Mechanism-based design and optimization of a catalytic electrophilic cyclopropanation without diazomethane

Article abstract of DOI:10.1021/acs.organomet.6b00531

Iodomethylboron compounds, either the trifluoroborate or a boronic ester, cyclopropanate electron-rich olefins and unprotected allylic alcohols with Pd catalysts according to a novel, designed catalytic cycle. Proposed intermediates in a "diverted Heck" mechanism are observed by means of spectroscopic studies and by isolation and X-ray crystallographic characterization, which together with reaction kinetics point to a separation of rate-determining and product-determining steps, and a mechanism-based optimization of the yield, selectivity, and scope of the catalytic electrophilic cyclopropanation. The reaction with crystalline, air-stable, nonhygroscopic, and nontoxic reagents provides an alternative to Simmons-Smith-type reactions, as well as cyclopropanation procedures that require the use of diazomethane.

Full text of DOI:10.1021/acs.organomet.6b00531

Article
pubs.acs.org/Organometallics
Mechanism-Based Design and Optimization of a Catalytic
Electrophilic Cyclopropanation without Diazomethane
Augustin A. S. W. Tchawou, Mihai Raducan, and Peter Chen*
Laboratorium fur Organische Chemie, Eidgenossische Technische Hochschule (ETH) Zurich, CH-8093 Zurich, Switzerland
̈
̈
̈
̈
S
* Supporting Information
ABSTRACT: Iodomethylboron compounds, either the trifluoroborate or a boronic
ester, cyclopropanate electron-rich olefins and unprotected allylic alcohols with Pd
catalysts according to a novel, designed catalytic cycle. Proposed intermediates in a
“diverted Heck” mechanism are observed by means of spectroscopic studies and by
isolation and X-ray crystallographic characterization, which together with reaction
kinetics point to a separation of rate-determining and product-determining steps, and
a mechanism-based optimization of the yield, selectivity, and scope of the catalytic
electrophilic cyclopropanation. The reaction with crystalline, air-stable, nonhygro-
scopic, and nontoxic reagents provides an alternative to Simmons−Smith-type
reactions, as well as cyclopropanation procedures that require the use of
diazomethane.
he synthetic preparation of cyclopropanes commonly
Trequires the use of a Simmons−Smith-type cyclo-
propanation1 or the metal-catalyzed decomposition of diazo
compounds.2 Alternative preparative pathways reported for
nucleophilic cyclopropanations include organocatalysis,3 the
Michael initiated ring closure (MIRC) reaction,4 or palladium
catalysis5 to produce cyclopropanes. These approaches,
however, are mostly limited to electron-deficient or polar
olefins. The necessity of developing a safer and catalytic
electrophilic strategy for the cyclopropanation of electron-rich
alkenes therefore exists.
(Scheme 1, eq 1). Using iodomethylboronic acid pinacol
ester (ICH2Bpin) instead of ICH2BF3K together with cesium
fluoride under the same conditions led to the formation of the
cyclopropane in 92% yield (Scheme 1, eq 2).
In that original work, other electron-rich alkenes, such as
cyclooctene (COE), showed low conversion and poor
Scheme 1. Previously Reported Pd-Catalyzed Methylenation
of NBE using Iodomethylboron Species
Palladium-catalyzed carbon−carbon bond formation via
cross-coupling chemistry has shown an enormous synthetic
potential in organic synthesis and has become widely used in
research, in both academia and industry.6 The mechanisms
have been extensively investigated over the past few decades.
Key elementary steps common to the major palladium-
catalyzed carbon−carbon bond-forming reactions include
oxidative addition, transmetalation, and reductive elimination.
The coupling partner associated with the Mizoroki−Heck-type7
reaction is an alkene, whereas for the other major coupling
methods (e.g., Suzuki,8 Negishi,9 Sonogashira,10 Hiyama,11
Stille,12 etc.) an organometallic reagent is utilized.
Combining the principles featured in these palladium-based
cross-coupling reactions, we recently disclosed a catalytic route
for the cyclopropanation of norbornene (NBE) using a
commercially available iodomethylboron species as the
methylene transfer agent.13 We found that in DMF/H2O in a
8/1 ratio at 90 °C, the Herrmann−Beller catalyst,14 widely used
in the Heck reaction, was particularly efficient in converting
norbornene (NBE) into the corresponding tricyclo(3.2.1.02,4)-
octane in the presence of potassium iodomethyltrifluoroborate
(ICH2BF3K), with potassium carbonate acting as a base
Special Issue: Hydrocarbon Chemistry: Activation and Beyond
Received: July 2, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
stirred for 10 min before and during the addition of a solution of
anhydrous dimethylacetamide (DMA, 0.6 mL) containing the Pd
catalyst [(o-tolyl)3P]2PdCl2 (4.6 mg, 5 mol %), ICH2BF3K (34.8 mg,
1.2 equiv), and cyclooctene (COE, 15.2 μL, 0.117 mmol). The rest of
the anhydrous DMA (2 × 0.5 mL) was utilized to rinse the vial
containing the previous solution and then added to the reaction
mixture. The tube was sealed, removed from the glovebox, and heated
to 75 °C for 22 h. The reaction mixture was then rapidly cooled to 0
°C in an ice bath for 15 min, and a workup was conducted.
selectivity for the desired cyclopropanation products over what
we identified to be β-hydride elimination products (Scheme 2).
Major questions concerning the mechanism by which the
reaction is taking place remained open.
Scheme 2. Previously Reported Pd-Catalyzed Methylenation
of COE using Iodomethyltrifluoroborate
Representative Methylenation Procedure in DMA/MeOH. A
5 mL oven-dried Schlenk (2.5 or 5 mL) equipped with a Teflon J.
Young valve and an oven-dried magnetic stir bar was charged in a N2-
filled glovebox with powdered K2CO3 (80.9 mg, 5.0 equiv) and
anhydrous MeOH (0.5 mL). The mixture was vigorously stirred for 10
min before, and during, the addition of a solution of anhydrous DMA
(0.6 mL) containing the Pd catalyst [(o-tolyl)3P]2PdCl2 (4.6 mg, 5
mol %), ICH2BF3K (34.8 mg, 1.2 equiv) and COE (15.2 μL, 0.117
mmol). The rest of the anhydrous DMA (2 × 0.5 mL) and anhydrous
MeOH (0.3 mL) was utilized to rinse the vial containing the previous
solution and then added to the reaction mixture. The tube was sealed,
removed from the glovebox, and heated to 75 °C for 22 h. The
reaction mixture was then rapidly cooled to 0 °C in an ice bath for 15
min, and a workup was conducted.
Representative Kinetic Experiment. Each time point in a kinetic
run was acquired with the following procedure. An oven-dried Schlenk
(2.5 or 5 mL) equipped with a Teflon Young valve and an oven-dried
magnetic stir bar was charged in a N2-filled glovebox with powdered
K2CO3 (80.9 mg, 5.0 equiv) and anhydrous EG (0.1 mL). The mixture
was vigorously stirred for 10 min before, and during, the addition of a
solution of anhydrous DMA (0.6 mL) containing either the Pd catalyst
[(o-tolyl)3P]2PdCl2 (4.6 mg, 5 mol %) or a mixture of (o-tolyl)3P
(1.78 mg, 5 mol %) and the Pd catalyst [COD]PdCl2 (1.67 mg, 5 mol
%), ICH2BF3K (34.8 mg, 1.2 equiv), and COE (15.2 μL, 0.117 mmol).
The rest of the anhydrous DMA (2 × 0.5 mL) was utilized to rinse the
vial containing the previous solution and then added to the reaction
mixture. The tube was sealed, removed from the glovebox, and heated
to 75 °C for the indicated reaction time. Then the reaction mixture
was rapidly cooled to 0 °C in an ice bath for 15 min and a workup was
conducted. The resulting solution was analyzed by GC-FID.
Procedure for Small-Scale Workup. The internal standard,
undecane (10 μL, around 0.40 equiv for 0.117 mmol scale of olefin),
was added to the tube with the reaction mixture, followed by Et2O (2
× 35 mL/mmol substrate). The solution was transferred to a 15 mL
glass vial. The Schlenk tube was rinsed with H2O (45 mL/mmol
substrate) or saturated NH4Cl (for the allylic alcohols). The aqueous
layer was transferred to the 15 mL glass vial, and the layers were
separated. The organic phase was transferred to another 15 mL glass
vial, washed with H2O (45 mL/mmol substrate) and brine (45 mL/
mmol substrate), dried over MgSO4, and then filtered through Celite.
The resulting solution was analyzed by GC-FID.
Representative Procedure for the Scale-Up of the Methyle-
nation Reaction of Olefins to Millimole (Gram-Scale) Amounts.
A 250 mL oven-dried round-bottom flask equipped with an oven-dried
magnetic stir bar was charged in a N2-filled glovebox with powdered
K2CO3 (4.04 g, 5.0 equiv), anhydrous EG (5 mL), and anhydrous
DMA (20 mL). The resulting mixture was stirred vigorously. [(o-
tolyl)3P]2PdCl2 (230 mg, 5 mol %), (o-tolyl)3P (89 mg, 5 mol %),
ICH2BF3K (2.18 g, 1.5 equiv), and cyclooctene (645 mg, 5.85 mmol)
were added to the mixture with stirring. A 60 mL portion of DMA was
used to rinse the vials utilized to weigh all the reagents and was
subsequently added to the reaction mixture. The round-bottom flask
was then closed and removed from the glovebox, and the contents
were stirred at 75 °C for the indicated time. At the end of the reaction,
which was monitored by GC/MS, the flask was cooled to 0 °C and its
contents were poured over Celite. The filter cake was washed with
Et2O (3 × 50 mL). The filtrate was transferred to a separatory funnel,
and H2O (200 mL) was added. The organic layer was separated, and
the aqueous layer was extracted with Et2O (2 × 100 mL). The
fractions were combined, dried over MgSO4, and filtered, and the ether
solution was removed by a rotary evaporator. The resulting oil residue
We report here experimental evidence for the “diverted
Heck” mechanism proposed in our earlier work and the
mechanism-based improvement in yield and substrate scope for
the palladium-catalyzed cyclopropanation of electron-rich
olefins, including strained and unstrained substrates, sterically
hindered olefins, and allylic alcohols.
EXPERIMENTAL SECTION
■
General Remarks. Unless otherwise stated, all reactions were
carried out under an argon or N2 atmosphere using standard Schlenk
or glovebox techniques with anhydrous solvents. Solvents and reagents
were acquired from Aldrich, Fluka, Acros Organics, ABCR Chemicals,
and TCI. Deuterated solvents were obtained from Amar Chemicals.
Extra molecular sieves were added to the purchased solvents and the
solvents stored in the glovebox. ICH2BF3K (>98.0%) was acquired
from TCI, further purified by washing several times with Et2O, dried
under high vacuum, and then stored in the glovebox. NMR data were
recorded at 300, 400, or 600 MHz (Bruker AV300, AV400, and AV600
spectrometers) with CDCl3, CD3OD, THF-d8, or DMF-d7 as solvent.
Chemical shifts (δ) are reported in ppm with the residual solvent
1
signal as internal standard (chloroform at 7.26 and 77.00 ppm for H
and 13C spectroscopy, respectively; methanol at 3.31 and 49.00 ppm
for 1H and 13C spectroscopy, respectively; tetrahydrofuran at 3.58 and
1
67.57 ppm for H and 13C spectroscopy, respectively; dimethylforma-
mide at 2.75, 2.92, and 8.03 ppm for H spectroscopy). 13C NMR
1
spectra were recorded with complete 1H decoupling. Service
measurements were performed by the NMR service team of the
Laboratorium fur Organische Chemie, ETH Zurich, by Mr. Rainer
̈
̈
Frankenstein, Mr. Rene
́
Arnold, and Mr. Stephan Burkhardt under the
supervision of Dr. Marc-Oliver Ebert. X-ray structural analysis was
performed by the X-ray crystallography group of the Laboratorium fur
Organische Chemie, ETH Zurich. High-resolution mass spectra were
recorded by the mass spectrometry service of the ETH Zurich
Laboratorium fur Organische Chemie with a Micromass (Waters)
̈
̈
̈
̈
AutoSpec Ultima (EI) or a Bruker solariX (MALDI-FTICR)
instrument. GC-FID analysis was performed on a Finnigan Focus
GC instrument with a Zebron ZB-5MS 30 m × 0.25 mm column or a
Restek Rtx-Wax 30 m × 0.25 mm column using a flame ionization
detector. Qualitative gas-phase studies were performed on a Thermo
Finnigan TSQ Quantum ESI-MS/MS instrument. Charged species
were generated by electrospray ionization (ESI) from dilute (∼10−5
mol/L) solutions freshly prepared in the glovebox, taken into a
Hamilton gastight syringe, and used immediately. All samples were
electrosprayed with a flow rate of 5 mL/min.
Representative Methylenation Procedure in DMA/EG or
THF/EG. An oven-dried Schlenk (2.5 or 5 mL) equipped with a Teflon
J. Young valve and an oven-dried magnetic stir bar was charged in a
N2-filled glovebox with powdered K2CO3 (80.9 mg, 5.0 equiv) and
anhydrous ethylene glycol (EG, 0.1 mL). The mixture was vigorously
B
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
3
was purified by flash column chromatography over silica gel using
pentane as eluent. Slow evaporation of the solvent yielded 654 mg
(90%) of a mixture of bicyclo[6.1.0]nonane (2a) and methylenecy-
clooctane (3a) (50:1 ratio of products in favor of the cyclopropane) as
a colorless oil. MS (EI): 124 (32, [M]+), 96 (94), 81 (77), 67 (100),
54 (73), 41 (27). HRMS (EI, ([M]+)): m/z calcd for C9H16, 124.1247;
found, 124.1247. See the Supporting Information for further details
related to each isolated cyclopropane.
6H, iPr CH3), 1.48 (d, 3JHP = 15.1 Hz, 18H, tBu), 1.36 (d, JHH = 6.9
3
3
Hz, 6H, iPr CH3), 1.19 (qd, JHF = 8.0 Hz and JHP = 1.6 Hz, 2H,
PdCH2BF3), 0.95 (d, JHH = 6.6 Hz, 6H, iPr CH3). 31P NMR (162
3
MHz, CH2Cl2-d2): δ 75.42 (q, JPF = 32.6 Hz). 19F NMR (376 MHz,
4
CH2Cl2-d2): δ −140.53 (s, PdCH2BF3). 11B NMR (160 MHz,
CH2Cl2-d2): δ 5.38 (s, PdCH2BF3). HRMS (MALDI-FTICR/DCTB,
([M-BF3]+)): m/z calcd for C30H47PPd, 544.2456; found, 544.2456.
All spectra are available in the Supporting Information.
Control Experiments. The methylenation procedures were
followed, and only the reagents needed for each experiment were
charged in the oven-dried Schlenk tube. The tube was sealed, removed
from the glovebox, and heated to 75 °C for 16 h. At the end of the
indicated time, the Schlenk tube was cooled to 0 °C for 15 min, dried,
and transferred to the glovebox. A 0.6 mL portion of the solution was
filtered and transferred into an oven-dried NMR tube placed in an
external NMR tube containing the reference solvent DMF-d7. The
tubes were closed, and NMR was taken at room temperature.
Gas-Phase Studies. In the glovebox, 0.1 mL of the reaction
mixture was diluted to ∼10−5 M in dry MeCN or dry THF and used
immediately for gas-phase studies. Conditions: spray voltage 5 kV,
collision offset 35 V, tube lens offset 100 V, and capillary temperature
50 or 170 °C. Collision-induced dissociation (CID) experiments were
done with an argon pressure of 0.5 mTorr in the collision cell.
Sample Preparation for NMR Methylenation Reaction
Monitoring. An oven-dried NMR tube equipped with a Teflon
screw cap was transferred into a N2-filled glovebox and charged with
powdered K2CO3 (0.195 mmol, 5.0 equiv). A 1.5 mL oven-dried vial
was charged with [(o-tolyl)3P]2PdCl2 (1.53 mg, 5 mol %), ICH2BF3K
(9.67 mg, 1.2 equiv), COE (5.1 μL, 0.039 mmol), anhydrous EG-d6
(0.33 mL), and anhydrous DMA-d9 (2.33 mL). The solution was then
transferred to the NMR tube and the vial rinsed with anhydrous DMA-
d9 (2 × 1.5 mL). The rinse solution was placed in the NMR tube. The
NMR tube was sealed, shaken, and maintained at 75 °C for 16 h.
Oxidative Addition of ICH2Bpin in THF. An oven-dried NMR
tube was brought into a N2-filled glovebox and charged with Pd(dba)2
(30 mg, 0.052 mmol, 1.05 equiv), (o-tolyl)3P (15 mg, 0.049 mmol, 1.0
equiv), and anhydrous THF-d8 (0.4 mL). The tube was then closed
and agitated for ∼5 min. A 1.5 mL oven-dried vial was charged with
ICH2Bpin (33 mg, 0.123 mmol, 2.5 equiv) and anhydrous THF-d8
(0.2 mL). The solution was then transferred to the NMR tube. The
tube was closed with a cap and agitated for 15 min, and the NMR
spectrum was taken at room temperature. The dark red solution
turned yellow with deposition of a greenish solid.
Oxidative Addition of ICH2Bpin in the Presence of NBE in
THF. An oven-dried NMR tube was transferred into a N2-filled
glovebox and charged with Pd(dba)2 (30 mg, 0.052 mmol, 1.05 equiv),
(o-tolyl)3P (15 mg, 0.049 mmol, 1.0 equiv), and anhydrous THF-d8
(0.4 mL). The tube was closed and agitated for ∼5 min. NBE (23.07
mg, 0.245 mmol, 5 equiv) in anhydrous THF (0.2 mL) was added, and
an NMR of the resulting orange solution was taken 30 min later. The
tube was then transferred back to the glovebox and charged with
ICH2Bpin (33 mg, 0.123 mmol, 2.5 equiv). Another NMR spectrum
was taken, and the mixture was heated to 60 °C for 10 min and then
cooled to room temperature, diluted, and used immediately for gas-
phase studies.
RESULTS
■
As will be detailed in the Discussion, the expansion of the
substrate scope and improvement of product selectivity were
based on mechanistic and kinetic studies, proceeding in tandem
with the preparative work. The preparative results are shown
first, though, to indicate the degree to which the optimization
succeeded. The experimental results pertinent to the reaction
mechanism follow in the text, although the two processes
actually ran concurrently.
Reaction Optimization. Cyclooctene (COE) was chosen
as a model substrate for the development of a generalized
procedure for the methylenation of electron-rich alkenes
because, as we had shown in the previously published work
(Scheme 2), COE gave only modest yields of both cyclo-
propanation and side products. Accordingly, as a test substrate,
there was ample room for improvement in yield and selectivity.
Table 1 summarizes the results of the optimization of the
cyclopropanation of COE. [(o-tolyl)3P]2PdCl2 (Cat1) proved
to be the most active catalyst in the conversion of COE among
the catalysts we tested. At 75 °C, it outperformed the
previously utilized Herrmann−Beller catalyst (Cat2) when
used in the presence of 1.2 equiv of the methylenating agent
ICH2BF3K and 3.0 equiv of the base K2CO3 in (DMA/MeOH,
2/1) and without an additional free phosphine ligand (entries 1
and 2). Under these conditions, the expected bicyclo[6.1.0]-
nonane (2a) product was prepared in 74% yield, a satisfactory
mass balance being obtained, with the balance of the material
being accounted for by the side product methylenecyclooctane
(3a) and recovered starting material. Increasing the amount of
the base to 5 equiv significantly improved the selectivity (entry
3). Interestingly, substituting CH3OH with CD3OD appeared
to give even better results (entry 4), although it is not clear if
the putative isotope effect exceeds the range of variation in the
experiment.
Oxidative Addition of ICH2BF3K in DMF. An oven-dried NMR
tube was brought into a N2-filled glovebox and charged with the Pd
complex t-BuXPhos Pd G3 (4.8 mg, 6.0 μmol, 1.0 equiv), ICH2BF3K
(3.0 mg, 1.2 μmol, 2.0 equiv), and powdered K2CO3 (1.7 mg, 1.2
μmol, 2.0 equiv), successively dissolved in anhydrous DMF-d7 (0.6
mL). The tube was then closed and agitated for ∼5 min, and the NMR
spectrum was taken at room temperature. The mixture was diluted and
immediately used for gas-phase studies.
Preparation of the Pd Complex 8. A 25 mL oven-dried round-
bottom flask was brought into a N2-filled glovebox and charged with
K2CO3 (136 mg, 4.0 equiv), ICH2BF3K (183.4 mg, 3.0 equiv), and
DMF (1 mL). The Buchwald palladium precatalyst tBuXPhos Pd G3
(196 mg, 0.246 mmol, 1.0 equiv) in DMF (1 mL) was added to this
mixture with stirring at ambient temperature, and the vial containing
the precatalyst was rinsed with another 1 mL of DMF that was then
transferred to the reaction mixture. Stirring was continued for 10 min
before removal of the solvent by rotary evaporation. A DCM/pentane
(5 mL/10 mL) mixture was added to the resulting residue, and this
mixture was filtered over Celite. The Celite filter cake was washed with
pentane (5 mL). The solvents were slowly evaporated on a rotary
evaporator under moderate vacuum (no water bath) until a great deal
of clear green microcrystals appeared in the flask. Evaporation was
stopped, and the crystals were isolated by filtration, washed with a
small volume of cold pentane, and dried under high vacuum. These
crystals were suitable for X-ray analysis. A 126 mg amount of product
The yield of 2a was further improved to 94% by replacing the
former solvent mixture with DMA/EG in a ratio of 16/1 (entry
5). Lowering (8/1) or increasing (32/1) this ratio led to
decreased product yields of 84% and 87%, respectively (entries
6 and 7). No products were observed when pinacol was used
instead of EG (entry 8). The use of glycerol gave 71% of 2a
(entry 9). Substituting THF for DMA led to a reduced
1
was obtained (NMR reveals the presence of 9H-carbazole). H NMR
3
4
(400 MHz, CH2Cl2-d2): δ 7.99 (tq, JHH = 6.5 Hz and JHP = 2.2 Hz,
1H, aryl CH), 7.57−7.51 (m, 2H, aryl CH), 7.28−7.27 (m, 2H, aryl
CH), 7.19−7.11 (m, 1H, aryl CH), 3.02 (hept, 3JHH = 6.9 Hz, 1H, iPr
CH), 2.45 (hept, 3JHH = 6.7 Hz, 2H, iPr CH), 1.56 (d, 3JHH = 6.8 Hz,
C
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
leading to the main products which were determined to have
the same stereochemical configuration as those produced by a
control Simmons−Smith-type cyclopropanation. The products
were obtained in moderate to excellent yields, an exception
being camphene (1l), for which only the starting material was
recovered at the end of the reaction (Scheme 3). In general, the
Table 1. Optimization Experiments for the Pd-Catalyzed
Methylenation Reaction of COE (1)
a
Scheme 3. Pd-Catalyzed Methylenation Reaction: Scope of
ab
,
the Alkenes and Allylic Alcohols
variation from “standard
conditions”
2
(%)
3
(%)
b
b
no.
solvent (ratio)
cd
,
1
2
3
4
Cat2 (5 mol %)
3 equiv of K2CO3
none
DMA/MeOH
(2/1)
63.6
74.0
76.5
82.9
0.8
1.1
0.5
0.6
DMA/MeOH
(2/1)
DMA/MeOH
(2/1)
none
DMA/CD3OD
(2/1)
c
c
c
5
6
7
8
9
none
DMA/EG (16/1)
DMA/EG (8/1)
DMA/EG (32/1)
DMA
93.6
83.8
86.5
0.3
2.3
3.2
2.3
0.9
5.3
0.4
1.7
0
none
none
pinacol (15 equiv)
glycerol (10 equiv)
none
DMA
71.9
12.9
81.7
0
10
11
12
13
14
15
16
17
18
19
THF/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
DMA/EG (16/1)
ICH2BF3K (1.0 equiv)
BrCH2BF3K
e
in situ made catalyst
85.5
71.2
86.1
53.7
70.1
90.4
96.2
2.1
1.7
2.1
1.2
2.0
2.5
2.6
Cat1 (1 mol %)
Cat1 (1 mol %)
Cat3
f
Cat4
65 °C
65 °C for 36 h
a
Standard conditions: 0.117 mmol (15.2 μL, 1.0 equiv) of alkene 1a,
ICH2BF3K (34.8 mg, 1.2 equiv), K2CO3 (80.9 mg, 5.0 equiv), Cat1
(4.6 mg, 5 mol %), 75 °C, 22 h.
b
c
d
Yields were evaluated by GC-FID. ∼0.069 M solution. In the
e
presence of extra L1 (15 mol %). Mixture of L1 (10 mol %) and
[COD]PdCl2 (5 mol %). In the presence of extra L1 (1 mol %).
f
conversion of COE (entry 10). The yield also dropped to 81%
when exactly 1.0 equiv of ICH2BF3K was utilized (entry 11).
Use of BrCH2BF3K as the methylenating agent resulted in a full
recovery of the starting material (entry 12).
a
Conditions: 0.117 mmol (1.0 equiv) of alkene 1, ICH2BF3K (34.8
mg, 1.2 equiv), K2CO3 (80.9 mg, 5.0 equiv), Cat1 (4.6 mg, 5 mol %),
b
solvents DMA/EG (1.6 mL/0.1 mL, 16/1). Yields were evaluated by
GC-FID. 1.5 equiv of ICH2BF3K used. Isolated yields are given in
parentheses (5.85 mmol of alkene 1 used). In the presence of extra
L1 (5 mol %). Commercial mixture of cis and trans (2:1 trans:cis).
c
d
In accordance with previous reports, the in situ generated
catalyst gave a lower yield (entry 13). Reducing the amount of
catalyst to 1 mol % decreased the yield of the cyclopropane to
71% (entry 14). The addition of 1 mol % extra ligand to the
latter improved the catalytic coupling (entry 15). The use of
the bulky and electron-rich Cat3 resulted in only a moderate
yield (entry 16), which was slightly improved with its bidentate
ferrocenyl analogue Cat4 (entry 17). Interestingly, lowering the
temperature to 65 °C for an extended reaction time (36 h)
slightly increased the yield of formation of the product 2a to
96% (entry 19).
e
f
mass balance in these cyclopropanations was good, with the
material not accounted for as cyclopropane consisting of the
starting materials and/or β-hydride elimination products. GC-
FID yields with detector calibration using independently
synthesized reference compounds (see the Supporting
Information) are given; selected substrates were cyclopropa-
nated on the millimole, i.e. gram, scale, the isolated yields for
which are given in parentheses in Scheme 3. Under the given
reaction conditions, α-pinene, isolated in only 12% yield
following the Simmons−Smith procedure,15 gave up to 64%
Substrate Scope and Limitations. Having optimized the
reaction conditions, the scope of this transformation was
explored with various unactivated olefins and allylic alcohols,
D
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
GC yield of the cyclopropane 2d. It is noteworthy that only one
stereoisomer was obtained in this case, which was confirmed to
be the diastereomer 2d through nuclear Overhauser effect
spectroscopy (NOESY) experiments (Figure SI-20 in the
Supporting Information). Although the reaction is stereo-
specific, a significant difference in selectivity and reactivity is
observed between (E)- and (Z)-2-hexenol (73:1 dr and 22:1 dr,
respectively). trans-Cyclododecene (1k; 2:1 trans:cis in the
commercial mixture used) is mostly converted into the
cyclopropane 2k. The tricyclic ketone 2m was prepared in
good yield as a 7:1 mixture of its two isomers. The less
abundant isomer corresponds to that having its cyclopropyl
protons close to the carbonyl, leading to a downfield shift of δ
0.62 ppm. The cyclopropanation reaction of oxabenzonorbor-
nadiene (1n), which does not cyclopropanate under Simmons−
Smith conditions at all,16 led specifically to the exoisomer 2n in
only 29% GC yield despite the complete consumption of the
starting alkene.
reservoir (vide infra) has to be cleaved by an olefin and only
then can the catalyst decompose and (2) the methylenating
agent ICH2BF3K slowly decomposes, or converts into a less
reactive reagent, under the reaction conditions when no olefin
is available into which insertion can occur. This should
represent a limitation of the present reaction for less reactive
olefins. Finally, more reactive olefins, for which the cleavage of
the dimeric reservoir species is more efficient, e.g. trans-5-
decene, do show formation of palladium black, which signals
that the reaction has ended with catalyst decomposition. For
the unprotected allyl alcohols, an aldehydic oxidation (side)
product was identified in the case of myrtenol (1f; see the
Supporting Information).
Spectroscopic Studies. For a better understanding of the
reaction, relevant reaction intermediates and/or products
needed to be identified or isolated. Therefore, the reaction of
COE with 1.2 equiv of ICH2BF3K in the presence of 5 equiv of
K2CO3 catalyzed by 5 mol % of Cat1 in DMA-d9/EG-d6 was
1
Methylenation Reaction of 5-Decene. A discrimination in
the methylenation reaction of the isomers of 5-decene is
observed (Scheme 4). Under the same reaction conditions,
conducted in an NMR tube and was monitored by H and 11
B
spectroscopy at 75 °C for 16 h (Figures SI-27 and SI-28 in the
Supporting Information and Figure 1). It has to be noted that
a
Scheme 4. Methylenation Reaction of 5-Decene
a
Conditions: 0.117 mmol (1.0 equiv) of alkene, ICH2BF3K (1.2
equiv), K2CO3 (5.0 equiv), Cat1 (5 mol %), L1 (5 mol %), solvents:
DMA/EG (16/1), 75 °C, 24 h.
Figure 1. 11B NMR spectrum of the methylenation reaction of COE
recorded at 75 °C and isotope pattern of the bis(ethylene
glycolato)borate anion in the ESI-MS.
trans-5-decene (1s) and cis-5-decene (1t) were converted into
the corresponding cyclopropanes in 75% and 13% yields,
respectively. Furthermore, the cyclopropane 2s was formed in
62% yield and >99:1 dr with an observed deposit of metallic
black palladium, while 2t was obtained in 9% yield and 12:1 dr
with no similar catalyst aggregation and deposition observed.
Methylenation Reaction of a Linear Terminal Alkene. The
cyclopropanation reaction of 1-octene (1o) presented in Table
SI-9 in the Supporting Information highlights the strong
influence of the β-hydride elimination on the catalytic outcome,
as multiple consecutive methylene-inserted products, both
branched and linear, are observed alongside the corresponding
cyclopropanes.
Limitation on Yields. Control experiments were performed
to ascertain the cause of lower cyclopropanation yield for some
of the substrates. The low yield obtained with α-pinene,
cyclohexene, and cis-5-decene, particularly, arises because the
reaction is simply slow with these substrates. No deposition of
palladium black was present after 22 h at 75 °C. With α-pinene,
an additional 1 equiv of ICH2BF3K added to the reaction after
18 h at 75 °C resulted in an increase of cyclopropanation yield
to 74% 24 h later (against 55% after 22 h at the same
temperature; see Table SI-5 in the Supporting Information,
entry 3), indicating that the catalyst remained active. Second, a
control reaction under otherwise “standard conditions”, but
with no olefinic substrate (D-3 in the Supporting Information),
gave no palladium black after 16 h at 75 °C. Subsequent
addition of 1 equiv of cyclooctene into that flask gave a 53%
yield of the cyclopropane after 24 h (against 94% yield at the
same temperature and time when cyclooctene is added from
the beginning). This experiment shows that (1) the dimeric
this NMR study was hampered by the poor solubility of K2CO3
due to the lack of stirring in the tube. Consequently, the
cyclopropane was produced in only 60% yield. 31P NMR
spectra were also taken at the beginning and the end of the
reaction. No phosphorus signal was observed at the onset of the
reaction because of the low solubility of the catalyst. At the end
of the reaction, the 31P NMR spectrum contained only one
singlet at δ −30.4 ppm corresponding to the free tri-o-
tolylphosphine ligand L1. Similarly, the 11B NMR spectrum
contained only two major peaks, one quartet at δ 2.6 ppm and
one broad singlet at δ 9.3 ppm. The quartet signal corresponds
to the starting ICH2BF3K. The broad singlet is assigned to the
potassium bis(ethylene glycolato)borate having bound EG
exchanging in solution with free EG and/or fluoride. Of note is
the absence of a signal in the region of the spectrum where
signals of trivalent boron should be present (δ 20−40 ppm;17
Figure 1). The species observed in NMR were confirmed by
ESI-MS experiments showing a peak in the spectrum at m/z
−131 (Figure 1). Identical observations were made when an
independently prepared synthetic standard was used.
Similar peaks were observed in the 11B NMR spectrum of a
control reaction consisting of a mixture of 10 mol % of Cat1,
1.2 equiv of ICH2BF3K and 5 equiv of K2CO3 in DMA/EG (2/
1) at 75 °C for 16 h in the fume hood (Figure SI-43 in the
Supporting Information). No traces of palladium black were
observed in the final reaction mixture under those conditions.
The 31P NMR spectrum in Figure SI-42 in the Supporting
Information shows the signal of the free L1 ligand (δ −30.4
ppm) and two small broad peaks at δ 35.6 and 40.4 ppm which
E
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
may correspond to cis and trans isomers of a bridged, dinuclear
Pd(II) species (vide infra).
ESI-MS analysis of a reaction mixture, prepared analogously
to the NMR samples, allowed the detection, in the positive
mode, of a peak at m/z 595 that is assigned as the cationic
palladium intermediate IM1, whose collision-induced dissoci-
ation (CID) is shown in Figure 2.
Figure 2. CID on the cation at m/z 595. Collision gas pressure: 0.5
mTorr, Collision energy: 20 V (center-of-mass reference frame). The
structure in the inset represents the composition and makes no claim
about whether the EG is bound monodentate or bidentate. The
otherwise neutral Pd species are observed by MS as their K+ adducts.
Figure 3. (A) Full scan in the positive mode showing products of
oxidative addition and substitution by MeCN. (B) CID on the cation
at m/z 592. Collision gas pressure: 0.5 mTorr, Collision energy: 10 V
(center-of-mass reference frame).
Direct Oxidative Addition to Pd(0). Oxidative addition of
ICH2Bpin occurs at room temperature in THF-d8 to the Pd(0)
species, formed in situ from 1.05 equiv of Pd(dba)2 and 1 equiv
of L1, to generate two major phosphine-containing species,
observed in the 31P NMR spectrum shown in Figure SI-29 in
the Supporting Information. The two broad singlets appeared
at δ 32.2 and 31.2 ppm. These peaks are most likely signals of
phosphine ligands in the cis and trans isomers of bridged,
dinuclear Pd species which would be formed after oxidative
addition of Pd(0) to the C−I bond of ICH2Bpin. Analysis of
the NMR tube content using ESI-MS, sprayed after dilution
with acetonitrile, permitted the observation of undesired
products (m/z 319 and 431) as well as intermediate IM2
(m/z 592; Figure 3) as the main palladium-containing species
in the positive mode. Acetonitrile cleaves the presumed
bridged, dinuclear Pd species to generate the observed
mononuclear cationic palladium species. When the reaction
was repeated in the presence of excess NBE, a new peak
corresponding predominantly to the NBE-coordinated palla-
dium species IM3 was observed at m/z 645 after 10 min
(Figure 4). Nevertheless, a small peak was observed in the CID
spectrum at m/z 537, which most likely corresponds to the
release of tricyclo(3.2.1.02,4)octane from IM4, a β-hydride
elimination being disfavored for NBE.
The direct oxidative addition of ICH2BF3K to a “(o-
tolyl)3PPd0” intermediate at ambient temperature was not
observed by NMR or ESI-MS, the reaction evidently requiring
the higher temperatures specified in the preparative runs. In
contrast, spectroscopic observations (NMR and ESI-MS)
clearly show that a direct oxidative addition of ICH2BF3K to
an in situ generated Pd(0) complex stabilized by the ligand t-
BuXPhos in DMF-d7 does occur at room temperature. A
quartet with a coupling constant of 32.1 Hz at δ 75.5 ppm was
observed in the proton-decoupled 31P NMR spectrum. The
observed quartet should arise from the fluorine−phosphorus
coupling. The resulting palladium complex 8 was stable enough
to be isolated and characterized by X-ray single-crystal
Figure 4. CID on the cation at m/z 645. Collision gas pressure: 0.5
mTorr, Collision energy: 25 V (center-of-mass reference frame).
diffraction. The X-ray structure of 8 depicted in Figure 5, the
first-formed intermediate in the proposed catalytic cycle,
confirms the presence of a coordination to the Pd metal of
one of the fluorides (2.1833 Å) attached to boron. A markedly
small B1−C1−Pd1 angle (93.17°) is therefore observed.
Furthermore, the additional C2−Pd (2.400 Å) interaction
leads to a Pd1−F1 bond sitting on top of the triisopropylben-
zene group participating in the stabilization of the whole
structure.
Kinetic Studies. To assess the rate law of the reaction,
kinetic investigations were conducted. Using the “different
excess” protocol of reaction progress kinetic analysis (RPKA)
developed by Blackmond,18 the data obtained from the given
methylenation reaction conditions and for two different known
concentrations of the reagent of interest were analyzed. The
consumption of COE was followed over time, and yields were
obtained by GC-FID. The results are reported in the
Supporting Information. The overall reaction displayed zero-
order kinetics with respect to ICH2BF3K and COE. Reaction
orders of 0.5 and −0.5 with respect to Pd and EG, respectively,
were also observed (see Figure 6 and the Supporting
F
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
species, the resting state in the catalytic cycle, to a kinetically
competent monomeric species.
DISCUSSION
■
The nucleophilic cyclopropanation of electron-deficient or
polar olefins is widely documented in the literature.4 Among
the reported approaches, the most important preparative
method reported is the MIRC, in which α,β-unsaturated
carbonyl, sulfonium, or nitrogen ylides react with styrenes,19
stilbenes, or a Michael acceptor. For the electrophilic
cyclopropanation of electron-rich olefins, the options are
decidedly more limited,20 the leading methods being zinc
carbenoid mediated cyclopropanations and metal-catalyzed
cyclopropanation with diazo compounds. We had originally
reported the Pd-catalyzed cyclopropanation of norbornene with
ICH2BF3K. We had further presented a heuristic classification
of possible mechanisms for the catalytic electrophilic cyclo-
propanation and proposed that a “diverted Heck” (Scheme 5,
Figure 5. ORTEP diagram of C30H47BF3PPd (8) with 50% probability
thermal ellipsoids. Hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (deg) of 8: Pd1−P1 = 2.2302(8); Pd1−
C1 = 2.045(3); Pd1−F1 = 2.1833(17); Pd1−C2 = 2.400(3); B1−C1−
Pd1 = 93.17(19); F1−Pd1−P1 = 167.56(5).
Scheme 5. Proposed “Diverted Heck” Mechanism for the
Cyclopropanation of Olefins
mechanism V in the heuristic13) was most probable, on the
basis of the observation of terminal methylene side products for
other substrates. Our original work, while mechanistically novel,
suffered from the practical disadvantage of a limited substrate
scope. Having a working hypothesis for the mechanism of
cyclopropanation, the immediate goals of the study were
• confirmation of the proposed mechanism by observation
of putative intermediates
Figure 6. (A) Initial reaction rates for the conversion of COE over
time in the presence of 5 mol % of Cat1 (•) and 10 mol % of Cat1
(red •). (B) Initial reaction rates for the conversion of COE over time
in the presence of EG (0.1 mL, 0.69 M, black •) and EG (0.05 mL,
0.71 M, red •).
• determination of the kinetic order of the reaction with
respect to the different components
• mechanism-based expansion of the substrate scope to
preparatively useful levels
The originally proposed “diverted Heck” mechanism, in a
simplified form (without possible side equilibria and without
hydrolysis or alcoholysis of the borate), is depicted in Scheme
5.
Information). The influence of the phosphine ligand on the
initial reaction rate was found to be negligible. These results
suggest a turnover-limiting cleavage of a dinuclear Pd(II)
G
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
An expectation that two aspects of the reaction, its overall
rate and the selectivity for the cyclopropanation over terminal
methylene formation, could be separately optimized was given
by the observation that the rate-determining step and the
product-determining step are necessarily distinct. The kinetic
studies indicated that the former involved a cleavage of a
dinuclear Pd species. The latter, on the other hand, is
determined by the branching between transmetalation and β-
hydride elimination in a Pd alkyl species. The tuning of the two
different chemical events, however, requires some indications of
the reactive intermediates in the reaction.
In principle, spectroscopic studies should seek one or more
of the intermediates: the product of oxidative addition, the
olefin complex, the product of migratory insertion, and/or the
product of intramolecular transmetalation. While THF turned
out to be inferior to DMA as a solvent for the cyclo-
propanation, THF was sufficiently dipolar aprotic and/or
sufficiently coordinating that the cyclopropanation did, in fact,
run in that solvent, albeit with a reduced yield. It was the NMR
solvent of choice, though, and the NMR experiments gave
some evidence for intermediates, especially in the case of
oxidative addition to ICH2Bpin. While the 31P and 11B spectra
had limited utility in confirming specific structures, they did
indicate that the number of species formed in the reaction of
the methylenation agent with Pd(0), with or without olefin, was
small. Having only one, or a few, species formed in solution
suggested that ESI-MS studies could be informative.
are different from those during a preparative reaction, the
observation and isolation of 8 mean that there is a good
probability that the oxidative addition reaction of Pd(0) and the
primary alkyl iodide does indeed constitute the first step in the
catalytic cycle, as proposed.
ESI-MS of reactive solutions prepared with ICH2Bpin as
methylenation agent and norbornene as substrate display ions
which are assignable to either the olefin π complex or the
product of migratory insertion. CID of the ion produces
primarily loss of the olefin, which suggests that the ions are
most likely the π complexes of norbornene on Pd(II), although
collision-induced β-hydrocarbyl elimination from the insertion
product would have been possible, in principle. We had
observed β-methyl elimination from cationic Pd(II) neopentyl
complexes in the gas phase,21 but this reaction, the microscopic
reverse of a migratory insertion, was competitive only in the
case that a β-hydride elimination is strongly disfavored or
impossible. The loss of norbornene in CID (Figure 4) is most
consistent with the parent ion being the simple π complex. Of
note is the qualitative order of reactivity for the different olefins
(norbornene > cycloheptene > cyclooctene > cyclohexene) in
the preparative reactions, which seems to follow their
propensity to form π complexes. This suggests, of course,
that ESI-MS observation of the adducts with olefins less
coordinating than norbornene might be difficult, which
corresponds in fact with our experience. Nevertheless, the
ESI-MS experiment, this time with ICH2Bpin and an olefin,
provides strong evidence supporting yet another step in the
proposed “diverted Heck” mechanism. Moreover, the slower
rate of cyclopropanation observed in the case of cis-5-decene
with respect to that of trans-5-decene seems to arise from the
steric hindrance induced by the concerted migratory syn
insertion of coordinated cis-5-decene into the palladium−
alkylboron bond in the transition state. Similar observations
were reported by James et al.22 in the carbomethoxypalladation
reaction of trans- and cis-2-butene.
Finally, the CID of the norbornene adduct with the oxidative
addition product, m/z 645 in Figure 4, gives a small product
peak at m/z 537, which would be consistent with intra-
molecular transmetalation, followed by reductive elimination of
the cyclopropane product. Intramolecular transmetalation in
solution is presumably much accelerated by nucleophilic
addition to the boronic ester, converting it to an “ate”
complex,23 but one may interpret the CID experiment to
mean that, even without nucleophilic assistance, the proposed
intramolecular transmetalation and subsequent reductive
elimination of the cyclopropane product is a plausible reaction
when nothing more facile intervenes. In solution, cyclopropane
formation via reductive elimination from proposed in situ
generated or isolated palladacyclobutane24 intermediates,
generally upon coordination of π-acceptor ligands, has already
been reported in the literature. The intermediates observed by
ESI-MS, e.g. IM1 in Figure 2, clearly indicate that alcoholysis of
the trifluoroborate is important. Some assisted version of this
reaction may proceed in solution in the final step in the
catalytic cycle. The tetracoordinate borate ester in Figure 1 is
the expected product.
ESI-MS studies were done for both ICH2BF3K and
ICH2Bpin as methylenation agents. The former reagent is
likely to be the reagent of choice in preparative applications on
the laboratory scale because it is an easily handled, air-stable,
nonhygroscopic solid, but the putative Pd(II) intermediates in
solution would be natively uncharged, meaning that electro-
spray ionization would require coordination of a proton or
alkali-metal cation, with potentially increased complexity in the
spectrum and likely reduced intensity. The alternative
methylenation agent, ICH2Bpin, had been shown to be
comparably effective in the cyclopropanation of norbornene,
but it is obtained as a moisture-sensitive oil and is therefore less
attractive as a synthetic reagent; the presumed Pd(II)
intermediates would be monocationic, though, and therefore
ideal for ESI-MS. Comparing the two, we find that the
electrospray of solutions prepared with Pd(0) and either
ICH2BF3K or ICH2Bpin yields ions assignable to the product of
the proposed initial oxidative addition step, the latter
methylenation agent unsurprisingly giving much better spectra.
In any case, the direct observation of IM1 (Figure 2) and IM2
(Figure 3) strongly supports a mechanism with the oxidative
addition preceding transmetalation. We should mention that
the special case of the very sterically demanding ligand t-
BuXPhos (Figure 5) gives an intermediate which does not
proceed to turn over, at least under the conditions of its
formation. The NMR indicated that it is formed as the
predominant species in solution. Presumably, the pendant
2,4,6-triisopropylphenyl ring of the ligand occupies the last
available coordination site on the Pd(II) center, chelation
making substitution by an incoming olefinic substrate kineti-
1
cally unfavorable. The spin−spin couplings in the H and 31P
While the NMR and ESI-MS investigations of the
intermediates in the catalytic cycle support the proposed
“diverted Heck” mechanism, the kinetics of cyclopropanation,
as determined by the RPKA, present a challenge to that
hypothesis, as noted at the beginning of the Discussion. In the
most likely interpretation, the overall kinetics of the reaction
spectra (Figure SI-36 in the Supporting Information) are fully
consistent with the proposed structure, which is also supported
independently by the ESI-MS spectrum in Figure SI-37 in the
Supporting Information and the single-crystal X-ray diffraction
result in Figure 5. While the conditions for the formation of 8
H
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 6. Proposed Detailed Mechanism for the Pd-Catalyzed Cyclopropanation of Alkenes on the Basos of the Kinetics and
the Observed or Isolated Intermediates (R = CH2CH2OH)
are determined by the rate of transformation of a dinuclear Pd
species bearing at least one molecule of EG into a monomeric,
active Pd complex, with release of one molecule of EG. This
transformation would be unsurprisingly dependent on ligand
and solvent, which offers an opportunity for the engineering of
a catalytic reaction with broader substrate scope and
preparatively useful yields. Targets for a mechanism-based
reaction engineering were 2-fold: suppression of the identified
side reaction in which a terminal methylene group, rather than
cyclopropanation, is formed and an increase in the rate of
reaction so that it can handle substrates less reactive than
norbornene.
The branching between the cyclopropane, the “diverted
Heck” product, and the terminal methylene, the product of
protodeboronation after a normal Heck reaction, would be
determined by the relative rates of intramolecular trans-
metalation versus β-hydride elimination for the intermediate
formed by migratory insertion of the olefinic substrate. In the
original report, norbornene already gave excellent yields of
cyclopropanation, but β-hydride elimination would have had to
re-form the strained double bond and can be expected to be
disfavored. For other olefins, in our case cyclooctene, or acyclic
α-olefins, the β-hydride elimination is less disfavored, leading to
significant formation of the terminal methylene product after
protodeboronation. β-hydride elimination and the deactivation
of the Pd-based catalysts are often more strongly temperature
dependent14,25 than other elementary steps in the catalytic
cycle: in particular, migratory insertion or catalyst activation.
Systematic reduction in the temperature for the Pd-catalyzed
cyclopropanation slowed the overall reaction, but, at least for
the initial experiments with H2O or CH3OH as hydroxylic
additive, it also selectively suppressed the terminal methylene
side product, improving the ratio of “diverted Heck” to normal
Heck (Table SI-3 in the Supporting Information, entries 1 and
2, in comparison to ref 13).
Alternatively, the branching between intramolecular trans-
metalation and β-hydride elimination could be shifted by
selectively accelerating the transmetalation. Transmetalation of
the trifluoroborates has been reported to require prior
hydrolysis or alcoholysis, and an acceleration by nucleophilic
assistance is well-documented.23 The nucleophilic solvents
DMF and DMA, the hydroxylic additives ranging from water to
methanol to ethylene glycol, and the base, either fluoride or
carbonate, are all already components in the successful
cyclopropanation. A reported nucleophilicity order of −OH
≪ MeO−< EtO− < n-PrO− < i-PrO− in ROH−MeCN
mixtures26 is consistent with our observation that alcohols, as
additives, provide higher selectivity for cyclopropanation than
does water. Moreover, EG presumably affords a faster
alcoholysis than a comparable monobasic alcohol with the
same nucleophilicity because of chelate effects. The latter effect
is further consistent with the decreased advantage offered by
lower temperature in the DMA/EG solvent system, a
presumably chelate promoted alcoholysis accelerating trans-
metalation sufficiently that it outcompetes β-hydride elimi-
nation even at higher temperatures.
Engineering a broader substrate scope for olefins less reactive
than norbornene is made more difficult by the temperature
reduction used to improve selectivity for cyclopropanation. The
kinetic studies, however, find that the overall rate appears to be
I
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 7. Proposed Concerted Metalation−Deprotonation Mechanism and Comparison of the π-Complexes Involved
controlled by the cleavage of a presumed dinuclear Pd complex
into an active monomeric species. An analogous interpretation
had been given for the observations by van Leeuwen, as well as
Pfaltz and Blackmond, of 0.5 kinetic order in Pd for Heck
reactions.27 The zeroth order found for the olefin, COE, and
the methylene transfer agent, as well as the inverse half-order in
EG, together with the intermediates observed in our
investigations, can be rationalized by the detailed proposed
mechanism shown in Scheme 6, which we present as a
mechanism consistent with our experimental observations, but
not necessarily exclusively. After the oxidative addition of
iodomethylboron species to the Pd(0) complex and sub-
stitution of the halide to VI, the dinuclear structure VI would
inhibit the occurrence of an undesired, and catalytically
unproductive, β-hydride elimination that one would ordinarily
expect from a Pd(II) complex with a primary alkoxide ligand
under the reaction conditions.28 An associative displacement of
one of the two oxygens of EG on VI by the olefin generates the
postulated dinuclear Pd(II) resting state VII. The bridging
electron-withdrawing alkoxide facilitates the olefin coordination
by increasing the electrophilicity of the metal center. Electron-
rich alkenes that are not sterically hindered such as COE should
thus bind favorably to Pd(II) under the reaction conditions.
Mechanistic investigations on the opening of halo-bridged
palladacycles by pyridine have shown the reaction to follow a
fast and bimolecular associative pathway.29 The cleavage of the
dinuclear Pd resting state should operate through a reversible
and concerted alkoxide-assisted deprotometalation30 mecha-
nism on the o-methyl C(sp3)−H bond of the ligand L1, leading
to the release of the bridging alkoxide and formation of the
palladacycle intermediate VIII. The release of one molecule of
EG would produce the −0.5 kinetic order in EG observed
experimentally. A strong base such as t-BuONa had already
been used to mediate the C(sp3)−H functionalization at a
benzylic sp3 position.30c,d Additionally, the close proximity
between the Pd metal and the o-methyl C(sp3)−H bond of the
ligand L1 (Tolman cone angle31 of 194°) supports the
occurrence of the concerted deprotometalation process. This
could explain why catalysts able to form alkyl-phosphapallada-
cycles work so well in this reaction. A similar cyclometalation to
a palladacycle is known from Pd(II) polymerization catalysts.32
The palladacycle intermediate VIII formed after the deproto-
nation step is destabilized by the new C(sp3)−Pd bond trans to
the alkylboron moiety (Scheme 7).
zero-order dependence observed for both COE and ICH2BF3K,
this migratory insertion must be the turnover-limiting step.
Similar zero-order dependence on the olefin had been reported
by Brookhart for Pd(II)-catalyzed olefin polymerization, for
which the resting state had been identified to be the olefin π
complex and the turnover-limiting step was the migratory
insertion.32 The resulting palladacycle product IX could either
undergo an intramolecular transmetalation via the transition
state TS[IX-X′] depicted in Scheme 6 to give the
palladacyclobutane X′ or proceed with a β-hydride elimination
to generate the intermediate X. Reductive elimination from
these two intermediates will regenerate the Pd(0) complex V.
The solvents utilized in this reaction are good coordinating
nucleophiles, and their coordination to the organometallic
reagents helps stabilize the metal catalyst and thus extend its
lifetime. Another major role played by DMA or DMF is to
ensure the solubility of the metal catalyst, the methylene
transfer agent, as well as the inorganic base during the
methylenation reaction.
The nucleophilicity is not the sole relevant performance
metric for a given additive. For example, although we observe a
general trend in which the reaction is improved as the
conjugate base of the hydroxylic component (H2O, MeOH,
other aliphatic alcohols) becomes more nucleophilic in
reference reactions,26 we observed only traces of products
when EG was replaced by pinacol, even in the presence of 2.0
equiv of methanol (used to ensure the reduction of Pd(II) to
Pd(0)). We attribute this result to the steric hindrance of
pinacol, which makes the alcoholysis of the trifluoroborate
extremely slow (Figure 7), counteracting the favorable effect of
what would have been an otherwise more nucleophilic,
chelating alkoxide.
Similarly, while EG helps to improve the selectivity for
transmetalation over β-hydride elimination, the observed −0.5
kinetic order with respect to EG means that increasing the EG
concentration slows the overall turnover frequency. Accord-
ingly, we find a nonmonotonic dependence of the cyclopropane
yield on the EG content in the reaction mixture. These
examples serve to illustrate the complicated interplay of factors
which make reaction optimization less transparent than one
would have wished.
CONCLUSION
■
We report reagents and conditions for a catalytic, electrophilic
cyclopropanation reaction without diazo compounds that
works efficiently from di- and trisubstituted, electron-rich
olefins and allylic alcohols. We further report NMR and ESI-
Such a primary alkyl anion possesses a stronger trans effect in
comparison to the phosphine in the palladacycle,33 which
should accelerate the migratory insertion.34 On the basis of the
J
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Perman, J. A.; Zhang, X. P. J. Am. Chem. Soc. 2010, 132, 12796−12799.
(f) Doyle, M. P. Angew. Chem., Int. Ed. 2009, 48, 850−852.
(g) Emerzian, M. A.; Davenport, W.; Song, J.; Li, J.; Erden, I. Adv.
Synth. Catal. 2009, 351, 999−1004.
(3) For examples of organocatalytic cyclopropanations, see:
(a) Wang, J.; Liu, X.; Dong, S.; Lin, L.; Feng, X. J. Org. Chem.
2013, 78, 6322−6327. (b) Biswas, A.; De Sarkar, S.; Tebben, L.;
Studer, A. Chem. Commun. 2012, 48, 5190−5192. (c) Cheng, Y.; An,
J.; Lu, L.-Q.; Luo, L.; Wang, Z.-Y.; Chen, J.-R.; Xiao, W.-J. J. Org.
Chem. 2011, 76, 281−284. (d) Rueping, M.; Sunden, H.; Hubener, L.;
Sugiono, E. Chem. Commun. 2012, 48, 2201−2203. (e) Raveendran, A.
E.; Paul, R. R.; Suresh, E.; Nair, V. Org. Biomol. Chem. 2010, 8, 901−
905. (f) Alba, A. N.; Companyo, X.; Viciano, M.; Rios, R. Curr. Org.
Chem. 2009, 13, 1432−1474. (g) Kunz, R. K.; MacMillan, D. W. C. J.
Am. Chem. Soc. 2005, 127, 3240−3241.
Figure 7. 19F NMR spectra of the alcoholysis of ICH2BF3K in DMA,
(4) For examples of MIRC cyclopropanations, see: (a) del Villar, I.
S.; Gradillas, A.; Dominguez, G.; Perez-Castells, J. Tetrahedron Lett.
2010, 51, 3095−3098. (b) Abramovitch, A.; Fensterbank, L.; Malacria,
M.; Marek, I. Angew. Chem., Int. Ed. 2008, 47, 6865−6868. (c) Sun, X.
L.; Tang, Y. Acc. Chem. Res. 2008, 41, 937−948. (d) Gaunt, M. J.;
Johansson, C. C. C. Chem. Rev. 2007, 107, 5596−5605. (e) Caine, D.
Tetrahedron 2001, 57, 2643−2684. (f) Hakam, K.; Thielman, M.;
Thielman, T.; Winterfeldt, E. Tetrahedron 1987, 43, 2035−2044.
(g) Huang, Y.-Z.; Shen, Y. Adv. Organomet. Chem. 1982, 20, 115−157.
(h) Little, R. D.; Dawson, J. R. Tetrahedron Lett. 1980, 21, 2609−2612.
(i) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353−
1364. (j) Bestmann, H. J.; Seng, F. Angew. Chem. 1962, 74, 154−155.
(k) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1962, 84, 867−868.
(5) For example of Pd-catalyzed cyclopropanations, see: (a) Wu, W.;
Jiang, H.; Gao, Y.; Huang, H.; Zeng, W.; Cao, D. Chem. Commun.
2012, 48, 10340−10342. (b) Lyons, T. W.; Sanford, M. S. Tetrahedron
2009, 65, 3211−3221. (c) Welbes, L. L.; Lyons, T. W.; Cychosz, K. A.;
Sanford, M. S. J. Am. Chem. Soc. 2007, 129, 5836−5837.
(6) (a) Johansson Seechurn, C. C. C.; Kitching, M. O.; Colacot, T. J.;
Sniekus, V. Angew. Chem., Int. Ed. 2012, 51, 5062−5085; Angew. Chem.
2012, 124, 5150−5174. (b) Nicolaou, K. C.; Bulger, P. G.; Sarlah, D.
Angew. Chem., Int. Ed. 2005, 44, 4442−4489; Angew. Chem. 2005, 117,
4516−4563.
in the presence of pinacol (A) and EG (B).
MS observation of intermediates in the reaction, isolation, and
crystallographic characterization of a key model intermediate, as
well as kinetics of the cyclopropanation, which support the
“diverted Heck” mechanism, on which basis we broadened the
substrate scope of the new reaction to preparatively useful
levels.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.6b00531.
■
S
Synthesis details, characterization results, additional
screening tables, GC-FID, kinetic studies, gas-phase
observations of intermediates and/or products, and
ORTEP representation and crystallographic data of
complex 8 (PDF)
X-ray crystallographic data (CIF)
(7) (a) Heck, R. F.; Nolley, J. P., Jr. J. Org. Chem. 1972, 37, 2320−
2322. (b) Mizoroki, T.; Mori, K.; Ozaki, A. Bull. Chem. Soc. Jpn. 1971,
44, 581.
AUTHOR INFORMATION
Corresponding Author
*E-mail for P.C.: peter.chen@org.chem.ethz.ch.
■
(8) (a) Suzuki, A. Acc. Chem. Res. 1982, 15, 178−184. (b) Miyaura,
N.; Yamada, K.; Suzuki, A. Tetrahedron Lett. 1979, 20, 3437−3440.
(9) Negishi, E.; King, A. O.; Okukado, N. J. Org. Chem. 1977, 42,
1821−1823.
(10) Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975,
16, 4467−4470.
(11) Hatanaka, Y.; Hiyama, T. J. Org. Chem. 1988, 53, 918−920.
(12) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1978, 100, 3636−
3638.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We acknowledge financial support from the ETH Zurich and
the Swiss National Science Foundation. We acknowledge
■
̈
helpful discussions with Prof. P. Pregosin (ETH Zurich). We
thank Dr. N. Trapp and M. Solar for the X-ray crystal structure
determination.
̈
(13) den Hartog, T.; Sarria Toro, J. M.; Chen, P. Org. Lett. 2014, 16,
1100−1103.
̈
(14) Herrmann, W. A.; Brossmer, C.; Ofele, K.; Reisinger, C.-P.;
Priermeier, T.; Beller, M.; Fischer, H. Angew. Chem., Int. Ed. Engl.
1995, 34, 1844−1848. From the X-ray structure of the palladacycle,
the distance of the P−O bond trans to the C(sp3) (214.7(3) pm) is
longer than the distance of the P−O bond trans to P (211.1(3) pm).
(15) Filliatre, C.; Gueraud, C. C.R. Acad. Sci. Paris Ser. C 1971, 273,
1186−1189.
REFERENCES
■
(1) (a) Shitama, H.; Katsuki, T. Angew. Chem. 2008, 120, 2484−
2487; Angew. Chem., Int. Ed. 2008, 47, 2450−2453. (b) Charette, A. B.;
Juteau, H. J. J. Am. Chem. Soc. 1994, 116, 2651−2652. (c) Furukawa,
J.; Kawabata, N. Adv. Organomet. Chem. 1974, 12, 83−134.
(d) Simmons, H. E.; Smith, R. D. J. Am. Chem. Soc. 1958, 80,
5323−5324.
(16) McKee, M.; Haner, J.; Carlson, E.; Tam, W. Synthesis 2014, 46,
1518−1524.
(17) Shaw, W. J.; Linehan, J. C.; Szymczak, N. K.; Heldenbrandt, D.
J.; Yonker, C.; Camaioni, D. M.; Baker, R. T.; Autrey, T. Angew. Chem.,
Int. Ed. 2008, 47, 7493−7496.
(18) Mathew, J. S.; Klussmann, M.; Iwamura, H.; Valera, F.; Futran,
A.; Emanuelsson, E. A. C.; Blackmond, D. G. J. Org. Chem. 2006, 71,
4711−4722.
(19) Sarria Toro, J. M.; den Hartog, T.; Chen, P. Chem. Commun.
2014, 50, 10608−10610.
(2) For recent examples of metal-catalyzed cyclopropanation with a
diazo species, see: (a) Shen, J.-J.; Zhu, S.-F.; Cai, Y.; Xu, H.; Xie, X.-L.;
Zhou, Q.-L. Angew. Chem., Int. Ed. 2014, 53, 13188−13191.
́
(b) Lev
́
esque, E.; Goudreau, S. R.; Charette, A. B. Org. Lett. 2014,
16, 1490−1493. (c) Lindsay, V. N. G.; Fiset, D.; Gritsch, P. J.; Azzi, S.;
Charette, A. B. J. Am. Chem. Soc. 2013, 135, 1463−1470. (d) Xu, X.;
Lu, H. J.; Ruppel, J. V.; Cui, X.; de Mesa, S. L.; Wojtas, L.; Zhang, X. P.
J. Am. Chem. Soc. 2011, 133, 15292−15295. (e) Zhu, S. F.; Xu, X.;
K
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(20) For a recent example showing the limitation of the metal-
catalyzed cyclopropanation of electron-rich olefins, see: Kunzi, S. A.;
̈
Sarria Toro, J. M.; den Hartog, T.; Chen, P. Angew. Chem., Int. Ed.
2015, 54, 10670−10674.
(21) Zhugralin, A. R.; Kobylianskii, I. J.; Chen, P. Organometallics
2015, 34, 1301−1306.
(22) James, D. E.; Hines, L. F.; Stille, J. K. J. Am. Chem. Soc. 1976, 98,
1806−1809.
(23) (a) Lennox, A. J. J.; Lloyd-Jones, G. C. Angew. Chem., Int. Ed.
2013, 52, 7362−7370. (b) Amatore, C.; Jutand, A.; Le Duc, G. Angew.
Chem., Int. Ed. 2012, 51, 1379−1382; Angew. Chem. 2012, 124, 1408−
1411. (c) Amatore, C.; Jutand, A.; Le Duc, G. Chem. - Eur. J. 2011, 17,
2492−2503. (d) Ishiyama, T.; Murata, M.; Miyaura, N. J. Org. Chem.
1995, 60, 7508−7510. (e) Moriya, T.; Miyaura, N.; Suzuki, A. Synlett
1994, 1994, 149−151.
(24) (a) Horino, Y.; Homura, N.; Inoue, K.; Yoshikawa, S. Adv.
Synth. Catal. 2012, 354, 828−834. (b) Grigg, R.; Kordes, M. Eur. J.
Org. Chem. 2001, 2001, 707−712. (c) Denmark, S. E.; Stavenger, R.
A.; Faucher, A.-M.; Edwards, J. P. J. Org. Chem. 1997, 62, 3375−3389.
(d) Hoffmann, H. M. R.; Otte, A. R.; Wilde, A.; Menzer, S.; Williams,
D. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 100−102.
(25) (a) Crabtree, R. H. Chem. Rev. 2015, 115, 127−150. (b) Rezabal,
E.; Asua, J.; Ugalde, J. M. Organometallics 2015, 34, 373−380 and the
references therein..
(26) Phan, T. B.; Mayr, H. Can. J. Chem. 2005, 83, 1554−1560.
(27) (a) Rosner, T.; Le Bars, J.; Pfaltz, A.; Blackmond, D. G. J. Am.
Chem. Soc. 2001, 123, 1848−1855. (b) van Strijdonck, G. P. F.; Boele,
M. D. K.; Kamer, P. C. J.; de Vries, J. G.; van Leeuwen, P. W. N. M.
Eur. J. Inorg. Chem. 1999, 1999, 1073−1076.
(28) For reviews with excellent references, see: (a) Schultz, M. J.;
Sigman, M. S. Tetrahedron 2006, 62, 8227−8241. (b) Schultz, M. J.;
Hamilton, S. S.; Jensen, D. R.; Sigman, M. S. J. Org. Chem. 2005, 70,
3343−3352. (c) Nishimura, T.; Uemura, S. Synlett 2004, 0201−0216.
(d) Muzart, J. Tetrahedron 2003, 59, 5789−5816.
(29) Ryabov, A. D.; Kuzmina, L. G.; Polyakov, V. A.; Kazankov, G.
M.; Ryabova, E. S.; Pfeffer, M.; Vaneldik, R. J. Chem. Soc., Dalton Trans.
1995, 999−1006.
(30) (a) Balcells, D.; Clot, E.; Eisenstein, O. Chem. Rev. 2010, 110,
749−823. (b) Boutadla, Y.; Davies, D. L.; Macgregor, S. A.; Poblador-
Bahamonde, A. L. Dalton Trans. 2009, 5820−5831. (c) Schipper, D. J.;
Campeau, L.-C.; Fagnou, K. Tetrahedron 2009, 65, 3155−3164.
(d) Campeau, L.-C.; Schipper, D. J.; Fagnou, K. J. Am. Chem. Soc.
2008, 130, 3266−3267.
(31) Tolman, C. A. Chem. Rev. 1977, 77, 313−348.
(32) Tempel, D. J.; Johnson, L. K.; Huff, R. L.; White, P. S.;
Brookhart, M. S. J. Am. Chem. Soc. 2000, 122, 6686−6700.
(33) Dupont, J.; Consorti, C. S.; Spencer, J. Chem. Rev. 2005, 105,
2527−2571.
(34) Backtorp, C.; Norrby, P.-O. J. Mol. Catal. A: Chem. 2010, 328,
̈
108−113.
L
DOI: 10.1021/acs.organomet.6b00531
Organometallics XXXX, XXX, XXX−XXX