Lett. 1971, 12, 2717−2720.
(9) For some recent theoretical investigations on regioselective [3 +
2] cycloadditions, see: (a) Li, Z. M.; Wang, Q. R. Int. J. Quantum
Chem. 2008, 108, 1067−1075. (b) Benchouk, W.; Mekelleche, S. M.;
Aurell, M. J.; Domingo, L. R. Tetrahedron 2009, 65, 4644−4651.
(c) Cheng, Y.; Wang, B.; Wang, X. R.; Zhang, J. H.; Fang, D. C. J. Org.
Chem. 2009, 74, 2357−2367. (d) Domingo, L. R.; Chamorro, E.;
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Basic Research
Program (Grant No. 2011CB808600) and the National Natural
Science Foundation of China (Grant Nos. 21290192 and
21273102). We are grateful to the High Performance
Computing Center of Nanjing University for doing the
quantum chemical calculations in this paper on its IBM Blade
cluster system.
Perez, P. Eur. J. Org. Chem. 2009, 3036−3044. (e) Arron
́
iz, C.; Molina,
J.; Abas, S.; Molins, E.; Campanera, J. M.; Luque, F. J.; Escolano, C.
́
Org. Biomol. Chem. 2013, 11, 1640−1649. (f) Rhyman, L.; Abdallah,
H. H.; Jhaumeer-Laulloo, S.; Domingo, L. R.; Joule, J. A.; Ramasami, P.
Tetrahedron 2011, 67, 8383−8391. (g) Bentabed-Ababsa, G.; Hamza-
́
Reguig, S.; Derdour, A.; Domingo, L. R.; Saez, J. A.; Roisnel, T.;
Dorcet, V.; Nassar, E.; Mongin, F. Org. Biomol. Chem. 2012, 10, 8434−
8444.
REFERENCES
■
(10) For several recent examples of mechanistic studies on noble
metal catalyzed [3 + 2] cycloaddition, see: (a) Menezes, F. M. C.;
Kuznetsov, M. L.; Pombeiro, A. J. L. Organometallics 2009, 28, 6593−
6602. (b) Hou, D.-R.; Kuan, T.-C.; Li, Y.-K.; Lee, R.; Huang, K.-W.
Tetrahedron 2010, 66, 9415−9420. (c) Jiao, L.; Lin, M.; Yu, Z. X. J.
(1) For the earliest studies on the [3 + 2] dipolar cycloaddition, see:
(a) Huisgen, R.; Bayer, H. O.; Schaefer, F. C.; Gotthardt, H. Angew.
Chem., Int. Ed. 1964, 3, 136−137. (b) Huisgen, R.; Otto, P.
Tetrahedron Lett. 1968, 4491−4495. (c) Huisgen, R.; Brunn, E.;
Gilardi, R.; Karle, I. J. Am. Chem. Soc. 1969, 91, 7766−7767.
(2) For recent leading reviews on the [3 + 2] dipolar cycloaddition:
(a) Padwa, A.; Pearson, W. H. Synthetic applications of 1,3-dipolar
cycloaddition chemistry toward heterocycles and natural products; Wiley:
Hoboken, NJ, 2002. (b) Gothelf, K. V.; Jorgensen, K. A. Chem. Rev.
1998, 98, 863−909. (c) Fruhauf, H.-W. Coord. Chem. Rev. 2002, 230,
79−96. (d) Kanemasa, S. Synlett 2002, 1371−1387. (e) Pandey, G.;
Banerjee, P.; Gadre, S. R. Chem. Rev. 2006, 106, 4484−4517. (f) Nair,
V.; Suja, T. D. Tetrahedron 2007, 63, 12247−12275. (g) Pellissier, H.
Tetrahedron 2007, 63, 3235−3285. (h) Rubin, M.; Rubina, M.;
Gevorgyan, V. Chem. Rev. 2007, 107, 3117−3179. (i) Patil, N. T.;
Yamamoto, Y. Chem. Rev. 2008, 108, 3395−3442. (j) Carson, C. A.;
Kerr, M. A. Chem. Soc. Rev. 2009, 38, 3051−3060. (k) Bartok, M.
Chem. Rev. 2010, 110, 1663−1705. (l) Xing, Y.; Wang, N.-X. Coord.
Chem. Rev. 2012, 256, 938−952.
Am. Chem. Soc. 2011, 133, 447−461. (d) García-Fandino, R.; Gulías,
̃
́
M.; Mascarenas, J. L.; Cardenas, D. J. Dalton Trans. 2012, 41, 9468−
̃
9481. (e) Liu, T.; Han, L.; Liu, Y.; Zhang, D.; Li, W.-Z. Comput. Theor.
Chem. 2012, 992, 97−102. (f) Liu, Y.; Zhang, D.; Bi, S.; Liu, C. Org.
Biomol. Chem. 2013, 11, 336−343. (g) Rodier, F.; Rajzmann, M.;
Parrain, J. L.; Chouraqui, G.; Commeiras, L. Chem. Eur. J. 2013, 19,
2467−2477.
(11) (a) Domingo, L. R.; Arno, M.; Merino, P.; Tejero, T. Eur. J. Org.
Chem. 2006, 3464−3472. (b) Zhang, J. S.; Shen, W.; Li, M. Eur. J. Org.
Chem. 2007, 4855−4866. (c) Frank, E.; Mucsi, Z.; Zupko, I.; Rethy,
B.; Falkay, G.; Schneider, G.; Wolfling, J. J. Am. Chem. Soc. 2009, 131,
3894−3904. (d) Zamfir, A.; Schenker, S.; Bauer, W.; Clark, T.;
Tsogoeva, S. B. Eur. J. Org. Chem. 2011, 3706−3709. (e) Rhyman, L.;
Ramasami, P.; Joule, J. A.; Sae
́
z, J. A.; Domingo, L. R. RSC Adv. 2013,
́
3, 447−457. (f) Sniezek, M.; Stecko, S.; Panfil, I.; Furman, B.;
̇
(3) For examples of dipolar cycloadditions of aldehydes, see:
(a) Christie, S. D. R.; Davoile, R. J.; Elsegood, M. R. J.; Fryatt, R.;
Jones, R. C. F.; Pritchard, G. J. Chem. Commun. 2004, 2474−2475.
(b) Nair, V.; Mathai, S.; Viji, S.; Mathew, S. Res. Chem. Intermediat.
2006, 32, 1−7. (c) Pohlhaus, P. D.; Sanders, S. D.; Parsons, A. T.; Li,
W.; Johnson, J. S. J. Am. Chem. Soc. 2008, 130, 8642−8650.
(d) DeAngelis, A.; Taylor, M. T.; Fox, J. M. J. Am. Chem. Soc. 2009,
131, 1101−1105. (e) Jiang, Z.; Wang, J.; Lu, P.; Wang, Y. Tetrahedron
2011, 67, 9609−9617. (f) Wu, X.; Li, L.; Zhang, J. Chem. Commun.
2011, 47, 7824−7826.
(4) For examples of dipolar cycloadditions of oxiranes, see:
(a) Bentabed, G.; Derdour, A.; Benhaoua, H. Synth. Commun. 2003,
33, 1861−1866. (b) Madhushaw, R. J.; Li, C. L.; Su, H. L.; Hu, C. C.;
Lush, S. F.; Liu, R. S. J. Org. Chem. 2003, 68, 1872−1877. (c) Huo, C.
D.; Jia, X. D.; Zhang, W.; Yang, L.; Lu, J. M.; Liu, Z. L. Synlett. 2004,
251−254. (d) Hilt, G.; Bolze, P.; Harms, K. Chem. Eur. J. 2007, 13,
4312−4325. (e) Pan, J.; Zhang, W.; Zhang, J.; Lu, S. Tetrahedron Lett.
2007, 48, 2781−2785. (f) Chen, Z. L.; Wei, L.; Zhang, J. L. Org. Lett.
2011, 13, 1170−1173. (g) Hou, X.-L.; Wu, W.-Q.; Ding, C.-H. Synlett.
2012, 23, 1035−1038. (h) Wang, L.; Li, Z.; Lu, L.; Zhang, W.
Tetrahedron 2012, 68, 1483−1491.
́
Urbanczyk-Lipkowska, Z.; Chmielewski, M. Tetrahedron: Asymmetry
2013, 24, 89−103.
(12) (a) Bentabed-Ababsa, G.; Derdour, A.; Roisnel, T.; Saez, J. A.;
Perez, P.; Chamorro, E.; Domingo, L. R.; Mongin, F. J. Org. Chem.
2009, 74, 2120−2133. (b) Lan, Y.; Houk, K. N. J. Am. Chem. Soc.
2010, 132, 17921−17927.
(13) Novikov, R. A.; Balakirev, D. O.; Timofeev, V. P.; Tomilov, Y. V.
Organometallics 2012, 31, 8627−8638.
(14) (a) Shi, F. Q.; Li, X.; Xia, Y.; Zhang, L.; Yu, Z. X. J. Am. Chem.
Soc. 2007, 129, 15503−15512. (b) Liang, Y.; Liu, S.; Xia, Y. Z.; Li, Y.
H.; Yu, Z. X. Chem. Eur. J. 2008, 14, 4361−4373.
(15) 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, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.;
1724
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725

Organometallics
Article
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.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision B.01, Gaussian, Inc.: Wallingford, CT, 2010.
(16) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785−789.
(17) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270−283.
(b) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284−298. (c) Hay,
P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299−310.
(18) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241.
(19) (a) Tamura, H.; Yamazaki, H.; Sato, H.; Sakaki, S. J. Am. Chem.
Soc. 2003, 125, 16114−16126. (b) Sakaki, S.; Takayama, T.;
Sumimoto, M.; Sugimoto, M. J. Am. Chem. Soc. 2004, 126, 3332−
3348.
́
(20) Fuentealba, P.; Preuss, H.; Stoll, H.; Von Szentpaly, L. Chem.
Phys. Lett. 1982, 89, 418−422.
(21) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378−6396.
(22) (a) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold,
F. NBO Version 3.1. (b) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem.
Rev. 1988, 88, 899−926. (c) Reed, A. E.; Weinstock, R. B.; Weinhold,
F. J. Chem. Phys. 1985, 83, 735−746.
(23) The activation energy barriers are calculated at the B3LYP/6-
31+G(d) level. In experiments, the C−C bond can be cleaved
thermally; see: (a) Wang, G. W.; Yang, H. T.; Wu, P.; Miao, C. B.; Xu,
Y. J. Org. Chem. 2006, 71, 4346−4348. (b) Bentabed-Ababsa, G.;
Derdour, A.; Roisnel, T.; Saez, J. A.; Domingo, L. R.; Mongin, F. Org.
Biomol. Chem. 2008, 6, 3144−3157. The C−O bond cleavages always
occur under photochemical conditions.4e,h Thus for the isolated
oxiranes, the C−C bond is weaker than the C−O bond.
(24) Yamamoto, H. Lewis Acids in Organic Synthesis; Wiley: Hoboken,
NJ, 2002.
(25) (a) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533−3543.
(b) Woodward, S. Tetrahedron 2002, 58, 1017−1050. (c) Olah, J.; De
Proft, F.; Veszpremi, T.; Geerlings, P. J. Phys. Chem. A 2005, 109,
1608−1615.
(26) (a) Olah, G. A.; Kobayashi, S.; Tashiro, M. J. Am. Chem. Soc.
1972, 94, 7448−7461. (b) Luo, Y. R.; Benson, S. W. Inorg. Chem.
1991, 30, 1676−1677. (c) Kobayashi, S.; Busujima, T.; Nagayama, S.
Chem. Eur. J. 2000, 6, 3491−3494. (d) Kobayashi, S.; Manabe, K. Acc.
Chem. Res. 2002, 35, 209−217.
(27) Yamamoto, Y. J. Org. Chem. 2007, 72, 7817−7831.
(28) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751−767.
1725
dx.doi.org/10.1021/om500068m | Organometallics 2014, 33, 1715−1725