Efficient and scalable protocol for the Z-selective synthesis of unsaturated esters by horner-wadsworth-emmons olefination

Article abstract of DOI:10.1002/adsc.200404338

The synthesis of a highly Z-selective Horner-Wadsworth-Emmons (HWE) reagent from cheap and readily available starting materials on a 200-g scale is described. The HWE reaction conditions have also been studied in order to design an efficient and scalable

Full text of DOI:10.1002/adsc.200404338

UPDATES
Efficient and Scalable Protocol for the Z-Selective Synthesis of
ꢀ
ꢀ
Unsaturated Esters by Horner Wadsworth Emmons Olefination
FranÅois P. Touchard,* Nicolas Capelle, Mathilde Mercier
`
Rhodia Recherches, Centre de Recherches de Lyon, 85 rue des freres Perret, BP 62, 69192 Saint Fons Cedex, France
Fax: (þ33)-4-72-89-68-94, e-mail: francois.touchard@eu.rhodia.com
Received: November 2, 2004; Accepted: February 7, 2005
Abstract: The synthesis of a highly Z-selective Horn-
er–Wadsworth Emmons (HWE) reagent from
hydes.[8] But in order to render the Z-selective synthesis
of unsaturated esters even more practical, this article
presents an improved and scalable synthesis of phospho-
nate 3 as well as optimisation of the HWE reaction con-
ditions.
ꢀ
cheap and readily available starting materials on a
200-g scale is described. The HWE reaction condi-
tions have also been studied in order to design an ef-
ficient and scalable protocol leading to Z-selectivi-
ties close to 95% with aromatic and aliphatic alde-
hydes.
ꢀ
Keywords: alkenes; Horner–Wadsworth Emmons
reaction; olefination; preparative scale; stereoselec-
tivity
Scheme 1. Z-selective HWE reagents.
Introduction
Results and Discussion
ꢀ
The Horner–Wadsworth Emmons (HWE) reaction is a
variant of the Wittig reaction that possesses a number of
significant advantages. Not only is the work-up greatly
Phosphonate Synthesis Improvement
facilitated since the phosphate salt formed is water-solu- This is an important topic to address because as selective
ble but the reactivity of HWE reagents allows them to as the HWE reaction may be, the syntheses of diaryl
effect transformations that are either difficult or impos- phosphonoacetates published so far have not always
sible using the analogous Wittig ylides. This reaction, been very practical, requiring moreover most of the
which usually proceeds smoothly and under mild condi- time chromatographic purifications.
tions with phosphonates bearing stabilising substituents,
Compound 3 had been primarily prepared by Arbu-
typically leads to the more stable E-disubstituted ole- sov chemistry according to Scheme 2 and obtained as a
fins.[1] As it is highly desirable to be able to prepare the white crystalline solid with 98% purity.[8]
E or Z isomer at will, Still,[2] in 1983, and Ando,[3,4] in
Even if the Arbusov reaction proved to proceed quite
1995, developed two Z-selective reagents: methyl bis- well, the phosphite synthesis relying on PCl2(OEt) as a
(trifluoroethyl)phosphonoacetate (1) and ethyl diphe-
nylphosphonoacetate (2), respectively (Scheme 1).[5]
However, the conditions required to obtain high Z/E ra-
tios were fairly stringent and usually necessitated work
at low temperatures.[6] With the aim of establishing an
efficient and scalable protocol for the Z-selective olefi-
nation of aldehydes, we have decided to reinvestigate
and modulate the Still/Ando procedures. Using Stillꢀs
reagent, we have already shown that it is possible to
work at 08C with K2CO3/10 mol % 18-crown-6 and to
obtain selectivities up to 98%. However, these condi-
tions are effective essentially with aromatic aldehydes.[7]
More recently, we have also reported the phosphonate
based upon 2-t-butylphenol (3; Scheme 1), to be highly
Scheme 2. Former synthesis of phosphite 4 and phosphonate
selective at 08C with both aromatic and aliphatic alde- 3.
Adv. Synth. Catal. 2005, 347, 707–711
DOI: 10.1002/adsc.200404338
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
707

FranÅois P. Touchard et al.
UPDATES
[4][BrCH2CO2Et], with k¼A exp (ꢀEa/RT); Aꢁ5.7ꢂ
109, Eaꢁ81 kJ/mol.
This kinetic law also proved to be quite in agreement
with experimental observations on a larger scale. We in-
deed observed conversions of over 95% in 8 h when
working at 1308C with 2 molar equivalents of ethyl bro-
moacetate. After the removal of excess BrCH2CO2Et,
the crude phosphonate was further purified to over
99 mol % by recrystallisation in heptane.[10] This has
been done twice on a 200-g scale leading to 440 g of 3
(see the Experimental Section).
Scheme 3. Improved synthesis of phosphite 4.
HWE Reaction
starting material is not suitable for a high-scale imple- As mentioned in the Introduction, we have already de-
mentation. This reagent is indeed not readily available. scribed that 3 was a phosphonate of choice leading to se-
So, we have decided to look at the preparation of 4 by lectivities close to 95% at 08C with benzaldehyde, cyclo-
first reacting 2-t-butylphenol with PCl3 and then treating hexanecarboxaldehyde and octanal (Scheme 4).[8]
the chlorophosphite thus formed with ethanol in a one-
pot, two-step procedure (Scheme 3).
With the objective of working under more industrially
realistic conditions, simplification of the basic system
This strategy proved to be especially effective thanks was studied. Beginning with KOH as the base, the reac-
to the steric crowding of the 2-t-butyl group; the reaction tion with benzaldehyde at 08C has first been carried out
has been carried out twice on a 3-L scale and the mixed in THF and other ether solvents (Figure 2).
phosphite was twice obtained with an unoptimised yield
of 95% and a purity higher than 98% (31P NMR). We
thus prepared close to 500 g of 4 without noticeable con-
tamination by other phosphites (see the Experimental
Section).[9]
Then, since the first time we carried out the Arbusov
reaction (Scheme 2),[8] 30 h were required for it to go
to completion, we decided to perform a kinetic study
1
to better control the phosphonate formation. H/31P
NMR experiments were thus carried out at several tem-
peratures and various initial compositions (Figure 1).
The data collected proved to be consistent with the re-
action being first order concerning both reagents : v¼k
Scheme 4. HWE reaction of 3 with aldehydes.
Figure 2. Reaction of 3 with benzaldehyde and KOH at 08C
in THF and other ether solvents. Conditions: 3 (0.5 mmol,
1.1 equivs), benzaldehyde (0.44 mmol, 1 equiv.), KOH
(0.70 mmol, 1.6 equivs.), solvent (10 mL), 08C. Selectivity,
Figure 1. Phosphonate formation during the Arbusov reac-
tion.
S¼Z/(ZþE), and benzaldehyde conversion were deter-
mined by GC.
708
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 707–711

Z-Selective Synthesis of Unsaturated Esters by HWE Olefination
Table 1. Reaction of 3 with NaOH or KOH at 08C in THF.[a]
UPDATES
octanal using either KOH or NaOH as bases are given
in Table 1.
Entry
Aldehyde
Base
Selectivity
[%][b]
Conversion
[%][c]
Completion was here observed in each case within 1 h,
with selectivities very similar to those already reported
with NaHMDS or KHMDS and NaI/TMG.[8] These re-
sults confirm that the cation plays a predominant role
and that the Z/E ratio is not sensitive to the base
strength. Unfortunately, the selectivity dropped dramat-
ically when we increased the concentration. Working at
1 M with benzaldehyde and KOH, only 75% Z-selectiv-
ity was observed at 08C and 85% at ꢀ108C.
1
2
3
4
5
6
PhCHO
PhCHO
CyCHO
CyCHO
n-Octanal
n-Octanal
KOH
NaOH
KOH
NaOH
KOH
NaOH
93
86
95
95
93
93
100
100
100
100
100
100
[a]
3 (0.5 mmol, 1.1 equivs.), aldehyde (0.44 mmol, 1 equiv.),
base (0.70 mmol, 1.6 equiv.), THF (10 mL), 08C, 1 h.
Selectivity S¼Z/(ZþE) calculated by GC.
We then carried on the evaluation of bases with
K2CO3, the reaction with benzaldehyde being first per-
formed in various aprotic solvents. Slow reaction rates
were observed but addition of two equivalents of water
proved to speed up the transformation, especially in
[b]
[c]
Aldehyde conversion calculated by GC.
Although completion was observed within 2 h in each THF, without any selectivity decrease. Both the selectiv-
case, only THF led to an acceptable Z/E ratio. The selec- ity and the conversion obtained in these conditions are
tivities and conversions then obtained in THF at 08C shown in Figure 3.
with benzaldehyde, cyclohexanecarboxaldehyde and
DMSO thus led to the fastest reaction but with rather
low selectivity (75%). DMF, DMAC and NMP then im-
proved the Z/E ratio close to 85/15. THF enabled us to
approach 90% and CH3CN to reach 93% Z-selectivity.
But reaction times higher than 48 h were required in
these conditions for conversions to exceed 80%.
Trying to increase the reaction rate, we then turned
our attention to Cs2CO3 and K3PO4. The former base en-
abled us to reduce noticeably the time to completion but
did not improve the selectivities: 91% Z-selectivity was,
for example, obtained with 99% conversion after 1 h in
CH3CN. K3PO4 led to the same observations but was fur-
ther evaluated due to its low price and availability. The
selectivities and conversions obtained with benzalde-
hyde, cyclohexanecarboxaldehyde and octanal in
DMF, CH3CN and THFare thus summarised in Table 2.
DMFalways gave the highest reaction rates with con-
versions close to 90% in 1 h and mean selectivities of
85%. CH3CN then led to the best selectivities with the
three aldehydes being, respectively, 93%, 92% and
91% with benzaldehyde, cyclohexanecarboxaldehyde
Figure 3. Reaction of 3 with benzaldehyde and K2CO3 at
208C. Conditions: 3 (0.5 mmol, 1.13 equivs.), benzaldehyde
(0.44 mmol, 1 equiv.), K2CO3(0.88 mmol, 2 equiv.), solvent
(10 mL), H2O (0.88 mmol, 2 equivs.), 208C. Selectivity, S¼
Z/(ZþE), and benzaldehyde conversion were determined
by GC.
Table 2. Reaction of 3 with K3PO4 at room temperature.[a]
Entry
Aldehyde
Solvent
Time [h]
Selectivity [%][b]
Conversion [%][c]
1
2
3
4
5
6
7
8
9
PhCHO
PhCHO
PhCHO
CyCHO
CyCHO
CyCHO
n-Octanal
n-Octanal
n-Octanal
DMF
CH3CN
THF
DMF
CH3CN
THF
DMF
CH3CN
THF
1
2
19
1
2
19
1
86
93
88
84
92
92
85
91
90
100
95
88
87
81
77
97
91
94
2
19
[a]
3 (0.5 mmol, 1.13 equivs.), aldehyde (0.44 mmol, 1 equiv.), K3PO4 (0.88 mmol, 2 equivs.), solvent (10 mL), 208C.
Selectivity S¼Z/(ZþE) calculated by GC.
[b]
[c]
Aldehyde conversion calculated by GC.
Adv. Synth. Catal. 2005, 347, 707–711
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
709

FranÅois P. Touchard et al.
UPDATES
In summary, we think we have developed a system to
make the Z-selective synthesis of unsaturated esters by
ꢀ
Horner–Wadsworth Emmons olefination quite practi-
cal. Extension of these conditions to phosphonates bear-
ing various stabilising groups would expand further the
scope of the reaction.
Experimental Section
General Remarks
All the commercially available reagents were used as received
without any further purification. The HWE reactions were per-
formed in Schlenk-like flasks with magnetic stirring unless oth-
erwise stated. All the compounds prepared (mixtures of olefins
and phosphonate 3) were already known in the literature and
their structures were confirmed by accordance with the pub-
lished data (1H, 13C, 31P NMR). The GC analyses were per-
formed with a Varian 3400Cx apparatus equipped with a
SGE BPX5(764) 25 mꢂ0.53 mmꢂ1 mm column and FID de-
tector. Helium (5 mL/min) was used as vector gas and N2
(25 mL/min) as make-up. Programme: 808C (4 min) !
1708C with heating rate 78C/min ! 3008C (4 min) with heat-
ing rate 308C/min. In these conditions, the retention times of
the analysed compounds are given hereafter.
Figure 4. Reaction of 3 with benzaldehyde, cyclohexanecar-
boxaldehyde and octanal in CH3CN at 08C. Conditions: 3
(35 mmol, 1.1 equivs.), aldehyde (31.8 mmol, 1 equiv.),
K3PO4 (64 mmol, 2 equivs.), CH3CN (35 mL), 08C. Selectivi-
ty, S¼Z/(ZþE), and aldehyde conversion were determined
by GC.
and octanal at room temperature. THF finally also gave
close to 90% Z-selectivity but with extended reaction
times.
However, in contrast with what had been observed
with KOH, we were delighted to observe that the selec-
tivity was maintained upon increasing the concentration
under the CH3CN/K3PO4 conditions. This was checked
Table 3.
with the three aldehydes by running the reaction at Compound
Rt (min)
08C, 1 M and on a 32-mmol scale. Z/E ratios close to
95/5 have been obtained with conversions of over 90%
within 6 h (Figure 4).
Octanal
The Z/E ratios may still be further improved by mod-
PhCHO
CyCHO
5.1
4.4
5.7
14.1
15.6
12.6
14.4
13.9
15.0
22.5
ꢀ
¼
ꢀ
Ph CH CH CO2Et (Z isomer)
ification of the conditions[11] but the system we have de-
veloped is a good starting basis for the practical synthesis
of unsaturated esters with high Z-selectivities.
Moreover, the phosphate salt formed during the
HWE reaction being only very slightly soluble in
CH3CN,[12] it can be easily removed by filtration from
the reaction medium leaving the clean mixture of olefins
in the filtrate for quantitative recovery (see the Experi-
mental Section).[13]
ꢀ
¼
ꢀ
Ph CH CH CO2Et (E isomer)
ꢀ
¼
ꢀ
Cy CH CH CO2Et (Z isomer)
ꢀ
¼
ꢀ
Cy CH CH CO2Et (E isomer)
ꢀ
¼
ꢀ
n-C7H15 CH CH CO2Et (Z isomer)
ꢀ
¼
ꢀ
n-C7H15 CH CH CO2Et (E isomer)
Phosphonate 3
Phosphonate 3[8]
In a 3-L vessel under nitrogen were introduced anhydrous tol-
uene (1.6 L) and PCl3 (100 g; 0.71 mol, 1 equiv.). The mixture
was then cooled to ꢀ158C and 2-t-butylphenol (215 mL,
1.39 mol, 1.95 equivs.) was introduced within 10 minutes. After
the addition of tripropylamine (418 mL, 2.16 mol, 3.02 equivs.)
over 2 h, the mixture was further stirred at ꢀ158C for 2 h. Ab-
solute ethanol (40 mL, 0.68 mol, 0.96 equivs.) was then poured
into the vessel over 30 minutes and the medium was slowly
warmed to room temperature overnight. The mixture was
then washed with water (775 mL) and the resulting organic
phase was passed over a pad of basic alumina. Solvent evapo-
ration under reduced pressure then led to a pale yellow oil;
yield: 246 g (96%); purity>98 mol % (NMR).
Conclusion
In this study, we have first developed an efficient and
scalable synthesis of 3: 440 g of the phosphonate have
thus been prepared in the laboratory from cheap and
readily available starting materials with an overall unop-
timised yield of 82%. Moreover, we have been able to
delineate very simple, mild, productive and inexpensive
conditions leading to selectivities close to 95% and
quantitative yields with benzaldehyde, cyclohexanecar-
boxaldehyde and octanal at 08C.
710
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 707–711

Z-Selective Synthesis of Unsaturated Esters by HWE Olefination
UPDATES
The above phosphite (231 g, 0.62 mol, 1 equiv.) was then
added to ethyl bromoacetate (140 mL, 1.2 mol, 1.99 equivs.)
over 1 h at 1308C and the mixture was further stirred for 8 h un-
der argon. The excess of ethyl bromoacetate was then removed
under reduced pressure and the crude phosphonate obtained
was further purified to over 99 mol % by recrystallisation in
heptane leading to a white solid; yield: 230 g (86%); purity>
99% (NMR).
the vinyl proton signals: 95% with both benzaldehyde and cy-
clohexanecarboxaldehyde and 93% with octanal.
The yields of the mixture of alkenes obtained in the reaction
with benzaldehyde, cyclohexanecarboxaldehyde and octanal
were determined to be 97%, 96% and 95%, respectively by
1H NMR analysis (300 MHz) using dimethyl terephthalate as
an internal standard.
Acknowledgements
HWE Reaction with K2CO3, Cs2CO3, K3PO4; General
Procedure (Figure 2, Table 2)
FranÅois Metz and Laurent Saint-Jalmes are gratefully ac-
knowledged for their advice.
Phosphonate 3, the base, the solvent and the aldehyde were
mixed and stirred at room temperature. The reaction was
then monitored by gas chromatography after quenching ali-
quots with saturated ammonium chloride and extracting the
mixture with MTBE.
References and Notes
[1] For a general review concerning the Wittig olefination
reaction and modifications see: B. E. Maryanoff, A. B.
Reitz, Chem. Rev. 1989, 89, 863–927.
[2] W. C. Still, C. Gennari, Tetrahedron Lett. 1983, 24, 4405–
4408.
HWE Reaction with KOH, NaOH; General
Procedure (Figure 3, Table 3)
[3] K. Ando, Tetrahedron Lett. 1995, 36, 4105–4108.
[4] Since the first publication of Ando,[3] many phosphonate
modifications have been reported: a) K. Ando, J. Org.
Chem. 1997, 62, 1934–1939; b) K. Ando, J. Org. Chem.
1998, 63, 8411–8416; c) K. Ando, J. Org. Chem. 1999,
64, 8406–8408; d) K. Ando, T. Oishi, M. Hirama, H.
Ohno, T. Ibuka, J. Org. Chem. 2000, 65, 4745–4749;
e) K. Kokin, J. Motoyoshiya, S. Hayashi, H. Aoyama,
Synth. Commun. 1997, 27, 2387–2392; f) K. Kokin, K. Ii-
take, Y. Takaguchi, H. Aoyama, S. Hayashi, J. Motoyosh-
iya, Phosphorus, Sulfur and Silicon 1998, 133, 21–40.
[5] For recent reviews, see: a) K. Ando, Yuki Gosei Kagaku
Kyokaishi 2000, 58, 869–876; b) J. Motoyoshiya, Trends
Org. Chem. 1998, 7, 63–73.
[6] The phosphonate based on 2,4-difluorophenol described
by Motoyoshiya et al.[4e, f] was until recently[8] the only
one reported leading to selectivities over 92% at 08C
with various aldehydes.
[7] F. P. Touchard, Tetrahedron Lett. 2004, 45, 5519–5523.
[8] F. P. Touchard, Eur. J. Org. Chem., accepted.
[9] P(OAr)3 and P(OEt)2(OAr), were only observed, if at
all, as traces in the crude compound.
[10] The solubility of 3 in heptane was determined to be 1.3%
(w/w) at 238C and 0.5% at 58C.
[11] For example, the selectivity was improved to 96% with
benzaldehyde by running the reaction at –108C.
[12] The solubility was determined to be 0.05% (w/w) at
room temperature.
[13] The yields of the mixture of alkenes obtained in the re-
action with benzaldehyde, cyclohexanecarboxaldehyde
and octanal were 97%, 96% and 95%, respectively.
Phosphonate 3, the base and the solvent were mixed at 08C.
The mixture was stirred until dissolution and the aldehyde
was then added. When the transformation was performed at
1 M (KOH/THF) on a 32-mmol scale, the aldehyde was intro-
duced over 1 h (exothermic addition). The reaction was then
monitored by gas chromatography after quenching aliquots
with saturated ammonium chloride and extracting the mixture
with MTBE.
HWE Reaction with K3PO4/CH3CN at 1 M; General
Procedure
The experiments were carried out under argon in a standar-
dised 100-mL reactor equipped with a pitched blade stirrer at
800 rpm. Acetonitrile (35 mL), K3PO4 (13.90 g, 64 mmol,
2.05 equivs.) and phosphonate 3 (15.15 g, 35 mmol, 1.1 equivs.)
were first introduced at room temperature (208C). The mix-
ture was then cooled to 08C with a cooling bath and the alde-
hyde (31.8 mmol, 1 equiv.) was slowly added over 30 minutes.
The temperature was maintained at 08C and the reaction was
then monitored by gas chromatography after quenching ali-
quots with saturated ammonium chloride and extracting the
mixture with MTBE. After completionof the reaction, the salts
were first removed by filtration and the reactor was rinsed with
toluene (50 mL). The salts were then washed twice with tol-
uene (2ꢂ50 mL) and the solvent was removed under reduced
pressure.
The phosphate formed during the HWE reaction was not de-
tected by NMR in the mixture, indicating the efficiency of the
filtration step. The selectivities were then determined by GC
1
and proved to be in accordance with H NMR integration of
Adv. Synth. Catal. 2005, 347, 707–711
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
711