A convenient and mild chromatography-free method for the purification of the products of Wittig and Appel reactions

Article abstract of DOI:10.1039/c2ob07074j

A mild method for the facile removal of phosphine oxide from the crude products of Wittig and Appel reactions is described. Work-up with oxalyl chloride to generate insoluble chlorophosphonium salt (CPS) yields phosphorus-free products for a wide variety of these reactions. The CPS product can be further converted into phosphine.

Full text of DOI:10.1039/c2ob07074j

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 3531
www.rsc.org/obc
PAPER
A convenient and mild chromatography-free method for the purification of
the products of Wittig and Appel reactions†
Peter A. Byrne, Kamalraj V. Rajendran, Jimmy Muldoon and Declan G. Gilheany*
Received 12th December 2011, Accepted 5th March 2012
DOI: 10.1039/c2ob07074j
A mild method for the facile removal of phosphine oxide from the crude products of Wittig and Appel
reactions is described. Work-up with oxalyl chloride to generate insoluble chlorophosphonium salt (CPS)
yields phosphorus-free products for a wide variety of these reactions. The CPS product can be further
converted into phosphine.
the addition of a hydrocarbon solvent and a carboxylic acid to
the crude product, allowing separation of alkene (in organic
Introduction
Phosphines and reagents derived from them are ubiquitous
reagents in organic chemistry used, for example, in the Wittig
reaction,1 in Appel2 and Mitsunobu3 conditions and in the aza-
Wittig reaction.4 However the phosphine oxide by-product of
these reactions poses significant problems, both of separation
and disposal, leading to poor atom efficiency5–7 and requiring
complex procedures,7,8 particularly for large-scale reactions. Dis-
posal of phosphine oxide may be achieved by incineration to
phosphorus pentoxide and hence calcium phosphate, but more
desirable is regeneration of the phosphine.9 The BASF process
for carotenoid synthesis10 provides a good illustration of the
issues. In it, the alkene product from a Wittig reaction is separ-
ated first from the bulk of the phosphine oxide by extraction with
hydrocarbon solvents, and then scrubbed of remaining oxide in a
second extractive column with aqueous alcohol. Regeneration of
phosphine is achieved by reaction of oxide with phosgene to
give triphenylphosphine dichloride (vide infra), which is then
treated with elemental phosphorus to give triphenylphosphine
and phosphorus trichloride, the latter being also a starting
material for the former. Triphenylphosphine dichloride can also
be converted to triphenylphosphine by mixing with finely
divided aluminium metal in chlorobenzene, which forms a tri-
phenylphosphine–aluminium trichloride complex that can be
decomposed hydrolytically. Phase separation allows isolation of
triphenylphosphine from the water-soluble aluminium-containing
product, Al(OH)Cl2.9 These processes involve the handling of
hazardous reagents and products, and are far from straight-
forward. Other methods for phosphine oxide removal include
phase) and phosphine oxide (in acid phase) by phase separ-
ation.11 In some cases, the alkene can be crystallised, thus facili-
tating its separation from the other materials in the crude
product.12
On a laboratory scale, phosphine oxide removal generally
necessitates the use of column chromatography. This presents
problems if the product is sensitive to decomposition or isomeri-
sation through contact with the stationary phase or by light, or if
it is sensitive to decomposition in air. For example, certain
alkenes produced in Wittig reactions have been shown to be sen-
sitive to isomerisation by light, or by contact with silica or
alumina.12 Thus, there is a strong imperative for the development
of chromatography-free phosphine oxide removal. Reported
methods addressing this need include modification of phos-
phonium ylides to furnish a phosphine oxide by-product that
may be removed by aqueous work-up13,14 and the use of water-
soluble and hydrolytically stable ylides to allow reactions to be
conducted in water, giving water-soluble phosphine oxide and
insoluble alkene product, which precipitates and can be isolated
by filtration.15,16 An alternative approach involves polymer-sup-
ported ylides such as those of Westman17 and Leung et al.18 in
which the ylide is generated in situ from a polymer supported
phosphine. These reactions often have long reaction times and in
some cases poor yield due to problems with polymer swelling.
Some measure of success has been achieved for one-pot Wittig
reactions involving in situ generation of stabilised ylides by the
use of a bifunctional polymer containing phosphine and amine
groups.19 In all of the foregoing, the starting phosphine (free or
polymer-bound) is either more expensive than triphenylphos-
phine, and/or must be synthesised by non-trivial routes. More
seriously, the modified phosphonium entity may not have the
same reactivity profile (e.g. may give different stereoselectivity
in the Wittig), and the phosphorus substituents cannot easily be
tuned to counter this without the undertaking of further onerous
synthetic procedures.
Centre for Synthesis and Chemical Biology, School of Chemistry and
Chemical Biology, University College Dublin, Belfield, Dublin 4,
Ireland. E-mail: declan.gilheany@ucd.ie; Fax: +353-1-7162127; Tel:
+353-1-7162308
†Electronic supplementary information (ESI) available: Experimental
procedures and characterization data for alkenes, alkyl chlorides and
phosphines. See DOI: 10.1039/c2ob07074j
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 3531–3537 | 3531

View Online
A promising potential solution to the problem of phosphine
oxide removal has been the development of processes that are
catalytic in the phosphorus-containing entity. O’Brien, Chass
and co-workers have recently reported the first catalytic Wittig
reaction by using diphenylsilane to reduce the phosphine oxide,
importantly leaving the carbonyl compound unaffected.20 The
reaction requires the use of a cyclic phosphine but, being cataly-
tic, this is less problematic. Two protocols for carrying out
Appel reactions that are catalytic in phosphine oxide have been
reported recently.21–23 In the process of Denton and co-workers,
oxalyl chloride reacts with the oxide to give chlorophosphonium
salt (vide infra), which in turn reacts with alcohol to give alkoxy-
phosphonium salt, resulting in alkyl chloride and regenerating
phosphine oxide. Denton et al. have also developed a catalytic
phosphine oxide-mediated dichlorination of epoxides using
oxalyl chloride,24 and an aza-Wittig reaction that is catalytic in
phosphine oxide has also been reported.25 A recent report
focuses on utilising the phosphine oxide by-product of the
Wittig reaction as a catalyst or co-catalyst (produced in situ) in a
subsequent step of a tandem reaction sequence, thus improving
the atom economy of the overall process.6 The conversion of
phosphine oxide to phosphine has also been realised by treat-
ment with oxalyl chloride followed either by reaction of the
resulting CPS (vide infra) with aluminium metal combined with
a catalytic metal salt,26 or by electrochemical reduction of the
CPS.27
Finally, we note that it is not uncommon that the method used
to remove phosphine oxide may result in isomerisation of a Z-
alkene product of a Wittig reaction. This is the case, for example,
in the catalytic version referred to above.20 Isomerisation can be
induced by the presence of acids,28 strong bases,29 the chromato-
graphic stationary phase used,12 the solvent, heat and sunlight.
The Z-enones synthesised during this project were found to be
extremely sensitive to isomerisation, such that samples contain-
ing only Z-enone isomerised to the E-isomer when simply
allowed to stand, probably induced by light. For cases such as
these, where the alkene cannot be stored or even exposed to
chromatographic stationary phase, the development of a quick
and easy chromatography-free method of separation of alkene
and phosphine oxide is particularly important.
and covalent phosphorane forms co-exist.36 A crystal structure of
the compound produced by the reaction of Ph3P and Cl2 in
dichloromethane showed it to have a dinuclear structure
[Ph3PCl⋯Cl⋯ClPPh3]Cl·CH2Cl2.33 The compound has been
observed by us and others to have limited solubility in diethyl
ether (at room temperature)34 and THF (with the small quantity
that is soluble being in phosphorane form),34 and very low solu-
bility in alkane solvents and cold diethyl ether. This last point
provides the basis of the new work-up procedure.
Wittig Reactions
For the production of phosphorus-free alkene by the Wittig reac-
tion, the reaction itself may be carried out using whatever start-
ing materials and reaction conditions are required to give the
desired product with the desired stereochemistry. For the reac-
tions in this study, the phosphonium ylide was generated in situ
from dry precursor phosphonium salt in dry THF using
NaHMDS or KHMDS as base, followed by addition of alde-
hyde, which in general was carried out at 20 °C (in some cases
the aldehyde was added at −20 °C or −78 °C). The crude
product can then be treated in any of the following four ways to
result in the separation of phosphine oxide and aldehyde from
the alkene product:
(i) The solvent is removed in vacuo and oxalyl chloride is
added, resulting in the evolution of CO and CO2 gas and,
depending on the case, the formation of a pale yellow solid or
oil. When bubbling has ceased, cyclohexane is added, and the
upper phase formed is decanted off and then filtered.
(ii) Oxalyl chloride is added directly to the reaction mixture
(THF solvent) and stirred, causing the evolution of CO and CO2.
When effervescence ceases, the solvent is removed in vacuo,
cyclohexane is added, and the upper phase formed is decanted
off and then filtered.
(iii) The solvent is removed in vacuo, cyclohexane is added to
the crude product, and oxalyl chloride is added, causing CO and
CO2 to be produced. When effervescence ceases, the upper
phase formed is decanted off and then filtered.
(iv) As in (i), the solvent is removed in vacuo and oxalyl
chloride is added resulting in the evolution of CO and CO2
gas.37 When effervescence ceases, the mixture is cooled to
−78 °C, dry degassed Et2O is added and the supernatant solution
is removed by cannula filtration. To ensure isolation of phos-
phorus-free alkene, only ca. 85% of the supernatant is removed,
and the residue is washed three times with dry degassed Et2O.
In all four methods, the oxalyl chloride reacts with phosphine
oxide to form chlorophosphonium salt and release CO and
CO2.30 The cyclohexane used was not purified, dried or
degassed in any way prior to use. The biphasic mixture that
forms in the presence of cyclohexane consists of an upper phase
of alkene and residual oxalyl chloride dissolved in cyclohexane
(this is typically clear or pale yellow), and a solid or viscous oil
that sits at the bottom of the vessel. We found, for Wittig reac-
tions that we studied (shown in Table 1), that remaining aldehyde
starting material usually does not dissolve in the cyclohexane
phase, but stays in the residue along with the chlorophospho-
nium salt. Washing of the organic solution (after filtration) with
aqueous base and then aqueous acid results in the removal of
We now report a convenient work-up procedure for both
Wittig and Appel reactions that allows such a rapid separation of
the reaction product from phosphine oxide, and facilitates the
reduction of the latter to phosphine. In the case of the Wittig
reaction, the method is also successful in removing the aldehyde
starting material and the conjugate acid of the ylide-generating
base from the alkene and phosphine products of the process.
Our method is based on the aforementioned reaction of phos-
phine oxide with oxalyl chloride to generate chlorophosphonium
chloride.30 We were already interested in such salts because of
their involvement in other chemistry in our laboratory.31,32 They
can also be generated by the reactions of phosphine with chlor-
ine (in various solvents),33–36 and phosphine oxide with sulfuryl
chloride, methanesulfonyl chloride, thionyl chloride or trichloro-
methyl chloroformate.30,32 The structure of the entity produced
seems to depend heavily on the solvent. It adopts a chlorophos-
phonium chloride structure in acetonitrile and chlorinated sol-
vents,34 and a dichlorotriphenylphosphorane structure in benzene
and diethyl ether.34 During its hydrolysis in C6D6, both the ionic
3532 | Org. Biomol. Chem., 2012, 10, 3531–3537
This journal is © The Royal Society of Chemistry 2012

View Online
Table 1 Wittig reactions for which oxalyl chloride treatment of the
crude product followed by filtration yields alkene product free of
aldehyde and phosphorus.a See ESI† for full details of the reactions,
work-up procedures and analyses
Table 2 Appel-type reactions for which oxalyl chloride treatment of
the crude product followed by filtration yields alkyl chloride product
free of phosphorus.a See ESI† for full details of the reaction, work-up
and analyses
Carbonyl
compound
Phosphonium salt
Yieldb Alkene
(%)
#
R3
Ph
R2
R
H
R1
Z : E ratioc
1d
i-Pr
92
100 : 0
2e,f
3
4
i-Bu i-Pr
H
H
H
H
H
H
H
H
H
H
H
H
H
Ph
93
90
87
96g
96g
84
99
93
96
90
87
22 : 78
10 : 90
10 : 90
12 : 88
100 : 0
15 : 85
12 : 88
94 : 6k
17 : 83
82 : 18
6 : 94
Me
Me
CO2Et
Ph
c-C5H9
c-C5H9
c-C5H9
c-C5H9
CH3
Ph
2-FC6H4
Ph
Yield (%)
neomenthyl chloride
Yield
(%) Ph3P
5
i-Bu i-Pr
#
CPS Generation
6h
7i,j
8j
Ph
i-Pr
Me
Me
Ph
2-OMeC6H4
2-OMeC6H4
2-FC6H4
2-FC6H4
i-Pr
1
2
3
Ph3P + CCl4
Ph3P + HCAb
Ph3PO +
88
85
92
86
86c
80
9d
10e
Me
(COCl)2
11d,i Ph
2-BrC6H4
Ph
2-MeC6H4 88
12i,j
13i,j
14l
Ph
Ph
Ph
Me
Ph
CO2t-Bu
CO2t-Bu
C(O)Me
Ph
a Yields based on isolated material unless otherwise indicated. b HCA =
hexachloroacetone c Yield estimated from 1H NMR of crude product
after filtration to remove chlorophosphonium salt. This crude product
contains pentachloroacetone (by-product from chlorinating agent).
3 : 97
Ph
87
92
93
3 : 97
15
Ph CF3
Ph CF3
12 : 88k
15 : 85
16i,j
CO2t-Bu
a Reaction temperature (at which addition of aldehyde occurred) was
20 °C, and the solvent used for the filtration step was cyclohexane,
unless otherwise specified. b Based on isolated yield of alkene. c Z : E
aliphatic aldehydes, and with ketones. In all cases, comparison
of the Z : E ratio of the crude alkene product with that of the
product after treatment with oxalyl chloride showed it to be un-
affected by the procedure.38 This is particularly striking for the
cases of Z-2,2′-difluorostilbene (entry 9) and 1-cyclopentyl-3-
methylbut-1-ene (entries 5 and 6). The former is known to be
extremely sensitive to isomerisation to the E-isomer, especially if
subjected to column chromatography.12 The latter alkene (both
isomers) spontaneously degrades under ambient conditions
simply if allowed to stand. Thus neither alkene is readily accessi-
ble in pure form by methods that demand chromatography. The
use of our procedure enabled the isolation of each compound in
sufficiently pure form for full NMR characterisation, and in par-
ticular allowed the very challenging task of establishing the
stereochemistry of the isomers of 1-cyclopentyl-3-methylbut-1-
ene to be accomplished by the 1D DPFGSE-NOE39 NMR tech-
nique (see ESI† for details).
1
ratio determined by H NMR and 19F NMR (where applicable) analysis
of the crude product after aqueous work-up. d Reaction (addition of
aldehyde) carried out at −78 °C. e Filtration was carried out using dry
degassed Et2O at −78 °C. f Stated yield takes account of residual
aldehyde. g This product could not be completely purified due to
decomposition, and was contaminated with a small amount of starting
aldehyde. Complete conversion (based on consumption of phosphonium
salt) was confirmed by 1H and 31P NMR of the crude product.
h Reaction (addition of aldehyde) carried out at −20 °C. i Parent
phosphine was regenerated from the chlorophosphonium salt contained
in the residue after filtration of the alkene solution by dissolving the
residue in dry THF and treating with LiAlH4. For entry 7, MePh2P was
isolated in a yield of 98%. For entries 11, 12, 13, and 16, Ph3P was
isolated in yields of 80%, 88%, 95%, and 88% respectively.
j Phosphonium chloride salt was used as the ylide pre-cursor. k Z : E ratio
determined exclusively by 19F NMR analysis of the crude product
obtained after aqueous work-up. l Pre-formed commercial ylide
acetonylidene-triphenylphosphorane (Aldrich) was used.
In triphenylphosphine derived cases, the residue remaining
after filtration of the cyclohexane solution of alkene was shown
to contain the derived CPS by NMR (δP = 65 ppm)34 and fre-
quently also a small amount of aldehyde. On exposure to air this
residue was usually found to gradually convert to phosphine
oxide over time.
residual oxalyl chloride and also the conjugate acid of the base
used to generate the phosphonium ylide (hexamethyldisilylamine
in this case), furnishing a high yield of alkene that is free of both
phosphorus and aldehyde (Table 1). The use of certain other bases
to generate the phosphonium ylide (or indeed using a pre-gener-
ated stabilised ylide) would obviate the need for this acid wash.
From Table 1, it can be seen that this procedure was successful
for the removal of triphenylphosphine oxide, iso-butyldiphenyl-
phosphine oxide, and methyldiphenylphosphine oxide from
Wittig reaction crude products, and is applicable to the crude
products of the reactions of all phosphonium ylide types (non-
stabilised, semi-stabilised and stabilised) with both aromatic and
Appel Conditions
The applicability of this method of purification was also investi-
gated for the product of reactions under Appel and Appel-type
conditions. Denton and co-workers21,22 have used a sequence
based on the treatment of phosphine oxide with oxalyl chloride
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 3531–3537 | 3533

View Online
as the basis for their catalytic Appel reactions. They reported that
their protocol was less efficient for reactions involving secondary
alcohols.21 Therefore we chose to use menthol in Appel pro-
cesses to see if our work-up procedure might provide a comp-
lementary methodology to theirs (Table 2).
Chlorotriphenylphosphonium salt was generated in three
ways: reaction of phosphine with CCl4 (Table 2 entry 1) or hexa-
chloroacetone (HCA, Table 2, entry 2), and reaction of phos-
phine oxide with oxalyl chloride (Table 2 entry 3), all in toluene
solvent. The salt so formed was treated with menthol and
worked up by evaporation of toluene and addition of cyclo-
hexane followed by oxalyl chloride (method (iii) above). After
effervescence had ceased, the supernatant liquid was filtered by
cannula from the white CPS-containing solid and, after aqueous
work-up and evaporation of the cyclohexane, was found to
contain neomenthyl chloride in good yield (see Table 2). The
absence of any phosphorus-containing material was confirmed
by 31P NMR of the product. In the case of Table 2 entry 2, in
which HCA was used as the chlorinating agent, the neomenthyl
chloride was contaminated with pentachloroacetone.
Chart 1 Phosphine chalcogenides that could be reduced by sequential
treatment with oxalyl chloride and LiAlH4.a
The use of an alkane solvent or cold Et2O for the filtration
step after chlorophosphonium salt generation is important for the
success of the separations achieved in both the Wittig and Appel
reactions. We have found that the entity generated by treatment
of phosphine oxide with oxalyl chloride is at least sparingly
soluble in toluene, THF and Et2O (at 20 °C), perhaps due to iso-
merisation between chlorophosphonium salt and dichloropho-
sphorane.34 The procedures described above were tested with dry
degassed THF (20 °C) and dry degassed Et2O (20 °C) in place
of cyclohexane as the filtration solvent. Using these conditions,
most of the phosphorus-containing material is successfully
removed, but a small quantity of phosphine oxide is invariably
present in the product. It may be possible to extend the applica-
bility of the protocol to purification of compounds that are not
soluble in cyclohexane by using mixed solvents such as cyclo-
hexane–Et2O at room temperature.
oxygen-free atmosphere) followed by solvent removal in vacuo
furnished the phosphine product.
In cases where a small amount of phosphine oxide contami-
nant was shown to be present in the sample of phosphine,
elution through a short plug of silica using cyclohexane, ethyl
acetate or dichloromethane removed the phosphine oxide and
furnished clean phosphine. In general, it was observed that the
yield of phosphine was higher and that side-product formation
was suppressed if the degassed dichloromethane–NH4Cl solution
work-up was employed. If ordinary bench solvents containing
dissolved oxygen are used for the work-up then a very signifi-
cant amount of phosphine oxide is formed. The high yields of
regenerated phosphine from the Wittig and Appel-type reactions
(where applicable) are shown in Table 1 (entries 7, 11, 12, 13,
and 16—see footnote i) and Table 2 (all entries) respectively. For
phosphine regeneration from Wittig reactions, the best results
were obtained if the initial crude product was subjected to
aqueous work-up (before any oxalyl chloride addition) using
dichloromethane as the organic solvent. This removed inorganic
salts and products derived from the hexamethyldisilazide base,
while ensuring that all phosphine oxide (much of it derived from
the Wittig reaction, but some also from hydrolysis of unreacted
ylide or phosphonium salt) remained present in the crude
product after the aqueous work-up. In some cases, the mixture
produced by the addition of degassed ethyl acetate to the THF
solution of phosphine and LiAlH4 (after reduction of the chloro-
phosphonium salt) was observed to contain small quantities of
unreacted phosphonium salt starting material by 31P NMR. In
these instances, the mixture was filtered prior to the work-up
with degassed NH4Cl solution, since phosphonium salt hydroly-
sis would result in contamination of the phosphine product with
phosphine oxide.
Regeneration of Phosphine
With a method for isolation of the alkene in hand, we turned our
attention to the possibility of reduction of the CPS to phosphine.
We were encouraged in this because we had recently discovered
that both chlorophosphonium salts and alkoxyphosphonium salts
could be subjected to reduction and in situ boranation using
NaBH4, thus providing facile conversion from phosphine oxide
to phosphine-borane.32 Drawing on this work and the work of
Imamoto and co-workers,40 we investigated LiAlH4 as reductant,
and found it to be successful for the overall reduction of the
phosphine chalcogenides shown in Chart 1.
Finally, the separation and phosphine regeneration procedures
were combined for a number of Wittig and Appel reactions.
Thus the residue remaining after separation of the cyclohexane
phase from the insoluble material was dissolved in THF and
treated with LiAlH4 at low temperature. The reaction mixture
was quenched either by addition of degassed ethyl acetate fol-
lowed by addition of degassed NH4Cl solution or by quickly
pouring the reaction mixture into a separatory funnel containing
a degassed mixture of dichloromethane and aqueous NH4Cl
under an inert atmosphere. Standard aqueous work-up (under an
Conclusions
We have reported a short procedure that, in principle, can be
used to remove the phosphine oxide by-product from any Wittig
3534 | Org. Biomol. Chem., 2012, 10, 3531–3537
This journal is © The Royal Society of Chemistry 2012

View Online
or Appel reaction, so long as the respective alkenes or alkyl
chlorides are soluble in cyclohexane (or other alkanes) or cold
diethyl ether. It may also be applied to remove most of the phos-
phine oxide if the alkene/alkyl chloride is soluble in diethyl
ether (at room temperature) or THF, especially if the filtrations
are carried out at low temperature. The method allows the clean
isolation of alkenes that are very sensitive either to Z : E isomeri-
sation and/or polymerisation. Therefore it is potentially extre-
mely useful preparatively.
Thus the troublesome problem of phosphine oxide removal
may be easily circumvented without having to resort to modifi-
cation of the starting materials. We envisage that the method
can also be applied to other reactions where phosphine oxide
removal is a problem e.g. the Mitsunobu and Staudinger
reactions.
(10 ml) was added, and after swirling of the mixture, the liquid
was decanted off. The residue was washed a further five times in
this manner with cyclohexane (5 ml). The combined cyclo-
hexane phases were filtered through a cannula using Whatman
grade GFD glass microfibre filter paper. The filtered solution
was washed with saturated NaHCO3 solution (2 × 5 ml) and
aqueous HCl solution (1 M, 2 × 5 ml), dried over MgSO4 and
concentrated in vacuo to yield the phosphorus-free alkene
(0.15 g, 88%), as confirmed by 1H and 31P NMR. The Z : E ratio
of the alkene was determined to be 3 : 97 by comparison of the
relative integrations of the signals assigned to each isomer.
1H NMR of purified product (integrations relative to 1 H of
E-isomer): δH (500 MHz, CDCl3) 7.88 (1 H, d, J 15.9, E-H-3),
7.55–7.53 (1 H, m, E-ArH), 7.27–7.22 (contains CHCl3 signal,
m, E-ArH), 7.21–7.16 (2 H, m, E-ArH), 7.02 (0.03 H, d, J 12.1,
Z-H-3), 6.29 (1 H, d, J 15.9, E-H-2), 5.95 (0.03 H, d, J 12.1, Z-
H-2), 2.43 (3 H, s, E-CH3), 2.27 (0.1 H, s, Z-CH3), 1.54 (9 H, s,
E-C(CH3)3), 1.30 (0.27 H, s, Z-C(CH3)3).42
The Schlenk flask containing the residue from the filtration
was charged with nitrogen. Dry THF (2 ml) was added, and the
resulting mixture was cooled to −78 °C and LiAlH4 in THF (1
M, 1.2 ml, 1.2 mmol) was added dropwise, resulting in the evol-
ution of hydrogen gas and the dissolution of all of the solid
material in the THF. The reaction mixture was removed from the
cold bath and stirred for 10 minutes, then quickly poured into a
degassed mixture of CH2Cl2 (30 ml) and saturated NH4Cl sol-
ution (15 ml) contained in a separatory funnel under an inert
atmosphere. After swirling, the CH2Cl2 phase was allowed to
drain off. The aqueous phase was extracted with further degassed
CH2Cl2 (15 ml). The combined CH2Cl2 phases were dried over
MgSO4 and filtered. The solvent was removed under vacuum
yielding clean Ph3P (0.20 g, 95%), δP (121 MHz, CDCl3) −5.3
(lit.43 −4.8).
We have also noted that the phosphine used for the Wittig or
Appel reaction can be regenerated from the crude separated
chlorophosphonium salt by treatment with LiAlH4 as reductant.
This is not significant for the case of triphenylphosphine but
could be most advantageous for cases where a non-standard
phosphine has to be used in the Wittig reaction.
Experimental
Representative experimental procedures for the Wittig and Appel
reactions, phosphine oxide removal and phosphine regeneration
are given below. For details of the other reactions in Tables 1
and 2 and Chart 1 see the ESI.†
Synthesis and purification of tert-butyl 3-(2-methylphenyl)-
prop-2-enoate (Table 1 entry 13)
Reaction of triphenylphosphine oxide + oxalyl chloride +
(−)-menthol (Table 2 entry 3)
(tert-Butoxycarbonylmethyl)triphenylphosphonium
chloride
(328 mg, 0.794 mmol) and KHMDS (167 mg, 0.834 mmol)
were added to a flame-dried Schlenk flask under an atmosphere
of argon in a glove box. The Schlenk flask was attached to a
Schlenk manifold and charged with nitrogen by the standard
pump-and-fill technique.41 Dry THF (4 ml) was added, and the
resulting mixture was stirred for one hour at 20 °C. 2-Methylben-
zaldehyde (95 mg, 0.79 mmol) was added by nitrogen-
flushed syringe. The reaction mixture was stirred for 12 hours.
The THF solvent was removed in vacuo. A small sample was
removed for analysis by NMR (Z : E ratio of alkene in crude
product was 3 : 97). The residue was dissolved in CH2Cl2
(15 ml) and washed with saturated NH4Cl solution (5 ml). The
phases were separated, and the aqueous phase was washed twice
with CH2Cl2 (10 ml each). The combined isolated CH2Cl2
phases were dried over MgSO4, filtered and concentrated
in vacuo.
Triphenylphosphine oxide (1.0 g, 3.5 mmol) was dissolved in
dichloromethane (10 ml) in a flame dried Schlenk tube under an
atmosphere of nitrogen. Oxalyl chloride (0.30 ml, 3.5 mmol)
was added dropwise to the reaction mixture resulting in the for-
mation of a white precipitate (31P NMR δ 64.5 ppm).
(−)-Menthol (0.56 g, 3.5 mmol) was added to the reaction
mixture through a powder funnel, causing the reaction mixture
to become clear. The complete formation of alkoxyphosphonium
salt was confirmed by 31P NMR of a small sample of the
Oxalyl chloride (0.10 ml, 1.2 mmol) was added to the residue,
causing the formation of a pale yellow solid. Cyclohexane
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 3531–3537 | 3535

View Online
reaction mixture (31P NMR δ 58.9). The reaction was allowed to
stir for 30 minutes. Completion of the reaction was confirmed by
31P NMR analysis of a small sample of the reaction mixture (31P
NMR δ 25.2). The toluene reaction solvent was removed in
vacuo and cyclohexane (10 ml) was added to the residue. Oxalyl
chloride (0.30 ml, 3.5 mmol) was added dropwise to the reaction
mixture which caused the formation of a white precipitate of
chlorophosphonium salt. The supernatant liquid was removed by
cannula filtration. The residue was washed (2 × 10 ml) twice
with cyclohexane and the washings removed by cannula fil-
tration. The combined cyclohexane filtrate was washed with
water (20 ml × 2). The separated organic phase was then dried
over anhydrous MgSO4. Filtration to remove the drying agent
and evaporation of the solvent in vacuo yielded neomenthyl
chloride (0.49 g, 80%).
pp. 387–431; (b) R. Appel, Angew. Chem., Int. Ed. Engl., 1975, 14, 801;
(c) I. M. Downie, J. B. Holmes and J. B. Lee, Chem. Ind. (London),
1966, 900.
3 (a) O. Mitsunobu and M. Yamada, Bull. Chem. Soc. Jpn., 1967, 40,
2380; (b) T. Mukaiyama, R. Matsueda and H. Maruyama, Bull. Chem.
Soc. Jpn., 1970, 43, 1271; (c) D. L. Hughes, in Organic Reactions, ed.
P. Beak et al., Wiley, New York, 1992, vol. 42, pp. 335–656;
(d) K. C. K. Swamy, N. N. B. Kumar, E. Balaraman and K. V. P.
P. Kumar, Chem. Rev., 2009, 109, 2551; (e) O. Mitsunobu, Synthesis,
1981, 1; (f) S. Dandapani and D. P. Curran, Chem.–Eur. J., 2004, 10,
3130.
4 (a) See ref. 1b (b) F. Palacios, C. Alonso, D. Aparicio, G. Rubiales and J.
M. de los Santos, Tetrahedron, 2007, 63, 523.
5 D. J. C. Constable, P. J. Dunn, J. D. Hayler, G. R. Humphrey, J.
L. Leazer Jr., R. J. Linderman, K. Lorenz, J. Manley, B. A. Pearlman,
A. Wells, A. Zaks and T. Y. Zhang, Green Chem., 2007, 9, 411.
6 J. J. Cao, F. Zhou and J. Zhou, Angew. Chem., Int. Ed., 2010, 49, 4976.
7 See ref. 3f.
8 I. J. S. Fairlamb, ChemSusChem, 2009, 2, 1021.
9 D. Stockburger, in Production-Integrated Environmental Protection and
Waste Management in the Chemical Industry, ed. C. Christ, Wiley-VCH,
Weinheim, Germany, 1999, pp. 90–93.
10 H. Pommer and A. Nürrenbach, Pure Appl. Chem., 1975, 43, 527.
11 T. Fukumoto and A. Yamamoto, US Pat., 5292973, March 8th 1994,
Chem. Abs., 1994, 533844; CAN 121, 133844.
δH (300 MHz, CDCl3, 25 °C): 4.77 (m, 1 H), 2.02–1.91 (m, 1
H), 1.88–1.78 (m, 1 H),1.68–1.56 (m, 2 H), 1.48–1.39 (m, 2 H),
1.33–1.28 (m, 2 H), 1.04–1.01 (m, 1 H) 0.85 (d, 3JH,H = 6.4 Hz,
3
3
3 H, CH3), 0.83 (d, JH,H = 6.7 Hz, 3 H, CH3), 0.70 (d, JH,H
=
6.9 Hz, 3 H, CH3).44
12 E. C. Dunne, E. J. Coyne, P. B. Crowley and D. G. Gilheany, Tetrahedron
Lett., 2002, 43, 2449.
The solid residue remaining after cannula filtration of the reac-
tion mixture was dissolved in THF (10 ml) and cooled to 0 °C.
A solution of LiAlH4 in THF (1.0 M, 3.5 ml, 3.5 mmol) was
added dropwise to the mixture, which was then stirred for
30 minutes while warming to room temperature. Ethyl acetate
(10 ml) was added to quench the LiAlH4, followed by saturated
aqueous NH4Cl (10 ml). The reaction mixture was transferred to
a separatory funnel and the layers were separated. The aqueous
phase was washed twice more with ethyl acetate (10 ml each
wash). The combined organic phases were dried over MgSO4,
filtered through a silica plug to remove residual phosphine oxide
and concentrated in vacuo to give phosphine. (0.78 g, 86%). δP
(162 MHz): −4.5 (lit.43 −4.8).
13 J. McNulty and K. Keskar, Tetrahedron Lett., 2008, 49, 7054.
14 (a) Q. Wang, M. Khoury and M. Schlosser, Chem.–Eur. J., 2000, 6, 420;
(b) S. Trippet and D. M. Walker, J. Chem. Soc., 1961, 2130;
(c) H. Daniel and M. Le Corre, Tetrahedron Lett., 1987, 28, 1165.
15 M. G. Russell and S. Warren, Tetrahedron Lett., 1998, 39, 7995.
16 J. McNulty and P. Das, Eur. J. Org. Chem., 2009, 4031.
17 F. Westman, Org. Lett., 2001, 3, 3745.
18 P. S.-W. Leung, Y. Teng and P. H. Toy, Synlett, 2010, 1997.
19 P. S.-W. Leung, Y. Teng and P. H. Toy, Org. Lett., 2010, 12, 4996.
20 C. J. O’Brien, J. L. Tellez, Z. S. Nixon, L. J. Kang, A. L. Carter, S.
R. Kunkel, K. C. Przeworski and G. C. Chass, Angew. Chem., Int. Ed.,
2009, 48, 6836. See also ref. 8. for a review on the topic of the phos-
phine-catalysed Wittig reaction.
21 R. M. Denton, J. An and B. Adeniran, Chem. Commun., 2010, 46, 3025.
22 R. M. Denton, J. An, B. Aderinan, A. J. Blake, W. Lewis and A.
M. Poulton, J. Org. Chem., 2011, 76, 6749.
23 H. A. van Kalkeren, S. H. A. M. Leenders, C. R. A. Hommersom, F. P. J.
T. Rutjes and F. L. van Delft, Chem.–Eur. J., 2011, 17, 11290.
24 R. M. Denton, X. Tang and A. Przeslak, Org. Lett., 2010, 12, 4678.
25 S. P. Marsden, A. E. McGonagle and B. McKeever-Abbas, Org. Lett.,
2008, 10, 2589.
26 T. Yano, M. Hoshino, M. Kuroboshi and H. Tanaka, Synlett, 2010, 801.
27 T. Yano, M. Hoshino and H. Tanaka, Tetrahedron Lett., 2010, 51, 698.
28 C. Harcken and S. F. Martin, Org. Lett., 2001, 3, 3591.
29 J. Cotter, A.-M. L. Hogan and D. F. O’Shea, Org. Lett., 2007, 9, 1493.
30 (a) M. Masaki and K. Fukui, Chem. Lett., 1977, 151; (b) K. Fukui and
N. Kakeya, U. S. Pat., 4,301,301, 1981.
31 E. Bergin, C. T. O’Connor, S. B. Robinson, E. M. McGarrigle, C.
P. O ’Mahony and D. G. Gilheany, J. Am. Chem. Soc., 2007, 129, 9566.
32 K. V. Rajendran and D. G. Gilheany, Chem. Commun., 2012, 48, 817.
33 S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard and J. M. Sheffield,
Chem. Commun., 1996, 2521.
Acknowledgements
We thank sincerely the Irish Research Council for Science,
Engineering and Technology (IRCSET) for an EMBARK Scho-
larship to PAB and Science Foundation Ireland (SFI) for a scho-
larship to KVR (under grants 08/RFP/CHE1251 and 09/IN.1/
B2627). We are also very grateful to UCD Centre for Synthesis
and Chemical Biology (CSCB) and the UCD School of Chem-
istry and Chemical Biology for access to their extensive analysis
facilities and especially to Dr Yannick Ortin for NMR.
34 S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard and J. M. Sheffield,
Chem. Commun., 1998, 921.
Notes and references
35 M. A. H. A. Al-Juboori, P. N. Gates and A. S. Muir, J. Chem. Soc.,
Chem. Commun., 1991, 1271.
36 N. C. Gonnella, C. Busacca, S. Campbell, M. Eriksson, N. Grinberg,
T. Bartholomeyzik, S. Ma and D. L. Norwood, Magn. Reson. Chem.,
2009, 47, 461.
37 For cases where the Wittig reaction is performed in Et2O, it was found in
practice that the reproducibility of phosphorus removal was far superior if
the oxalyl chloride was added to the neat crude product rather than to the
solution of the crude product in Et2O.
38 The Z : E ratio of the alkene was assigned by integrationof characteristic
signals in the 1H NMR and/or 19F NMR, for both the crude and the
purified products.
1 (a) B. E. Maryanoff and A. B. Reitz, Chem. Rev., 1989, 89, 863;
(b) A. W. Johnson, Ylides and Imines of Phosphorus, Wiley, New York,
1993, pp. 221–305; (c) E. Vedejs and M. J. Peterson, in Topics in
Stereochemistry, ed. E. L. Eliel and S. H. Wilen, Wiley, New York, 1994,
vol. 21, pp. 1–158; (d) A. D. Abell and M. K. Edmonds, in Organo-
phosphorus Reagents, ed. P. J. Murphy, Oxford University Press,
Oxford, U.K., 2004, pp. 99–127; (e) R. Schobert, in Organophosphorus
Reagents, ed. P. J. Murphy, Oxford University Press, Oxford, U.K.,
pp. 12–148; (f) A. D. Abell and M. K. Edmonds, in Modern Carbonyl
Olefination, ed. T. Takeda, Wiley–VCH, Weinheim, Germany, 2004,
pp. 1–17; (g) E. Vedejs and M. J. Peterson, in Advances in Carbanion
Chemistry, ed. V. Snieckus, JAI, Greenwich, CN, 1996, vol. 2, pp. 1–85.
2 (a) R. Appel and M. Halstenberg, in Organophosphorus Reagents in
Organic Synthesis, ed. J. I. G. Cadogan, Academic Press, London, 1979,
39 DPFGSE-NOE stands for Double Pulse Field Gradient Spin Echo
Nuclear Overhauser Effect NMR spectroscopy.
3536 | Org. Biomol. Chem., 2012, 10, 3531–3537
This journal is © The Royal Society of Chemistry 2012

View Online
40 T. Imamoto, S. Kikuchi, T. Miura and Y. Wada, Org. Lett., 2001, 3, 87.
41 D. F. Shriver and M. A. Drezdzon, The Manipulation of Air Sensitive
Compounds, McGraw-Hill, New York, 2nd edn, 1986.
43 T. Ramnial, S. A. Taylor, M. L. Bender, B. Gorodetsky, P. T. K. Lee, D.
A. Dickie, B. M. McCollum, C. C. Pye, C. J. Walsby and
J. A. C. Clyburne, J. Org. Chem., 2008, 73, 801.
42 K. S. Yoo, C. H. Yoon and K. W. Jung, J. Am. Chem. Soc., 2006, 128,
16384. E-isomer of alkene.
44 J. Sćianowski, Z. Rafinśki and A. Wojtczak, Eur. J. Org. Chem., 2006,
3216.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 3531–3537 | 3537