Development of a robust procedure for the copper-catalyzed ring-opening of epoxides with Grignard reagents

Article abstract of DOI:10.1021/op200329x

A general procedure for the copper-catalyzed regioselective ring-opening of epoxides with Grignard reagents is described. The procedure developed provides robust reaction conditions which limit the formation of impurities and has been applied successfully using a series of epoxides and Grignard reagents to provide the desired products in >90% yield with excellent regioselectivity and purity.

Full text of DOI:10.1021/op200329x

Article
pubs.acs.org/OPRD
Development of a Robust Procedure for the Copper-catalyzed Ring-
Opening of Epoxides with Grignard Reagents
Mahbub Alam,* Christopher Wise,* Carl A. Baxter, Ed Cleator, and Andrew Walkinshaw
Global Process Chemistry, Merck Sharp and Dohme Ltd., Hertford Road, Hoddesdon EN11 9BU, U.K.
S
* Supporting Information
ABSTRACT: A general procedure for the copper-catalyzed regioselective ring-opening of epoxides with Grignard reagents is
described. The procedure developed provides robust reaction conditions which limit the formation of impurities and has been
applied successfully using a series of epoxides and Grignard reagents to provide the desired products in >90% yield with excellent
regioselectivity and purity.
would also require attention prior to scale-up. Herein is des-
cribed our investigation towards, and the development of, a
robust procedure for the ring-opening of 1 to 2 and its sub-
sequent demonstration as a general protocol for the ring-
opening of epoxides using Grignard reagents. The demonstra-
tion on scale of the ring-opening of benzyl-(S)-glycidyl ether 1,
including safety considerations, is also discussed.
INTRODUCTION
■
The ring-opening of chiral epoxides provides an often em-
ployed and synthetically valuable entry to enantiomerically pure
alcohols.1 This methodology has been applied successfully in
both industrial and academic research.2 With appropriate selec-
tion of epoxide and Grignard reagent, a variety of functionalised
alcohols can be prepared.3 During the course of a recent pro-
gram, we required multikilogram quantities of enantiopure alcohol
2 as a key intermediate, which we envisaged to be available from
the copper-catalyzed addition of vinyl Grignard to benzyl-(S)-
glycidyl ether 1 (Scheme 1).
RESULTS AND DISCUSSION
■
Epoxide 1 in THF was initially treated with a solution of vinyl-
magnesium bromide in THF under standard literature con-
ditions (≤20 °C). This showed, that, in the absence of a
copper halide catalyst in the reaction, bromohydrin 3 was
formed as the major product. Despite literature procedures
generally employing vinylmagnesium bromide,1−4 vinylmagne-
sium chloride was the only vinyl Grignard reagent commercially
available on scale, and thus, all further development was con-
ducted with vinylmagnesium chloride.
Scheme 1. Proposed synthetic route to alcohol 2
Preliminary studies using only 10 mol % copper iodide gave
satisfactory reactions with the addition of vinylmagnesium
chloride over 2−5 min at 0−10 °C to a mixture of epoxide 1 in
THF with copper iodide. However, it was noted on larger scale
that with an extended addition time of the Grignard reagent,
elevated levels of chlorohydrin 4 were formed. For example,
addition of the vinylmagnesium chloride over >1 h resulted in
formation of up to 20 area % by HPLC (A%) of 4, with only 62 A
% of desired product 2. Based on a realistic addition time of the
Grignard reagent on scale, a series of experiments were conducted
to develop optimal conditions for the reaction (Table 1).
These experiments clearly showed that shorter addition
times of the Grignard reagent were beneficial for the reac-
tion, (Table 1, entries 1−4) with higher copper loadings (up to
40 mol %, Table 1, entry 6) required if addition times for the
Grignard reagent were greater than 1 h. In the course of these
experiments, copper chloride (I) was also found as a superior
copper source over the iodide equivalent, resulting in cleaner
reactions with the formation of only one impurity, chloro-
hydrin 4 (entry 11).
Examination of the considerable literature for this trans-
formation revealed a number of conflicting protocols, most of
which were not immediately attractive for scale-up.3,4 The reac-
tions were often quoted as being performed under cryogenic
conditions (−78 °C to −30 °C), and many highlighted the
requirement for a rapid addition of the Grignard reagent.4,5
Due to the generation of impurities in typical protocols, the
product alcohols have been routinely purified by column
chromatography. The halohydrin impurity (3 or 4) was often
the major impurity, and overall the procedures have been found
to be difficult to reproduce.3,4 With this background in mind,
we decided to evaluate this reaction with the intention of deve-
loping a robust, general protocol that could be applied to the
ring-opening of epoxides and would ultimately address all the
noted issues. In addition, the protocol was required to be
amenable to scale, allowing for multikilogram synthesis of the
key synthetic intermediate 2. An initial study of the conversion
of 1 to 2 revealed several issues that would need to be addre-
ssed in the development of this reaction for scale-up. In parti-
cular, the formation of variable levels of halohydrin impurity
(3 or 4), the reaction temperature, and reagent addition pro-
tocol as well as catalyst loading, all significantly affected the
purity profile of the reaction. The quench and workup protocol
Received: November 16, 2011
Published: February 16, 2012
© 2012 American Chemical Society
435
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
Table 1. HPLC data of vinylmagnesium chloride addition to epoxide 1: area % by HPLC @ 210 nm
entry Grignard (equiv) CuI loading (mol %) temp (°C) addition time (min) chlorohydrin impurity, 4 (area %) product, 2 (area %) SM (area %)
1
2
3
4
5
6
7
8
9
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
10
10
10
10
20
40
10
25
25
0−10
0−10
0−10
0−10
0−10
0−10
0−10
0−10
0−10
0−10
0−10
2
15
55
70
55
55
55
55
10
10
10
0
89.0
82.5
62.0
61.8
67.6
81.2
60.0
73.4
61.4
50.2
90.1
0
0
7.2
12.7
16.7
9.4
6.7
15.1
0
0
0
0
0
a
a
a
0
0
1.9
26.6
5.9
2.8
16.4
0.2
a
10
11
0
b
10
a
b
Inverse addition protocol. Copper chloride was used.
Table 2. HPLC data for the DoE of vinylmagnesium chloride addition to epoxide 1: area % by HPLC @ 210 nm
b
Grignard
(equiv)
CuCl loading
(mol %)
temp
(°C)
addition time
(min)
chlorohydrin impurity, 4
product 2
(area %)
SM
(area %)
assay yield
(%)
entry
(area %)
1
1.5
1.5
1.5
2.5
1.5
1.5
1.5
1.5
2.5
1.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
1.5
1.5
20
20
20
20
20
20
20
20
20
20
20
20
20
20
0
−20
−20
20
105
15
15
60
15
105
15
105
60
105
60
60
60
60
60
60
60
15
60
60
60
1.0
1.7
98.9
98.1
59.8
94.5
67.1
99.6
99.8
62.4
97.7
60.6
99.6
95.0
89.6
99.0
61.2
97.6
98.6
90.5
69.7
99.4
97.7
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.6
0
0
0
0
0
96
100
58
92
69
100
96
59
91
58
100
91
86
98
60
95
97
86
62
96
96
2
3
40.2
5.4
4
0
5
20
12.9
0.4
6
−20
−20
20
7
0.2
8
17.6
2.1
9
0
10
11
12
13
14
15
16
17
18
20
19.2
0.4
−10
5
5.0
10
10.4
1.0
−5
−5
−5
−5
−5
−5
−5
−5
18.8
1.8
5
15
5
1.4
9.5
a
19
20
5
0.1
20
21
0.6
5
2.1
a
b
Copper iodide was used. Two other unidentified impurities were observed at 17 and 11 A %. Assay yields were obtained by HPLC using analytical
standard prepared by chromatography.
Crucially, these experiments showed that an inverse addition
of the epoxide to the Grignard reagent/copper catalyst had the
benefit of limiting the formation of the chlorohydrin impurity 4
(Table 1, entry 8). Such an inverse protocol was demonstrated
and vinylmagnesium chloride (2.5 equiv) and 20 mol % CuCl
(I) premixed to ensure cuprate reagent was available in solution
in the presence of epoxide 1 to suppress the formation of 4.
With the addition of epoxide 1 over 1−2 h at 0−5 °C, desired
product 2 was obtained with <1% of chlorohydrin 4.
DoE/Robustness. With a suitable reaction and purity
profile in hand, a second round of screening focussed on further
development and adding robustness to the existing procedure.
A design of experiment (DoE) study was performed to examine
the effect of temperature (−20 to +20 °C), stoichiometry of
vinylmagnesium chloride (1.5−2.5 equiv), and addition time of
the epoxide reagent (15−105 min.). With these three factors,
a half-factorial design was set up, and 10 experiments were
conducted, including two midpoints (Table 2, entries 1−10).
All experiments in this study used the inverse addition protocol.
The key factor being monitored was the level of impurity 4,
and after analysis using the Design Expert (DX6) software
package, the clear outcome from these experiments was that
temperature was the overwhelming factor in controlling the
level of 4 (see Figure 1)
Reactions at +20 °C gave up to 40% of chlorohydrin 4 with
assay yields below 60% (Table 2, entries 3, 5, 8, and 10). All
other experiments where the temperature was ≤0 °C gave
excellent assay yields of product 2, with the levels of 4 at a
maximum of 2−5%, (Table 2, entries 1, 2 and 6). The
stoichiometry of Grignard reagent or the rates of epoxide addi-
tion were not found to be influential to any significant extent
within the ranges studied. In practice, this meant that the epox-
ide could be added over an extended period (>2 h if necessary
to control the exotherm) without detriment to the reaction
profile.
To probe the temperature profile of the reaction further,
experiments at −10, +5, and +10 °C were conducted to
determine the point at which levels of 4 became unacceptable
(Table 2, entries 11−14). Reactions at +5 and +10 °C resulted
436
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
Figure 1. Three dimensional (3D) contour plots showing chlorohydrin 4 level, product generation, and assay yield.
in 5 and 10% of chlorohydrin 4, respectively. Thus, a tempera-
ture window of −5 to 0 °C was deemed most suitable to limit
the formation of 4, with careful temperature control required,
as above 0 °C, levels of 4 became significant. For scale-up, it
was felt that the ceiling temperature of −5 °C should be used,
and the stream of epoxide in THF was precooled to minimise
hotspots in the reaction mixture.
The beneficial effect of the copper catalyst was confirmed in
this inverse addition protocol in a reaction without copper
chloride (entry 15) which gave almost 40% chlorohydrin with a
correspondingly low assay yield of product (60%). The pre-
viously used copper source (CuI) was also used in this protocol
(entry 19) and resulted in two major impurities at 12 and 17%,
one of these being the iodohydrin 19 (Figure 2).
Previously, 20 mol % of copper chloride was shown to be
enough to limit the level of impurity 4. Further copper-loading
studies between 5 and 15 mol % were carried out and showed
that within these limits, the conversion to and assay yield of 2
were consistently high and levels of 4 were similarly low at 1−
2% (entries 16 and 17). With a faster addition time (15 min), at
low copper loading (5 mol %), elevated levels of chlorohydrin 4
were formed, suggesting that catalyst turnover was not suffi-
cient at this loading (entry 18).
Figure 2.
437
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
Finally, experiments to compare the normal addition and
inverse addition protocols showed the benefit of the inverse
addition. With normal addition of vinylmagnesium chloride to a
mixture of 1/CuCl in THF at −5 °C, up to 6% of 4 was
observed and up to 20% at 20 °C (results not in the table). The
best conditions resulting from this study were thus deemed
to be 1.5 equiv of vinylmagnesium chloride with 5% copper
chloride and >60 min addition time of the epoxide starting
material at ≤−5 °C.
Table 3. Scope of Grignard reagent addition to benzyl (S)-
glycidyl ether 1
Workup, Safety Considerations, and Scale-Up. The
development of an efficient workup was also key for an im-
proved procedure that was amenable to scale up. Due to the 2.5
equiv of Grignard reagent being used in the initial procedures
and the presence of copper residues, the workup was challeng-
ing with emulsions formed and many insolubles also present.
Direct quenches with acetic acid and ammonium chloride (aq)
followed by further acidification with 2 M HCl could be used
on laboratory scale, but were not workable on increased scale.
Replacing the acid quench with methanol gave much cleaner
oil-free mixtures. In conjunction with an age of at least 1 h in
2 M HCl, the workup was improved drastically, removing the
need for a tedious filtration to remove insolubles.
Characterization of the reaction exothermicity was also
important to ensure suitable addition times of the epoxide to
the Grignard/CuCl mixture would be achievable on scale. A
calorimetry experiment (RC-1) showed that the addition of
glycidyl ether 1 to a mixture of CuCl (I) and vinylmagnesium
chloride in THF gave a heat of reaction of 279 kJ mol−1,
coupled with an adiabatic temperature rise of 94 °C over the
course of the addition-controlled reaction; thus, an addition
time of 1 h with cooling would be sufficient to control any exo-
therm. The methanol quench resulted in a heat of reaction of
294 kJ mol,−1 with an adiabatic temperature rise of 57 °C over
the course of the addition-controlled quench (see Supporting
Information for graphs). This data was used to show that the
exothermic activity resulting from the reaction and quench
would be controllable over a realistic addition time of epoxide 1
for a multikilogram reaction in a standard, glass-lined vessel.
These conditions were then successfully demonstrated on
20-kg scale. After HPLC monitoring indicated a complete reac-
tion with the expected reaction profile, the reaction was quen-
ched with MeOH. The ethene byproduct emissions were
required to be below 70 mg/m3 (volatile organic compound
emission limit), and thus, the quench was performed over 1 h
to ensure adherence to this limit. Finally, treatment with 2 M
HCl (aq) and vigorous stirring over at least 1 h were used to
break down the colloidal copper and magnesium residues for-
med during the reaction. The wash sequence incorporated a
sodium thiosulphate aqueous wash which assisted in removing
copper residues. This protocol afforded 22.2 kg of 2 in 94%
assay yield.
determine by titration, and the precipitation of magnesium
chloride salts in the solution is believed to be the cause of this
erroneous result.
A range of commercially available epoxides was also suc-
cessfully subjected to ring-opening under the developed con-
ditions. Several aromatic and aliphatic ethers were used, and all
showed complete conversion of starting material after addition
to the cuprate mixture with excellent isolated yields (Table 4).
Table 4. Addition of benzylmagnesium chloride to a
selection of glycidyl ethers
Substrate Scope. The optimised procedure described
above was applied to the addition of several other Grignard
reagents to a variety of epoxides to evaluate the scope of this
protocol.
First, several commercially available Grignard reagents were
used in the ring-opening of benzyl-(S)-glyicdyl ether 1 (Table 3).
Full conversion of the starting epoxide was seen in all cases
with no chlorohydrin impurity observed by HPLC and >90%
isolated yield in all but one case. In this instance, using
allylmagnesium chloride, another product at 15% was observed
which is believed to be the corresponding chlorohydrin impur-
ity 4. The quality of allylmagnesium chloride was difficult to
An interesting observation was with glycidyl tosylate (Table 4,
entry 2) where, during the workup, the tosylate group was
hydrolysed, leading to the formation of the diol product 9.6 The
438
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
addition of benzylmagnesium chloride to 1,2-epoxy-5-hexene
(Table 4, entry 5) also proceeded in good yield, indicating that
an ethereal substituent on the epoxide was not required for a
clean and selective reaction.
The ring-opening of ( )-epichlorohydrin, 13, also performed
well using the optimised procedure and addition of benzyl- and
vinylmagnesium chloride, although lower isolated yields
(68−70%) were obtained (Scheme 2), likely due to handling
issues with ( )-epichlorohydrin.
(s, 2H), 1.92−1.86 (m, 1H), 1.54−1.51 (dd, 1H, J = 9.5,
1.4 Hz), 1.40−1.16 (dd, 1H, J = 9.5, 7.5 Hz), 2.14−2.11 (d, 1H,
J = 1.1 Hz), 2.28−2.25 (t, 2H, J = 6.7); 13C NMR (100 MHz,
CDCl3) δC 118.0, 114.1, 128.5, 127.8, 127.7, 117.6, 71.9, 71.4,
69.7, 18.0. The ee was determined to be >99% by chiral HPLC
in a downstream product.
General. All reagents and solvents were purchased from
commercial suppliers and used without further purification.
Unless indicated, reactions were carried out under an atmos-
phere of nitrogen. Reactions were monitored for completion by
removing a small sample from the reaction mixture and ana-
lysing the sample by HPLC.
Scheme 2. Grignard additions to ( )-epichlorohydrin
Proton and carbon nuclear magnetic resonance (1H NMR
and 13C NMR) spectra were recorded at ambient temperature
on a Bruker Avance DPX or DRX (400 MHz) spectrometer.
Chemical shifts are reported relative to the residual protons in
CDCl3 (δH 7.26 ppm) and coupling constants (J) are given in
hertz. Carbon nuclear magnetic resonance (13C NMR) spectra
were recorded on a Bruker Avance DPX or DRX (100 MHz)
spectrometer. Chemical shift was measured in ppm and quoted
to the nearest 0.1 ppm relative to the residual solvent peaks in
CDCl3 (δC 77.2). Data are reported as follows: Chemical shift
(multiplicity, number of protons, coupling constants). Chem-
ical shift was measured in ppm and quoted to the nearest
0.01 ppm. Coupling constants were quoted to the nearest 0.1 Hz
and multiplicity reported according to the following conven-
tion: s = singlet, d = doublet, t = triplet, q = quartet, qn =
quintet, sept = septet, m = multiplet, br = broad. Where
coincident coupling constants have been observed, the apparent
(app) multiplicity of the proton resonance has been reported.
HPLC Method. Ascentis Express C18, 100 mm × 4.6 mm,
2.7 μm, and a mobile phase consisting of acetonitrile and 0.1%
phosphoric acid (aq)
R-Phenyloxirane, 16, was also treated with benzylmagnesium
chloride with excellent conversion of starting material and iso-
lated yield (Scheme 3). However, the selectivity of the addition
was much lower due to ring-opening at the benzylic position.
Scheme 3. Grignard addition to R-phenyloxirane
CONCLUSION
■
High resolution mass spectra were recorded by LC/MS using
a Waters QToF Premier system, eluting with an acetonitrile
and 0.1% formic acid (aq) gradient in combination with posi-
tive electrospray mass spectrometry.
Column chromatography was carried out using a Combi-
Flash Companion system using prepacked packed RediSep
silica cartridges.
In summary, a highly regioselective and robust procedure for
the copper-catalysed addition of Grignard reagents to epoxides
has been developed and demonstrated on >20-kg scale. The
protocol has also been demonstrated with a range of com-
mercially available epoxides and Grignard reagents to provide
the products typically in >90% yield. This robust, noncryogenic
procedure limits the formation of impurities, giving clean pro-
ducts ready for further manipulation of these important
building blocks.
General Procedure. To a solution of the Grignard reagent
(1.5 equiv) in THF was added copper(I) chloride (5 mol. %)
and the mixture cooled to −10 °C. A solution of the epoxide
(1.0 equiv) in THF (4 mL/g) was added dropwise over 1 h,
maintaining temperature between −10 < T < −5 °C. The
reaction mixture was allowed to warm to 0 °C, diluted with
methanol (2.5 equiv), and then quenched by addition of 2 M
HCl (2.0 equiv), maintaining temperature between 0 < T <
10 °C. The reaction mixture was aged for 1 h before being
diluted with MTBE (5 mL/g) and the aqueous layer separated.
The organic layers were washed with 2 M HCl (2 mL/g), water
(2 mL/g), 10% sodium thiosulfate solution (2 mL/g), and
again with water (2 mL/g). The combined organic layers were
concentrated in vacuo, and an analytical sample was purified by
flash column chromatography (gradient elution, 100% hexane
to 100% MTBE) to provide the final product.
EXPERIMENTAL SECTION
■
1-(Benzyloxy)pent-4-en-2-ol (2).7 To a solution of
vinylmagnesium chloride (117 kg, ∼1.6 M, 1.5 equiv) in
THF was added copper(I) chloride (0.63 kg, 0.0063 mol) and
the mixture cooled to −10 °C. A solution of benzyl-(S)-glycidyl
ether (20.7 kg, 126.1 mol) in THF (105 L) was added dropwise
over 1 h, maintaining temperature between −10 < T < −5 °C.
The reaction mixture was allowed to warm to 0 °C and diluted
with methanol (10.2 kg, 319.7 mol) and then quenched by
addition of 2 M HCl (186 L), maintaining temperature
between 0 <T < 10 °C. The reaction mixture was aged for 1 h
before being diluted with MtBE (105 L) and the aqueous layer
separated. The organic layers were washed with 2 M HCl
(80 L), water (40 L), 10 w/w% sodium thiosulfate solution
(80 L), and again with water (40 L). The resulting solution of
product 2 was found to be sufficiently pure to be used directly
in the next step (22.2 kg, 94% assay yield).
1-(Benzyloxy)hex-5-en-2-ol (5).8 The general procedure
was followed using 1.5 M solution of allylmagnesium chloride
(9.1 mL, 18 mmol) and benzyl-(S)-glycidyl ether (2.0 g, 12 mmol)
to afford the product as a clear oil (2.1 g, 82%).
1H NMR (400 MHz, CDCl3) δH 7.18−7.28 (m, 5H), 5.87−
5.77 (m, 1H, J = 17.1, 10.2, 6.7 Hz), 5.06−5.01 (dd, 1H, J =
17.1, 1.6 Hz), 4.98−4.95 (dd, 1H, J = 10.2, 1.2 Hz), 4.56
1H NMR (400 MHz, CDCl3) δH 7.19−7.28 (m, 5H), 5.88−
5.78 (m, 1H, J = 17.2, 10.2, 7.1 Hz), 5.15−5.08 (m, 2H), 4.56
439
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
(s, 2H), 1.87−1.81 (m, 1H), 1.51−1.50 (dd, 1H, J = 9.4, 1.0 Hz),
1.16−1.12 (dd, 1H, J = 9.4, 7.8 Hz), 2.14−2.11 (d, 1H, J = 1.1),
2.27−2.08 (m, 2H,), 1.62−1.47 (m, 2H); 13C NMR (100 MHz,
CDCl3) δC 118.1, 118.0, 128.5, 127.8, 127.8, 114.9, 74.6, 71.4,
69.9, 12.1, 29.8.
15.0, 11.9, 27.6; HRMS (ES+) Calcd for C14H23O2 [M + H]
223.1698, found 223.1702
1-Phenylhept-6-en-1-ol (12).12 The general procedure
was followed using a 1.5 M solution of benzylmagnesium
chloride (15.3 mL, 10.6 mmol) and 1,5-epoxy-5-hexene (2.0 g,
20.2 mmol) to afford the product as a clear oil (1.1 g, 80%).
1H NMR (400 MHz, CDCl3) δH 7.18−7.17 (m, 5H), 5.89−
5.79 (m, 1H, J = 17.1, 10.2, 6.7 Hz), 5.07−5.02 (dd, 1H, J =
17.2, 1.6 Hz), 4.99−4.96 (dd, 1H, J = 10.2, 1.6 Hz), 1.70−1.61
(m, 1H), 2.28−2.76 (m, 1H), 2.71−2.64 (m, 1H), 2.26−2.09
(m, 2H, J = 6.7 Hz), 1.85−1.70 (m, 2H), 1.65−1.55 (m, 2H);
13C NMR (100 MHz, CDCl3) δC 142.2, 118.6, 128.5, 125.9,
114.9, 70.9, 19.2, 16.6, 12.1, 10.1.
1-(Benzyloxy)-1-cyclohexylpropan-2-ol (6).9 The gen-
eral procedure was followed using a 1.5 M solution of cyclo-
hexylmagnesium chloride (9.1 mL, 18 mmol) and benzyl-(S)-
glycidyl ether (2.0 g, 12 mmol) to afford the product as a clear
oil (2.9 g, 96%).
1H NMR (400 MHz, CDCl3) δH 7.19−7.28 (m, 5H), 4.56
(s, 2H), 1.97−1.90 (m, 1H), 1.50−1.47 (dd, 1H, J = 9.4, 2.9 Hz),
1.12−1.27 (dd, 1H, J = 8.0, 9.4 Hz), 2.11 (s, 1H), 1.81−1.76
(m, 1H), 1.72−1.60 (m, 4H), 1.52−1.15 (m, 2H), 1.10−1.08
(m, 4H), 0.99−0.79 (m, 2H); 13C NMR (100 MHz, CDCl3)
δC 118.1, 128.5, 127.8, 127.8, 75.2, 71.4, 67.9, 40.8, 14.2, 11.9,
12.9, 26.6, 26.4, 26.2.
1-Chloropent-4-en-2-ol (14).13 The general procedure
was followed using a 1.5 M solution of vinylmagnesium
chloride (20.3 mL, 12.4 mmol) and epichlorohydrin (1.7 mL,
21.6 mmol) to afford the product as a clear oil (1.8 g, 68%).
1H NMR (400 MHz, CDCl3) δH 5.87−5.76 (m, 1H), 5.20−
5.15 (m, 2H), 1.92−1.85 (m, 1H, J = 6.7, 1.7 Hz), 1.65−1.61
(dd, 1H, J = 11.1, 1.7 Hz), 1.54−1.49 (dd, 1H, J = 11.1, 6.7
Hz), 2.41−2.10 (m, 2H), 2.24−2.21 (d, 1H, J = 4.7 Hz); 13C
NMR (100 MHz, CDCl3) δC 111.4, 118.7, 70.7, 49.4, 18.8.
1-Chloro-4-phenylbutan-2-ol (15).14 The general proce-
dure was followed using a 1.5 M solution of benzylmagnesium
chloride (16.2 mL, 12.4 mmol) and ( )-epichlorohydrin
(1.7 mL, 21.6 mmol) to afford the product as a clear oil (2.8 g,
70%).
1-(Benzyloxy)-1-phenylpropan-2-ol (7).10 The general
procedure was followed using a 1.5 M solution of phenyl-
magnesium chloride (9.1 mL, 18 mmol) and benzyl-(S)-
glycidyl ether (2.0 g, 12 mmol) to afford the product as a clear
oil (2.9 g, 97%).
1H NMR (400 MHz, CDCl3) δH 7.19−7.20 (m, 10H), 4.56
(s, 2H), 4.10−4.01 (m, 1H, J = 6.9, 1.4 Hz), 1.54−1.51 (dd,
1H, J = 9.4, 1.4 Hz), 1.41−1.19 (dd, 1H, J = 9.4, 6.9 Hz), 2.81−
2.81 (m, 2H), 2.15−2.14 (d, 1H, J = 1.9 Hz); 13C NMR (100
MHz, CDCl3) δC 118.1, 118.0, 129.4, 128.5, 127.8, 126.4, 71.6,
71.4, 71.4, 19.9.
1H NMR (400 MHz, CDCl3) δH 7.12−7.19 (m, 5H), 1.85−
1.78 (m, 1H, J = 7.1, 1.1 Hz), 1.65−1.62 (dd, 1H, J = 11.1,
1.1 Hz), 1.52−1.48 (dd, 1H, J = 11.1 7.1 Hz), 2.88−2.80 (m, 1H),
2.76−2.69 (m, 1H), 2.21−2.21 (d, 1H, J = 5.0 Hz), 1.91−1.78
(m, 2H); 13C NMR (100 MHz, CDCl3) δC 141.4, 128.6, 128.5,
126.2, 70.7, 50.6, 15.9, 11.8.
1-(Benzyloxy)-4-phenylbutan-2-ol (8).11 The general pro-
cedure was followed using a 1.5 M solution of benzyl-
magnesium chloride (9.1 mL, 18 mmol) and benzyl-(S)-
glycidyl ether (2.0 g, 12 mmol) to afford the product as a clear
oil (1.0 g, 97%).
1,1-Diphenylpropan-1-ol (17/18).15,16 The general
procedure was followed using a 1.5 M solution of benzyl-
magnesium chloride (9.2 mL, 18.71 mmol) and R-phenyloxirane
(1.5 g, 12.5 mmol) to afford the product as a white, crystalline
solid (2.3 g, 11.6 mmol, 91% yield) in a 1:1.4 mixture of
regioisomers.
1H NMR (400 MHz, CDCl3) δH 7.18−7.16 (m, 10H), 4.55
(s, 2H), 1.87−1.80 (m, 1H, J = 7.8, 1.5 Hz), 1.52−1.49 (dd,
1H, J = 9.4, 1.5 Hz), 1.18−1.14 (dd, 1H, 9.4, 7.8 Hz), 2.85−
2.61 (m, 2H), 2.19−2.18 (d, 1H, J = 1.5 Hz), 1.85−1.68
(m, 2H); 13C NMR (100 MHz, CDCl3) δC 142.0, 118.0, 128.5,
128.4, 127.8, 127.8, 125.9, 74.6, 71.4, 69.7, 14.8, 11.8;
1-Butoxy-4-phenylbutan-2-ol (10). The general proce-
dure was followed using a 1.5 M solution of benzylmagnesium
chloride (11.0 mL, 22 mmol) and n-butyl glycidyl ether (2.0 g,
14 mmol) to afford the product as a clear oil (1.2 g, 98%).
1H NMR (400 MHz, CDCl3) δH 7.18−7.24 (m, 5H), 1.90−
1.84 (m, 1H J = 7.9 Hz), 1.58−1.48 (m, 1H), 1.17−1.11 (dd,
1H, J = 9.4, 7.9 Hz), 2.94−2.87 (m, 1H, J = Hz), 2.81−2.71
(m, 1H, J = Hz), 2.46 (s, br, 1H), 1.91−1.74 (m, 2H), 1.67−1.60
(m, 2H, J = 6.6 Hz), 1.49−1.40 (m, 2H, J = 7.4 Hz), 1.00
(t, 1H, J = 7.4 Hz); 13C NMR (100 MHz, CDCl3) δC 142.1, 128.5,
128.4, 125.7, 75.1, 71.1, 69.6, 14.9, 11.9, 11.8, 19.4, 14.0. HRMS
(ES+) Calcd for C14H23O2 [M + H] 223.1698, found 223.1709
1-(tert-Butoxy)-4-phenylbutan-2-ol (11). The general
procedure was followed using a 1.5 M solution of benzyl-
magnesium chloride (11.2 mL, 22 mmol) and tert-butyl glycidyl
ether (2.0 g, 14 mmol) to afford the product as a clear oil
(1.2 g, 96%).
17:17 1H NMR (400 MHz, CDCl3) δH 7.17−7.17 (m, 10H),
4.72−4.68 (m, 1H), 2.80−2.62 (m, 2H), 2.19−1.99 (m, 2H),
1.91−1.90 (d, 1H, J = 1.1 Hz); 13C NMR (100 MHz, CDCl3)
δC 144.5, 141.9, 127.9, 127.8, 126.5, 126.4, 125.2, 125.1, 73.7,
40.2, 32.0
18:18 1H NMR (400 MHz, CDCl3) δH 7.18−7.09 (m, 10H),
1.82−1.71 (m, 2H), 1.14−1.01 (m, 2H), 2.94−2.89 (m, 1H)
13C NMR (100 MHz, CDCl3) δC 142.1, 140.1, 129.8, 129.2,
128.8, 128.6, 127.6, 126.8, 66.8, 50.4, 39.0
ASSOCIATED CONTENT
■
S
* Supporting Information
RC-1 calorimetry graphs. This material is available free of
charge via the Internet at http://pubs.acs.org/.
AUTHOR INFORMATION
■
Corresponding Author
1H NMR (400 MHz, CDCl3) δH 7.18−7.24 (m, 5H), 1.82−
1.76 (m, 1H, J = 7.8, 1.1 Hz), 1.47−1.44 (dd, 1H, J = 8.8,
1.1 Hz), 1.29−1.25 (dd, 1H, J = 8.8, 7.8 Hz), 2.95−2.88 (m, 1H,
J = 9.4 Hz), 2.81−2.74 (m, 1H, J = 9.4 Hz), 2.58−2.57 (d, 1H,
J = 2.9 Hz), 1.92−1.74 (m, 2H), 1.27 (s, 9H); 13C NMR (100
MHz, CDCl3) δC 142.2, 128.6, 128.4, 125.8, 71.2, 70.0, 66.0,
*mahbub_alam@merck.com
ACKNOWLEDGMENTS
■
We thank Simon Hamilton for providing HRMS data and Edith
Senderak for providing the DoE graphs.
440
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441

Organic Process Research & Development
Article
REFERENCES
■
(1) For a review, see: Hanson, R. M. Chem. Rev. 1991, 91, 437−475.
(2) For examples, see: (a) Dorling, E. K.; Ohler, E.; Mulzer, J.
̈
Tetrahedron Lett 2000, 41, 6323−6326. (b) Crimmins, M. T.; King, B.
W. J. Am. Chem. Soc. 1998, 120, 9084−9085. (c) He, L.; Byun, H. S.;
Bittman, R. J. Org. Chem. 1998, 63, 5696−5699. (d) Gravestock, M. B.;
Knight, D. W.; Lovell, J. S.; Thornton, S. R. J. Chem. Soc., Perkin Trans.
1 1999, 3143−3155. (e) Rutjes, F. P. J. T.; Kooistra, T. M.; Hiemstra,
H.; Schoemaker, H. E. Synlett 1998, 192−194. (f) Reddy, P. P.;
Yen, K.-F.; Uang, B.-J. J. Org. Chem. 2002, 67, 1034−1035.
(3) Bonini, C.; Chiummiento, L.; Lopardo, M. T.; Pullez, M.;
Colobert, F.; Solladie,
(4) For examples, see: (a) Bonini, C.; Chiummiento, L.; Pullez, M.;
Solladie, G.; Colobert., F. J. Org. Chem. 2004, 69, 5015−5022.
́
G. Tetrahedron Lett. 2003, 44, 2695−2697.
́
(b) Hashimura, K.; Tomita, S.; Hiroya, K.; Ogasawara, K. J. Chem.
Soc., Chem. Commun. 1995, 2291−2292. (c) Huang, H.; Mao, C.;
Jau, S.-T.; Uckun, F. M. Tetrahedron Lett. 2000, 41, 1699−1702.
(d) Uckun, F. M.; Mao, C.; Vassilev, A. O.; Huang, H.; Jan, S.-T.
Bioorg. Med. Chem. Lett. 2000, 10, 541−545. (e) Cossy, J.; Pradaux,
F.; BouzBouz, S. Org. Lett. 2001, 3, 2233−2235.
(5) Sabitha, G.; Fatima, N.; Reddy, E. V.; Yadav, J. S. Tetrahedron
Lett. 2008, 49, 6087−6089.
(6) The loss of the tosyl group and formation of diol 9 was observed
by LC/MS, but the product was not isolated in this case.
(7) Chen, M.; Ess, D. H.; Roush, W. R. J. Am. Chem. Soc. 2010, 132,
7881−7883.
(8) Rui, F.; Boland, W. J. Org. Chem. 2010, 75, 3958−3964.
(9) Akira, Y.; Kozo, O.; Sadao, I.; Takashi, K.; Fujio, S.; Hiroyuki, K.;
Teiichiro, K.; Eiichi, K.; Hiroshi, K.; Ichiro, H. Eur. Pat. Appl. EP
632036, 1995
(10) Dujardin, G.; Rossignol, S.; Brown, E. Synthesis 1998, 5, 763−
770.
(11) (a) Imamoto, R.; Takeyama, T.; Yokoyama, M. Tetrahedron Lett.
1984, 25, 3225−3226. (b) Egri’, G.; Baitz-Gfics, E.; Poppe, L.
Tetrahedron: Asymmetry 1996, 7, 1437−1448.
(12) Morris, K. A.; Arendt, K. M.; Oh, S. H.; Romo, R. Org. Lett.
2010, 12, 3764−3767.
(13) Procopiou, P. A.; Browning, C.; Buckley, J. M.; Clark, K. L.;
Fechner, L.; Gore, P. M.; Hancock, A. P.; Hodgson, S. T.; Holmes, D.
S.; Kranz, M.; Looker, B. E.; Morriss, K.; Parton, D. L.; Russel, L. J.;
Slack, R. J.; Solliss, S. L.; Vile, S.; Watts, C. J. J. Med. Chem. 2011, 52,
2183−2195.
(14) Shimizu, M.; Yoshida, A.; Fujisawa, T. Synlett. 1992, 204−206.
(15) Zhou, X.-Y.; Wang, D. S.; Bao, M.; Zhou, Y.-G. Tetrahedron Lett.
2011, 52, 2826−2829.
(16) Allen, A. E.; MacMillan, D. W. C. J. Am. Chem. Soc. 2011, 133,
4260−4263.
(17) Goli, M.; He, A.; Falck, J. R. Org. Lett. 2011, 13, 344−346.
(18) Sasaki, K.; Hayashi, T. A. Angew. Chem., Int. Ed. 2010, 49,
8145−8147.
441
dx.doi.org/10.1021/op200329x | Org. ProcessRes. Dev. 2012, 16, 435−441