The synthesis of stable 1,3,4-triphenylphospholes

Article abstract of DOI:10.1039/b000692k

The first syntheses of 1,3,4-triphenylphosphole, 2-bromo-l,3,4-phenylphosphole and 2,5-dibromo-l,3,4-triphenylphosphole are described. These phospholes, unlike most of the known derivatives, are difficult to oxidise and they and their corresponding oxides show little tendency to dimerise. The NBS mediated bromination of 1,3,4-triphenyl2,5-dihydrophosphole oxide led to the desired precursors of the corresponding phospholes and in one case unexpectedly yielded a phosphole oxide directly. Some reactions including the formation of organolithium reagents and their derivatisation are reported. McMurry reactions of derived formylphospholes to give coupled products are described and some copper mediated reactions are also discussed. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b000692k

View Article Online / Journal Homepage / Table of Contents for this issue
The synthesis of stable 1,3,4-triphenylphospholes
Tuula-A Niemi,a Paul L. Coe*a and Stephen J. Tillb
a School of Chemistry, The University of Birmingham, Edgbaston, Birmingham, UK B15 2TT
b DERA Malvern, Worcestershire, UK WR14 3PS
1
Received (in Cambridge, UK) 25th January 2000, Accepted 22nd March 2000
The first syntheses of 1,3,4-triphenylphosphole, 2-bromo-1,3,4-phenylphosphole and 2,5-dibromo-1,3,4-triphenyl-
phosphole are described. These phospholes, unlike most of the known derivatives, are difficult to oxidise and they
and their corresponding oxides show little tendency to dimerise. The NBS mediated bromination of 1,3,4-triphenyl-
2,5-dihydrophosphole oxide led to the desired precursors of the corresponding phospholes and in one case
unexpectedly yielded a phosphole oxide directly. Some reactions including the formation of organolithium reagents
and their derivatisation are reported. McMurry reactions of derived formylphospholes to give coupled products are
described and some copper mediated reactions are also discussed.
with excess triethylamine the phosphole oxide 1 was obtained.
Introduction
This method was subsequently adapted by Quin5 and by
Although there is a considerable literature1 concerned with five
Nakayama6 to prepare 1,3,4-triphenyl-3-phospholene-1-oxide
6 as shown in Scheme 2. Quin reacted 2,3-diphenylbutadiene
membered aromatic phosphorus heterocycles, the phospholes,
it has been found that the oxidative stability of most of the
known phospholes, except for certain penta-substituted com-
pounds, is very low. This has the consequence that because of
the formation of the phosphole oxides which readily dimerise,
there is little information on the isolation and chemistry of
stable derivatives. The first phosphole was in fact synthesised as
late as 1959.2 Most of the recent work on the chemistry of
phospholes has been carried out by the Mathey group with 3,4-
dimethyl-1-phenylphosphole and its 2-bromo derivative and
their work has been well reviewed.3 These workers were able, in
spite of the inherent problems of preparation and stability of
the starting material, to prepare a number of functionalised
compounds and to study their chemistry. We were interested
in the formation of stable phospholes which were substituted
at the 2 or 5 positions with functionality e.g. halogen which
was capable of being easily transformed in order to study
their chemistry. Westheimer4 reported the synthesis of 1-
phenoxy-3,4-diphenylphosphole 1-oxide 1 and showed that it
did not dimerise. Scheme 1 shows the reactions of 2,3-
Scheme 2
with dibromophenylphosphine in the presence of copper()
stearate followed by reaction of the intermediate phospholium
bromide 5 with aqueous sodium bicarbonate to give the desired
phosphole oxide 6. No further reactions of 6 have been
reported. Nakayama prepared 6 in a more vigorous reaction
using dichlorophenylphosphine and 2,3-diphenylbutadiene in
the presence of magnesium at 215 ЊC, and not surprisingly the
yield in this reaction was rather low and so we did not consider
it suitable for our purposes. We thus felt that a combination of
the Westheimer and Quin procedures might allow the possibil-
ity of preparing our desired 1,3,4-triphenylphosphole oxide.
Having obtained 6, which we believed would not dimerise
rapidly, we planned to convert it by the methods used by
Mathey to obtain the desired phospholes.
Results and discussion
The first problem we encountered in our synthetic plan was to
obtain 2,3-diphenylbutadiene 4. There are some reported routes
to 4 which when this work was commenced was not com-
mercially available and is still very expensive and impure. There
are a variety of routes available as shown below and we con-
sidered most of them before choosing the one we eventually
used. An obvious method to prepare 4 was to follow the prece-
dent used for the preparation of the dimethyl analogue: that
of dehydration of a suitably substituted butane-2,3-diol. This
method involves heating the diol with trace amounts of fused
potassium hydrogen sulfate. A worrying feature of these
descriptions (see below) seemed to be the variable yield
obtained (between 15 and 100%), depending on the author,
using essentially the same chemistry. Hayden and Alder7
Scheme 1
diphenylbutadiene with 2-chloroethyl dichlorophosphinite
and phosphorus trichloride and then that of the intermediate
1-chloro-3,4-diphenyl-2,5-dihydrophosphole 1-oxide 2, without
isolation, with phenol to give the phenoxy derivative 3. By a
sequence involving bromination and then dehydrobromination
DOI: 10.1039/b000692k
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528
This journal is © The Royal Society of Chemistry 2000
1519

View Article Online
claimed almost quantitative yields of the diene and so we
investigated their reaction conditions. In our hands, although
we were able to obtain very high carbon recoveries, we always
obtained the pinacolone rearrangement product 3,3-diphenyl-
butanone as by far the major product. Even under our best
conditions the yield of the diene never exceeded 25% of the
product and was usually not more than 10% in accord with the
results described by Dodson et al.8 In contrast Brettle9 reported
a 68% yield for the same reaction. Clearly there is an inherent
difficulty, which we could not master, in this reaction. An alter-
native route appeared to be the Wittig reaction of methylene-
triphenylphosphorane with benzil, as described by De Groot
et al.,10 but the yield here was only 15% and involved somewhat
tedious column chromatography to remove the triphenyl-
phosphine oxide formed in the reaction. Since we require
multi-gram quantities of the products this was not a viable
alternative.
A second possibility was the coupling of α-bromostyrene
using either a nickel or palladium catalyst. Such reactions have
been reported, the former by Caubere et al.11 and the latter by
Luo.12 The former report suggested to us that although the
desired diene could be obtained on a small scale the reaction
was very sensitive to the method of catalyst preparation, to
profound solvent effects, which led to an alternative product,
and to problems of scale. The Luo method again appears to be
attractive but we were concerned with the scale up and the high
cost of the catalyst. Thus, we needed to find an alternative route
which overcame these objections.
Fig.
1 X-Ray structure of 2,5-dibromo-1,3,4-triphenylphosphole
1-oxide.
functionalised phospholes we decided to attempt to prepare the
required bromodihydrophosphole oxides directly rather than in
a stepwise manner. Thus, we adapted Westheimers method
using 2 equivalents of NBS in the hope of preparing the
2,5-dibromo derivative in one step. The reaction of 6 with
NBS in DCM using purified benzoyl peroxide as catalyst at
40 ЊC afforded a mixture of three products in the ratio 1.4:1:1
(see Scheme 4). The mixture was readily separated by column
We eventually discovered a report in the Japanese patent
literature by Kondo13 who had shown that the pinacol–
pinacolone rearrangement could be avoided by using 2,3-
diphenyl but-3-en-2-ol 8 as starting material. We were able to
prepare 8 by reaction of 2-phenylethenylmagnesium bromide
(α-styrylmagnesium bromide) with acetophenone in 77%
isolated yield. Although 2-bromo-2-phenylethene is com-
mercially available we found that most commercial samples are
highly contaminated with the polymeric material and are very
difficult to purify. We thus used freshly prepared material14 in
all of our reactions. The dehydration of 8 was readily achieved
by heating the alcohol with freshly fused potassium hydrogen
sulfate at 120 ЊC followed by vacuum distillation of the diene 4
from the reaction mixture, as shown in Scheme 3. Having solved
Scheme 4
chromatography. The products were identified, using
a
combination of 1H, 13C and 31P NMR spectroscopy, mass
spectrometry and elemental analysis, as trans,trans-2,5-di-
bromo-1,3,4-triphenyl-2,5-dihydrophosphole oxide 9, cis,trans
2,5-dibromo-1,3,4-triphenyl-2,5-dihydrophosphole oxide 10,
and 2,5-dibromo-1,3,4-triphenylphosphole oxide 11. It was
possible to distinguish the pairs of diastereoisomers 9 and 10 by
means of the 1H and 13C spectra since the trans,trans isomer has
a plane of symmetry giving only therefore one NMR signal for
the 2 and 5 protons and also for the 2 and 5 carbon atoms.
Scheme 3
this problem we were able to repeat the reaction of 4 with
dibromophenylphosphine under Quins conditions to obtain the
desired phospholene oxide 6. Unfortunately, as was found by
Quin, the yields in the reaction were rather low. The phosphol-
ium bromide 5 crystallises from the petroleum ether solvent on
cooling and we found that by cropping this product each day
for up to ten days we could obtain an overall yield of 6,
after hydrolysis of 5, of 71%. We also found that the reaction
could be substantially scaled up and this allowed us to obtain
abundant supplies of the phosphole oxide 6 (25% overall from
styrene).
2
Further, there is only a small JPHα_coupling characteristic of
trans,trans isomers.15 The cis,trans isomer does not have the
symmetry plane and shows two signals for the 2 and 5 protons
and carbon atoms. The dibromophosphole oxide 11 was a
totally unexpected product and to confirm the structure we
carried out an X-ray analysis. The structure is shown in Fig. 1.
As can be seen the three phenyl rings and the oxide oxygen are
almost orthogonal to the planar phosphole and this fits well
with energy minimised MOPAC calculations.
The next stage in our synthesis was the conversion of 6 to
1,3,4-triphenylphosphole oxide using the Westheimer procedure
with NBS and base. Since our major objective was to prepare
This result raises some interesting mechanistic questions and
we decided to try to find an answer to the possible pathway
leading to the phosphole oxide 11. We carried out a series of
1520
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528

View Article Online
Scheme 5
Table 1 The reaction of 1,3,4-triphenylphospholene 1-oxide 6 with
N-bromosuccinimide
(β-cyanoethyl)phosphine to afford the corresponding phosp-
holes.15 During the course of this study Mathey showed that it
was possible to deoxygenate the phosphole oxides directly, in
good yield, by reaction with trichlorosilane in the presence of
pyridine,16 a reaction known to reduce phosphine oxides to
phosphines. We have now compared both methods using our
phosphole derivatives with some surprising results. The first
reactions we tried were the conversions of the dibromodihydro-
phosphole oxides 9 and 10 to the sulfides 13 and 14 respectively
(Scheme 6). We found that each diastereoisomer reacted
Yield (%)
Reaction
time/h
Entry
NBS
(PhCO)2O2
9
10
11
30.5
21
29
1
2
3
4
5
2.0 equiv.
2.0 equiv.
2.0 equiv.
3.0 equiv.
3.0 equiv.
cat.
18
18
18
18
240
55.5
59
57
14
20
14
12
excess
none
cat.
50
38
100
none
reactions varying the amount of catalyst to NBS ratio, and the
NBS to substrate ratio. The results are shown in Table 1 from
which it is clear that the amount of NBS used is important
in determining the products formed and that the catalyst, as
expected, merely speeds up the reaction. Interestingly the
results also show that either the rate of formation of 9 is faster
than that of 10 or that the rate of decomposition of 10 to 11 is
faster than 9 to 11. We believed from our original experiment,
where after 6 h the ratio of 9:10 was 1.4:1, that 10 reacted to
give 11 faster than 9. In a separate experiment using a 1:1
mixture of 9 and 10 and 2 equivalents of NBS the ratio of 9:10
was 1.6:1 after 12 h, the only other product being 11. The
proposed mechanism is shown in Scheme 5. We were unable to
obtain any evidence for the formation of the proposed tribromo
intermediate from NMR spectra of the reaction mixture and we
assume that the rate of dehydrobromination to 11 is very fast.
We have so far been unable to carry out the reaction within the
NMR probe. From a synthetic point of view however these
results were extremely useful. By using three equivalents of
NBS we were able to prepare multi-gram quantities of 11. Simi-
larly the use of two equivalents of NBS afforded a mixture of 9
and 10 and the use of one equivalent of NBS, after a simple
column separation, gave 2-bromo-1,3,4-triphenyl-2,5-dihydro-
phosphole oxide 12. The latter was characterised by standard
physical methods.
Scheme 6
smoothly in virtually quantitative yield to give the correspond-
ing sulfide with retention of configuration at phosphorus as was
reported by Mathey15 for the formation of the dimethyl com-
pound 15. The assignment of structure was based on the small
αH to P couplings characteristic of the trans arrangement and
on the appearance of only one signal in the 1H and 13 C NMR
spectra of the symmetrical trans,trans compound as outlined
above for the starting oxide.
We now had three potential precursors to the desired phos-
pholes. These results are shown in Scheme 5. Two problems now
remained to be solved, firstly to convert the dihydrophosphole
oxides to phosphole oxides and then to convert these by a
deoxygenation process to the phospholes. When we began this
work the favoured method of carrying out the latter process
was to convert the dihydrophosphole oxides by reaction with
phosphorus pentasulfide to the corresponding sulfides. After
dehydrobromination the resulting phosphole sulfides were
then treated with either triphenylphosphine or better with tris-
We next investigated the dehydrobromination of 13 and 14.
Westheimer4 had used triethylamine as the base to obtain 1
whereas Mathey15 had used methanolic potassium hydroxide
to dehydrobrominate 2,5-dibromo-3,4-dimethyl-1-phenyl-2,5-
dihydrophosphole 1-sulfide 15. We tried both of these reagents
quite successfully with the methanolic potassium hydroxide
giving the higher yield. The reaction of the trans,trans isomer
13 was much faster than the cis,trans derivative 14 (30 min as
against 4 h for completion of the reaction). Both 13 and 14
gave exactly the same product, as would be expected, namely
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528
1521

View Article Online
2-bromo-1,3,4-triphenylphosphole 1-sulfide 16. We were now
concerned with removal of the sulfide, and first tried the de-
sufurisation with triphenylphosphine which was quite success-
ful, all of the starting material being consumed as shown by
31P NMR spectroscopy, but unfortunately it was impossible to
separate the phosphole from unreacted triphenylphosphine.
We were, however, able to carry out a very straightforward des-
ulfurisation using tris(β-cyanoethyl) phosphine which afforded
2-bromo-1,3,4-triphenylphosphole 17 in 88% yield. These
results are shown in Scheme 6. The phosphole was character-
ised by physical means. The new phosphole showed no
tendency to dimerise and was found to oxidise only very slowly.
A sample kept in air afforded only a few percent of oxide after
six months and a sample in dry nitrogen appears to be totally
unchanged after two years. The small amount of oxide present
in the sample, unlike almost all other phosphole oxides
reported, does not appear to have dimerised.
the desired bromophosphole oxide 20 and 19 suggesting there is
indeed a competition between the two reactions. We were also
able to show that bromination of 20 with elemental bromine
followed by dehydrobromination in a “one pot” reaction
yielded the dibromophosphole oxide 11 identical to that
obtained from the NBS reaction of 6 above (Scheme 5).
To complete our objectives we wished to make 1,3,4-tri-
phenylphosphole. This should be available from the deoxygen-
ation of the corresponding phosphole oxide, which in turn
could be prepared by dehydrobromination of the dihydro-
phosphole oxide derivative 12. Reaction of the latter with
methanolic potassium hydroxide readily afforded the required
phosphole oxide in good yield (Scheme 9). We found no evi-
At this time in our work an improved deoxygenation route
described above became available and we decided to use this
method instead of the somewhat unpleasant sulfide route.
Thus, we reacted the phosphole oxide 11 with trichlorosilane
using the Mathey conditions (Scheme 7) and obtained an 81%
Scheme 9
dence for the nucleophilic substitution reaction which we had
observed with the dibromo analogue above. The phosphole
oxide, which was again characterised by the usual physical
means, showed no evidence of rapid dimerisation and
even after one year only trace amounts of the dimer were
evident from the 31P NMR spectrum. Deoxygenation of
the phosphole oxide with trichlorosilane–pyridine gave 1,3,4-
triphenylphosphole 21 in 79% yield. The phosphole was charac-
terised in the usual way. We had thus achieved the first objective
of the study with the preparation of the three stable phosp-
holes. It is important to note the crucial role of 31P NMR
spectroscopy in structural determination in this area. The
dihydrophosphole and phosphole sulfides were generally
observed to show phosphorus resonances at 55–60 ppm and ca.
50 ppm respectively, whereas the corresponding oxides had
resonances at ca. 40 and 35 ppm. The phospholes all showed
resonances at rather lower fields depending on their substitu-
tion but in the range of 5–22 ppm. This enabled us to monitor
reactions very easily and also to check on the ease of oxidation
or dimerisation of the products so that appropriate precautions
could be taken to counter these processes. A similar set of reac-
tions of the dihydrophosphole oxide 6 but with NIS failed to
yield the corresponding iodo compounds, only the starting
material being recovered.
We were now in a position to study further functionalisation
of the phospholes to enable us to proceed towards our objective
of synthesising macrocyclic derivatives. In the light of the work
of Mathey1 we first studied the formation and reactions of
organolithium derivatives as a means of functionalising our
phospholes. Since the dibromophosphole 18 was more readily
available and could possibly lead more directly to our objective,
we concentrated our initial studies on this derivative. We tried
a number of different conditions for the halogen–metal inter-
change and these results are summarised in Table 2. Aliquots
were taken at the reaction temperature indicated and the result-
ing solutions were examined by 31P NMR spectroscopy to
determine the approximate product ratios. We were somewhat
surprised to find that the reaction apparently did not take place
at some temperature between Ϫ75 and Ϫ30 ЊC and even at
Ϫ75 ЊC did not go to completion, whereas at Ϫ100 ЊC complete
conversion to the lithio compound was seen. As yet we have no
real explanation for this observation: it would imply the usual
halogen–metal interchange is reversed with increasing temper-
ature. In a simple experiment we found that hydrolysis of a
sample taken from a solution taken at Ϫ100 ЊC showed com-
plete conversion to the lithium derivative. After warming the
mixture to Ϫ75 Њ C and allowing it to stand at this temperature
Scheme 7
yield of a pale yellow solid which was shown by physical
methods to be the desired 2,5-dibromo-1,3,4-triphenylphos-
phole 18. This phosphole was also quite stable; it did not readily
oxidise in air, but as with 17, a sample left in air for some
months did show some oxidation, but in a nitrogen atmosphere
it seems to be unchanged after at least one year. In principle we
should have been able to prepare 17 directly via its oxide, which
in turn should be available from dehydrobromination of 9 and
10 as for the sulfide. Thus, we reacted 9 and 10 as separate
diastereoisomers. In each case we obtained largely one product
which showed a band at 31.5 ppm in its 31P NMR spectrum.
This was unlike any other product we had found previously in
1
this series. The H spectrum showed a doublet signal in the
methoxy region with a proton–phosphorus coupling of 11 Hz.
This suggested we had replaced a bromine atom with a methoxy
group in a simple nucleophilic displacement process. This was
confirmed by mass spectrometry, which showed the character-
istic isotopic distribution for a monobromo- compound. This
confirmed the product as 2-bromo-5-methoxy-1,3,4-triphenyl-
2,5-dihydrophospholene oxide 19. Thus, we were unable to pre-
pare 17 from the oxide (Scheme 8). This change in reactivity
Scheme 8
on going from the phosphole sulfide to the oxide was quite
unexpected from previous literature results. We presume that
this is due to the greater electronegativity of the oxygen atom
on phosphorus making the α-carbon atom more positive and
hence more reactive towards nucleophiles and therefore allow-
ing this reaction to compete with dehydrobromination.
In a separate experiment using a mixture of 9 and 10 under
slightly milder conditions we were able to obtain a mixture of
1522
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528

View Article Online
Table 2 The reaction of 2,5-dibromo-1,3,4-triphenylphosphole with
organolithium reagents
compared to two for the other route. It may be possible to
distinguish between the two pathways therefore by converting
28 to the corresponding chloride by exchange with PCl3 and
then to study the rate of its reaction with bromide ion using 31
P
Yield (%)
NMR spectrocopy. Equally the exchange reaction of PhPCl2
with bromide could be studied in the same way. Somewhat to
our surprise we have so far been unable to react 22 successfully
with carbon dioxide or with acyl halides.
Organolithium
T/ЊC
Ϫ100
Ϫ75
Ϫ30
20
Ϫ100
Ϫ75
Ϫ30
20
Ϫ100
Ϫ75
Ϫ30
20
18
17
21
n-BuLi
n-BuLi
n-BuLi
n-BuLi
MeLi–LiBr
MeLi–LiBr
MeLi–LiBr
MeLi–LiBr
tert-BuLi
tert-BuLi
tert-BuLi
tert-BuLi
15
85
50
none
none
none
none
none
none
none
none
20
50
100
100
94
Having obtained the aldehydes 23 and 24 we were now in a
position to attempt the McMurry reaction and we chose to do
this on the more readily available aldehyde 24. We tested out the
reaction conditions using furfural and were able to obtain the
desired coupled product in 93% yield as a single isomer shown
to be the E isomer in agreement with the literature.18 A similar
reaction with the aldehyde 24 afforded a pale orange solid in
71% yield which was characterised by physical means and was
shown to be the expected ethene 29. Both the 1H and 31P NMR
spectra revealed the presence of two diastereoisomeric mixtures
in the ratio of appoximately 90:10 due to the E and Z arrange-
ments of the phenyl groups attached to phosphorus. The
double bond was in the E-configuration, in common with the
furfural derived product. The mass spectrum showed an inter-
esting feature which was very useful in confirming the structure.
The mass of the molecule was so high (it contained more than
40 carbon atoms) that it was possible to distinguish not only the
bromine isotope ratios but also the natural abundance ratios of
the carbon atoms. The observed pattern was a complete match
to the computer generated unique pattern for the molecule. As a
further check on the method we prepared the corresponding
3,4-dimethyl derivative 30 from 2-bromo-3,4-dimethylphos-
phole-5-carbaldehyde16 which was again a diastereomeric mix-
ture with similar NMR characteristics to those found for 29.
We next investigated some copper coupling reactions of the
lithium reagent 22 and the dibromophosphole 18. In the light
of the work of Mathey15 we reacted 22 with copper() chloride
and obtained a mixture of products in moderate yield which
could be separated by column chromatography to yield the
expected dimer 31 (33%) the reduced dimer 32 (19%) unreacted
starting phosphole 18 (12%) and the phosphole (17) formed
from the hydrolysis of 22. The new products were characterised
by physical means. Reaction of 18 with copper() cyanide under
a variety of conditions always afforded mixtures of products
which we could not completely separate, but from which we
could obtain a mixture of the dimers 31 (50%) and 32 (16%) as
shown by GC/MS and a small (19%) yield of the desired nitrile
33, which was partially characterised by physical means. The
formation of coupled products in these copper assisted reac-
tions are quite commonly found19 and it would appear in our
case this reaction is the favoured one. These reactions are
summarised in Scheme 10. We have thus shown that the new
phospholes lead to a number of derivatives which have scope
for further developments in this interesting area of chemistry.
none
none
6
1.5
none
none
16
16
5
none
98.5
100
100
64
68
90
16
5
none
100
for a short time before sampling, starting material was present
to approximately 50% and on repeating this process at Ϫ20 ЊC
only starting material could be detected. A repeat of this
experiment confirmed this result. We also carried out the same
type of reaction using methyllithium–lithium bromide, tert-
butyllithium and methylmagnesium bromide. We found that in
general these reagents were less reactive than n-butyllithium but
when reaction did occur the same unusual temperature effects
were observed. Further study on this reaction may clear up this
anomaly.
As we could carry out the interchange to yield the lithium
reagent 22 quite successfully at Ϫ100 ЊC we proceeded with
reactions at this temperature. Since we planned to try to join
phospholes together with unsaturated carbon chains, we felt an
obvious route was to use the McMurry reaction which has been
applied very successfully in the preparation of macrocyclic
furan derivatives.17 Thus, we required the appropriate carbald-
ehydes for this coupling. We found the best conditions for the
reaction were to treat the lithium reagent 22 with a slight excess
of N-methylformanilide at Ϫ100 ЊC in THF as solvent. In this
way the bromophosphole 17 and the dibromophosphole 18
afforded the corresponding aldehydes 23 and 24 in fair yield. As
was discovered by Mathey16 for the corresponding 3,4-dimethyl
compound, the dibromophosphole did not form the dilithio
derivative. We found that under similar conditions 22 reacted
with propanal to yield the expected alcohol 25, but attempts to
isolate this material completely pure all yielded only the
dehydrated product 26. These compounds were characterised
by physical means. Reaction of 22 with iodine afforded an
inseparable mixture of the starting material 10% and the
bromoiodo compound 27 90%. These could be readily identi-
fied by GC/MS. Reaction of 22 with phenyldichlorophosphine
afforded a mixture of two products. The major component was
the phosphine 28, identified by physical means, but we were
unable to identify the minor component. The composition of
the phosphine 28 at first sight was unexpected but this can be
readily rationalised in terms of the known halogen exchange
reactions between phosphines. There are two possibilities which
could account for the formation of 28: either the phenyl-
dichlorophosphine exchanges chlorine for bromine with the
bromide ions present in the butyllithium, and the bromo-
phosphine then reacts with 22, or more likely the chloro-
phosphine formed by straightforward reaction of the
dichlorophenylphosphine with 22 exchanges with lithium
bromide. A suggested method of distinguishing between these
possibilities has been made† as follows: if the exchange pro-
ceeds to produce 28 via addition to give a phosphoranide
intermediate the first pathway would be expected to be faster
since the phosphorus atom would bear three halogen atoms
Experimental
NMR spectra were recorded using a Bruker AC-300 spec-
trometer. TMS was used as the internal standard for 1H (300.13
MHz) and 13C (75.47 MHz) spectra. H3PO4 (85%) was used as
external reference for 31P (121.50 MHz) spectra. Mass spectra
were recorded on a VG Prospec triple focussing spectrometer.
X-Ray spectra were recorded using a Rigaku Raxis 2C spec-
trometer. Matrix Silica 60 (35–70 micron) was used for column
chromatography (Fisher Scientific Co.). SilicaGel 60 covered
aluminium sheets were used for TLC.
Preparation of 2,3-diphenylbut-3-en-2-ol 8
1-Bromo-1-phenylethene (82.4 g, 0.45 mol) in dry ether (150
cm3) was added dropwise at such a rate to maintain gentle reac-
tion (over 2.5 h) to a suspension of magnesium turnings (13.1 g,
† We thank a referee for this suggestion.
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528
1523

View Article Online
Scheme 10 Reagents: i, BuLi; ii, PrCHO; iii, silica; iv, PhNHCHO; v, TiCl4–Zn; vi, CuCl; vii, CuI; viii, PhPCl2; ix, CuCN.
0.54 mol) in ether (30 cm3). When the addition was complete
the mixture was stirred for 40 min and then acetophenone (52.5
cm3, 0.45 mol) was added over 2 h. The reaction mixture was
stirred for a further 4 h when saturated ammmonium chloride
solution (70 cm3) was added and the mixture was neutralised
with HCl (3 M, 130 cm3). The ether layer was separated and
combined with the ether extracts of the aqueous layer (5 × 200
cm3). The dried (MgSO4) ether layers were evaporated and the
residue distilled in vacuo to yield 2,3-diphenylbut-3-en-2-ol 8
(77.5 g, 76%), bp 127–130 ЊC/0.5 mmHg (lit.,8 127–148 ЊC/0–1
Preparation of 1,3,4-triphenyl-2,5-dihydrophosphole 1-oxide 6
The butadiene 4 (25.5 g, 0.12 mol), copper() stearate (0.25 g)
and dibromophenylphosphole (40 g, 0.15 mol) in petroleum
ether (bp 60–80 ЊC, 170 cm3) were stirred at 60 ЊC for 16 days
but after each day the precipitated product was filtered off. The
combined precipitates were treated with ice cold saturated
sodium bicarbonate solution (150 cm3) The aqueous layer was
extracted with DCM (6 × 50 cm3), the combined extracts were
dried (MgSO4) and the solvent evaporated to yield 1,3,4-
triphenyl-2,5-dihydrophosphole 1-oxide 6 (29.3 g, 71%), mp
160–162 ЊC (from CCl4) (lit.,6 162–165 ЊC); δH (CDCl3) 7.9–7.8
(m, 2H, PhP), 7.6–7.5 (m, 3H, PhP), 7.3–7.1 (m, 10H, PhC),
3.5–3.2 (m, 4H, CH2); δP (CDCl3) 48.1; MS m/z 330 [Mϩ].
mmHg); δH (CDCl ), 7.4 (m, 10H, 2Ph), 5.7 (m, 2H, ᎐CH), 2.4
᎐
3
(s, 1H, OH), 1.8 (s, 3H, CH3); MS m/z 224 [Mϩ].
Preparation of 2,3-diphenylbuta-1,3-diene 7
Reaction of 1,3,4-triphenyl-2,5-dihydrophosphole 1-oxide with
N-bromosuccinimide
Compound 8 (64.9 g, 0.3 mol), freshly fused potassium hydro-
gen sulfate (1.5 g) and phenyl-β-naphthylamine (0.04 g) were
heated together for 40 min at 120 ЊC at 7–8 mmHg and the
mixture was distilled at 0.01 mmHg to yield 2,3-diphenylbuta-
1,3-diene 7 (27.3 g, 46%), mp 41–43 ЊC (lit.,8 46–47 ЊC) (from
methanol); δH (CDCl3) 7.2–7.5 (m, 10H, 2Ph), 5.6 (d, 2JHH 1.5,
2H, ᎐CH), 5.4 (d, 2J 1.5, 2H, ᎐CH); MS m/z 206 [Mϩ].
The phospholene oxide 4 (1.3 g, 3.9 mmol), N-bromo-
succinimide (1.4 g, 7.9 mmol) and benzoyl peroxide (0.15 g) in
chloroform (35 cm3) were heated together under reflux for 6 h.
The mixture was cooled to room temperature and the solvent
evaporated. The residue was shaken with a small amount of
᎐
᎐
HH
1524
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528

View Article Online
DCM and the resulting suspension filtered. The filtrate was
passed through a column of silica (eluant ethyl acetate–hexane
1:1) to yield three products identified as (i) trans,trans-2,5-
dibromo-1,3,4-triphenyl-2,5-dihydrophosphole 1-oxide 9 (0.34
g) as a gum (Found: C, 53.8; H, 3.4%. C22H17Br2OP requires C,
54.1; H, 3.5%); δH (CDCl3) 7.9 (m, 2H, PhP), 7.6 (m, 1H, PhP),
Cipso-PhC), 133 (Ar), 130.7 (d, 2JCP 8.5, PhP), 129.4 (Ar), 128.7
1 1
(Ar), 128.5 (Ar), 50.9 (d, JCP 69, CHBr), 35.3 (d, JCP 68.4,
CH2); δP (CDCl3), 41.5; MS m/z 410/408 (1:1) [Mϩ], 328
[M Ϫ Brϩ].
Synthesis of trans,trans-2,5-dibromo-1,3,4-triphenyl-2,5-
dihydrophosphole 1-sulfide 13
2
7.5 (m, 2H, PhP), 7.3 (m, 10H, PhC), 5.0 (d, 2H, JHP 3.3,
1
CHBr); δC (CDCl3), 141.2 (d, JCP 12.4, Cipso-PhP), 135.1 (d,
The phospholene oxide 9 (0.7 g, 1.43 mmol) and phosphorus
pentasulfide (0.32 g, 0.73 mmol) were refluxed together in
toluene–dichloromethane (50 cm3, 1:1) for 5 h. The mixture
was filtered through Celite and the solvents evaporated to
yield, as a single component by TLC analysis, trans,trans-
2,5-dibromo-1,3,4-triphenyl-2,5-dibromophosphole-1-sulfide 13
(0.72 g, 99%) as a gum (Found mass 501.913224. Required mass
(C22H1779Br2PS) 501.915533); δH (CDCl3) 8.0 (m, 2H, PhP), 7.6
(m, 1H, PhP), 7.5 (m, 2H, Ph-P), 7.3 (m, 10H, PhC), 5.2 (d, 2H,
2
2JCP 10.2, PhC᎐C), 133.6 (Ar), 131.1 (d, J 8.8, Cortho-PhP),
᎐
CP
1
129.2 (Ar), 127(Ar), 45 (d, JCP 70.1, CHBr); δP (CDCl3) 28.9;
MS m/z 490/488/486 (1:2:1) [Mϩ], 409/407 (1:1) [M Ϫ Brϩ],
328 [M Ϫ Br2ϩ]; (ii) cis,trans-2,5-dibromo-1,3,4-triphenyl-2,5-
dihydrophosphole 1-oxide 10 (0.25 g) as a gum (Found: C, 54.1;
H, 3.2%. C22H17Br2OP requires C, 54.1; H, 3.5%); δH (CDCl3)
7.9 (m, 2H, PhP), 7.6 (m, 1H, PhP), 7.5 (m, 2H, PhP), 7.2 (m,
2
6H, PhC), 7.1 (m, 4H, PhC), 5.5 (d, JHP 10.7, CHBr), 5.2 (s,
1H, CHBr); δC (CDCl3) 139.8 (d, 1JCP 11.9, Cipso-PhP), 134.9 (d,
2JHP 2.2, CHBr); δC (CDCl3) 142.3 (d, 2JCP 9, PhC᎐C), 135.2 (d,
᎐
2JCP 9, PhC᎐C), 134.4 (d, J 9, PhC᎐C), 133.7 (Cortho-PhC),
2
2
1JCP 8.5, Cipso-PhP), 133.1 (Cortho-Ph-C), 131.5 (d, JCP 9.6,
᎐
᎐
CP
2
1
132.1 (d, JCP 8.8, Cortho-PhP), 128.9 (Ar), 127.8 (Cipso-PhC),
46.3 (d, 1JCP 65.6, CHBr), 45.3 (d, 1JCP 67.3, CHBr); δP (CDCl3)
28.9 (m); MS m/z 490/488/486 (1:2:1) [Mϩ], 409/407
(1:1) [M Ϫ Brϩ], 328 [M Ϫ Br2]; and (iii) 2,5-dibromo-1,3,4-
triphenylphosphole 1-oxide 11 (0.26 g), mp 210 ЊC (decomp.)
(Found: C, 54.1; H, 3.3%. C22H15Br2OP requires C, 54.4; H,
3.1%); δH (CDCl3) 8.0–7.9 (m, 2H, PhP), 7.7–7.6 (m, 1H, PhP),
Cortho-PhP), 129.8 (Cipso-PhC), 129.2 (Ar), 51.3 (d, JCP 50.9,
CHBr); δP (CDCl3) 56.5; MS m/z 506/504/502 (1:2:1) [Mϩ],
425/423 (1:1) [M Ϫ Brϩ], 344 [M Ϫ Br2ϩ].
Synthesis of cis,trans-2,5-dibromo-1,3,4-triphenyl-2,5-dihydro-
phosphole 1-sulfide 14
In the same way as above, except that reflux was only necessary
for 90 min, the phospholene oxide 10 (0.8 g) afforded cis,trans-
2,5-dibromo-1,3,4-triphenyl-2,5-dihydrophosphole 1-sulfide 14
(0.75 g) as a gum (Found mass 501.91438. Required mass
(C22H1779Br2PS) 501.915533); δH (CDCl3), 8.2 (m, 2H, PhP),
7.6–7.5 (m, 2H, PhP), 7.3–7.2 (m, 6H, PhC), 7.1–7.0 (m, 4H,
2
PhC); δC (CDCl3) 153.2 (d, JCP 26, PhC᎐C), 133.5 (PhP), 133
᎐
1
2
(d, JCP 11.9, Cipso-PhP), 131.4 (d, JCP 10.5, Cortho-Ph-P), 129.1
(Ar), 128.9 (Cmeta-PhC), 124.4 (d, 1JCP 110.2, Cipso-PhP), 115.3 (d,
1JCP 102.3, CBr); δP (CDCl3), 32.4 (m); MS m/z 488/486/484
(1:2:1) [Mϩ], 407/405 (1:1) [M Ϫ Brϩ].
2
7.6 (m, 3H, PhP), 7.3 (m, 10H, PhC), 5.5 (d, 2H, JHP 6.9,
1
CHBr), 5.4 (s, 1H, CHBr); δC (CDCl3), 140.8 (d, JCP 9, Cipso
-
2
2
PhP), 135.2 (d, JCP 8.5, PhC᎐C), 134.6 (d, J 8.5, PhC᎐C),
133.3 (Cortho-Ph-C), 132.4 (d, JCP 9.6, Cortho-PhP), 129.2 (Cipso
PhC), 128.9 (Ar), 52.4 (d, JCP 58, CHBr), 51.8 (d, JCP 50.9,
CHBr); δP (CDCl3) 62; MS m/z 506/504/502 (1:2:1) [Mϩ],
425/423 (1:1) [M Ϫ Brϩ], 344 [M Ϫ Br2ϩ].
᎐
᎐
CP
X-Ray data for compound 11.‡ Empirical formula: C22H15-
Br2OP. M = 486.13. Crystal system orthorhombic, unit cell
dimensions a = 12.6240 Å, b = 24.696 Å, c = 6.000 Å, V = 1870
Å3, Z = 4, Dc = 1.726 mg mϪ3, µ = 4.428 mmϪ1. Crystal size
0.50 × 0.14 × 0.10 mm. F(000) 960. θ range for data collection
1.65 to 26.34Њ. Index ranges Ϫ15 ≤ h ≤ 15, Ϫ29 ≤ k ≤ 27,
Ϫ7 ≤ l ≤ 7. Reflections collected 11014. Independent reflections
3417 [R(int) = 0.422]. Refinement method full matrix least
squares on F 2. Data/restraints/parameters 3417/0/235. Good-
ness of fit on F 2 1.069. Final R indices [I > 2E(I)]. R1 = 0.0366,
wR2 = 0.0821. R indices all data R1 = 0.0422, wR2 = 0.0821.
2
-
1
1
Synthesis of 2-bromo-1,3,4-triphenylphosphole 1-sulfide 16
To the phospholene sulfide 13 (0.8 g, 1.58 mmol) in DCM (15
cm3) KOH (0.18 g, 3.17 mmol) in methanol (4 cm3) was added
dropwise over 20 min. The mixture was stirred for a further 30
min (in a similar experiment for 4 h with 14) and then water
(10 cm3) was added. The organic layer and the DCM extracts of
the aqueous layer (6 × 10 cm3) were combined, dried (MgSO4)
and the solvent removed to leave a pale yellow solid which on
recrystallisation from methanol–ether afforded 2-bromo-1,3,4-
triphenylphosphole 1-sulfide 16 (0.57 g, 85%) mp 155 ЊC
(decomp.) (Found: C, 62.2; H, 3.9%. C22H16BrPS requires C,
62.4; H, 3.8%); δH (CDCl3) 8.0–7.9 (m, 2H, PhP), 7.6–7.5 (m,
w = [E2(Fo ) ϩ (0.037P)2 ϩ 0.74P]. P = (F02 = 2Fc2)/3. Flack
2
parameter 0.008. Largest diff. peak and hole 0.331 and Ϫ0.461
e ÅϪ3
.
Synthesis of 2-bromo-1,3,4-triphenyl-2,5-dihydrophosphole
1-oxide 12
3H, PhP), 7.3–7.0 (m, 10H, PhC), 6.6 (d, 1H, 2JHP 28.7, CH᎐C);
Reaction of 6 (95.0 g, 15 mmol) with N-bromosuccinimide (2.7
g, 15 mmol) and benzoyl peroxide (0.15 g) as above afforded a
mixture of four products and some unreacted starting material.
The mixture was separated by column chromatography into
two fractions using toluene–ethyl acetate (2:5) as eluant. The
first fraction (2.5 g) was shown to be a mixture of 9, 10
and 11 above, the second fraction was shown to be a mixture
of starting material and a new component. This fraction
was re-separated using toluene–ethyl acetate (1:3) as eluant to
give unreacted starting material (0.6 g) and 2-bromo-1,3,4-
triphenyl-2,5-dihydrophosphole 1-oxide 12 (2.0 g, 33%) as gum;
(Found: C, 46.4; H, 4.2%. C22H18BrOP requires C, 46.6; H,
4.4%); δH (CDCl3) 7.9–7.8 (m, 2H, PhP), 7.6–7.5 (m, 3H, Ph-P),
7.3–7.1 (m, 10H, PhC), 5.0 (d, 2JHP 3.7, 1H, CHBr), 3.8–3.1(m,
᎐
δC (CDCl3) 157.2 (d, 1JCP 12.4, PhP), 150.3 (d, 2JCP 12.4, Ph-P),
3
3
135.2 (d, JCP 11.9, PhP), 133.6 (d, JCP 12.4, Ph-C), 130.9 (d,
2JCP 11.9, Ph-C), 129.3 (Ar), 128.4 (Ar), 128.1 (Ar), 125.3
(d, 1JCP 65.5, C5), 124.3 (d, 1JCP 83.6, CBr); δP (CDCl3) 51.3; MS
m/z 424/422 (1:1) [Mϩ], 390 [M Ϫ Sϩ] (Found mass 501.91438.
Required mass (C22H1779Br2PS) 501.915533).
Synthesis of 1,3,4-triphenylphosphole 1-oxide
2-Bromo-1,3,4-triphenyl-2,5-dihydrophosphole 1-oxide 12
(0.5 g, 1.22 mmol) in DCM (10 cm3) was refluxed with KOH
(0.14 g, 2.44 mmol) in methanol (4 cm3) for 5 h. The mixture
was treated with 0.5 M HCl (5 cm3) and the organic layer and
the combined DCM extracts of the aqueous layer (6 × 10 cm 3)
were dried (MgSO4). The solvent was removed to yield a pale
yellow solid which was purified by column chromatography
using toluene–ethyl acetate (1:3) as eluant to give 1,3,4-
triphenylphosphole 1-oxide (0.32 g, 79%), mp 165 ЊC (decomp.)
(Found: C, 80.3; H, 5.4%. C22H17PO requires C, 80.5; H, 5.2%);
δH (CDCl3) 7.5–7.4 (m, 2H, PhP), 7.3 (m, 3 H, PhP), 7.2–7.1 (m,
2
2H, CH2); δC (CDCl3) 138.9 (Cipso-PhP), 137.6 (d, JCP 14.1,
2
3
PhP), 136.4 (d, JCP 11.3, Ph-P), 129.1, 135.8 (d, JCP 9,
‡ CCDC reference number 207/422. See http://www.rsc.org/suppdata/
p1/b0/b000692k/ for crystallographic files in .cif format and tables of
crystal data, structure refinement, atomic coordinates, bond lengths
and angles.
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528
1525

View Article Online
10H, PhC), 7 (d, 2H, 2JHP 36.4, CH᎐C); δ (CDCl ) 152 (d, 2J
CP
C23H20BrO2P requires C, 62.9; H, 4.6%) δH (CDCl3) 7.9–7.8 (m,
2H, PhP), 7.5–7.3 (m, 3H, PhP), 7.2–7.00 (m, 10H, PhC), 6.8 (s,
1H, CHBr), 6.5 (d, 1H, JHP 13.6, CHOMe), 3.7 (d, JHP 10.9,
3H, OMe); δP (CDCl3) 31.5; MS m/z 359 [Mϩ], and (ii) 2-
bromo-1,3,4,-triphenylphosphole 3-oxide (0.1 g) 20 see below.
᎐
C
3
8.5, PhC᎐C), 138.4 (d, 3JCP 4, Ph-C), 134.4 (Ar), 133.9 (d, 2JCP
᎐
3
2
4
19.2, PhP), 130.5 (d, JCP 10.2, Ph-P), 129.5 (Ar), 128.8 (Ar),
128.4 (Ar), 128 (Ar), 127.3 (Ar), 125.3 (d, 1JCP 97.2, CH᎐C); δ
᎐
P
(CDCl3) 5.0 (Found mass 312.10663. Required mass (C22H17P)
312.10679).
Preparation of 2-bromo-1,3,4-triphenylphosphole-3-oxide 20
Synthesis of 2-bromo-1,3,4-triphenylphosphole 17
To a mixture of the dibromophopholene oxides 9 and 10 (3 g) in
DCM (50 cm3) at 0 ЊC was added KOH (0.68 g) in methanol (10
cm3) over 30 min at such a rate to maintain the temperature at
0 ЊC. The mixture was stirred at 0 ЊC for a further 4 h and was
then worked up as above to yield a mixture of largely one
product with traces of 19 and unreacted starting material as
indicated by 31P NMR spectroscopy. The mixture was purified
by column chromatography to remove the minor impurities to
afford 2-bromo-1,3,4-triphenylphosphole 1-oxide (1.9 g) as a
gummy solid (Found: C, 64.7; H, 3.7%. C22H16BrOP requires C,
2-Bromo-1,3,4-triphenylphosphole 1-sulfide 16 (1.0 g, 2.36
mmol) and tris(2-cyanoethyl)phosphine (2 g) in dry toluene (45
cm3) were refluxed for 6 h and then the cooled mixture was
allowed to stand overnight at room temperature. The mixture
was filtered through Celite and the solvent evaporated to yield a
pale yellow solid which was purified by column chromato-
graphy on silica using hexane–ethyl acetate (4:1) as eluant to
afford 2-bromo-1,3,4-triphenylphosphole 17 (0.8 g, 86%), mp
165–170 ЊC (decomp.) (Found: C, 67.2; H, 4.2%. C22H16BrP
requires C, 67.5; H, 4.1%); δH (CDCl3) 8.0–7.9 (m, 2H, PhP),
64.9; H, 4.0%); δH (CDCl3) 7.9–7.8 (m, 2H, Ar), 7.6–7.5 (m, 3H,
2
2
7.6–7.5 (m, 3H, PhP), 7.4–7.1 (m, 10H, PhC), 6.9 (d, 1H, JHP
Ar), 7.2–7.0 (m, 10H, Ar), 6.5 (d, 1H, JHP 27.6, CH᎐); δ
᎐
C
1
2
36.8, CH᎐C); δ (CDCl ) 15.9.
᎐
P
3
(CDCl3) 155.2 (d, JCP 12.4, PhP), 149.3 (d, JCP 12.4, Ph-P),
3
3
135.4 (d, JCP 11.9, PhP), 135.7 (d, JCP 12.4, Ph-C), 130.4 (d,
2JCP 11.9, Ph-C), 129.3 (Ar), 128.6 (Ar), 127.8 (Ar), 125.6 (d,
Synthesis of 2,5-dibromo-1,3,4-triphenylphosphole 18
1
Pyridine (5.0 cm3, 61.7 mmol) in DCM (30 cm3) was slowly
added to a solution of trichlorosilane (2.9 cm3, 28.3 mmol) in
DCM (20 cm3) under nitrogen at 0 ЊC. The white suspension
was stirred for a further 30 min and then 2,5-dibromo-1,3,4-
triphenylphosphole oxide 11 in DCM (30 cm3) was added drop-
wise over 30 min at 0 ЊC. After the addition was complete the
mixture was refluxed for 3 h, cooled and poured into saturated
sodium bicarbonate containing ice. The suspension formed was
filtered through Celite, the filtrate extracted with DCM (4 × 30
cm3) and the combined extracts and DCM washings (100 cm3)
of the Celite were dried (MgSO4). The solvent was evaporated
to give a yellow solid which was purified by column chromato-
graphy under nitrogen using hexane–ethyl acetate (4:1) as elu-
ant to yield 2,5-dibromo-1,3,4-triphenylphosphole 18 (1.9 g),
mp 170 ЊC (decomp.) (Found: C, 56.5; H, 3.3%. C22H15Br2P
requires C, 56.2; H, 3.2%); δH (CDCl3) 7.6 (m, 2H, PhP), 7.5 (m,
1JCP 65,5, C5), 124.1 (d, JCP 83.6, CBr); δP (CDCl3) 36.9; MS
m/z 407/405 [Mϩ].
A general method for the preparation of 5-bromo-2-lithio-1,3,4-
triphenylphosphole 22
The dibromophosphole 18 (1.0 g, 2.13 mmol) was dissolved in
THF–ether (2:1, 45 cm3) in a nitrogen atmosphere. The mixture
was cooled to Ϫ100 ЊC using a liquid nitrogen–ethanol slush
bath. N-Butyllithium (1.09 cm3, 1.89 M in hexanes) pre-cooled
to Ϫ78 ЊC was added dropwise with stirring and the mixture
stirred for 45 min at Ϫ100 ЊC. This solution was then used for
further reactions.
Temperature effects on the solutions of 5-bromo-2-lithio-1,3,4-
triphenylphosphole. A solution of the lithium reagent prepared
above was taken. After the stirring for 45 min stage, a sample
(2 cm3) was hydrolysed with 3 M HCl (0.5 cm3) and diluted with
DCM (3 cm3). The organic layer was washed with water (2 × 2
cm3), dried (MgSO4) and the solvent evaporated. The residue
was examined by 31P NMR spectroscopy. This procedure was
repeated with the solution at Ϫ75, Ϫ40 and ϩ20 ЊC. The results
we obtained are shown in Table 2 together with those using
different organometallic reagents.
3H, PhP), 7.3–7.1 (m, 10H, PhC); δC (CDCl3) 150.8 (d, 2JCP 7.9,
1
C᎐CPh), 135.6 (Ar), 134.3 (d, J 20.91, Ph-P), 131.1 (Ar),
᎐
CP
129.7 (Ar), 129.4 (d, 2JCP 8.5, Ph-P), 128.3 (d, 2JCP 13.6, Ph-P),
1
128 (Ar), 127.9 (Ar), 124.2 (d, JCP 20.9, CBr); δP (CDCl3)
22.4 (Found mass 467.929923. Required mass (C22H1579Br2P).
467.927811).
Synthesis of 1,3,4-triphenylphosphole 21
1,3,4-Triphenylphosphole 1-oxide (0.88 g, 2.68 mmol) was
reacted with trichlorosilane (0.69 cm3, 8.58 mmol) as described
above to yield 1,3,4-triphenylphosphole 21 (0.42 g, 50%) mp
155–160 ЊC (decomp.) (Found: C, 84.4; H, 5.8%. C22H17P
requires C, 84.6; H, 5.5%); δH (CDCl3) 7.5–7.4 (m, 2H, PhP), 7.3
Preparation of 1,3,4-triphenylphosphole-2-carbaldehyde 23
The lithium reagent from the bromophosphole 17 (0.25 g) at
Ϫ100 ЊC was treated with N-methylformanilide (0.09 cm3) in
THF (3 cm3), the mixture was stirred for 30 min and then
worked up by hydrolysis with 3 M HCl and extraction with
DCM (4 × 5 cm3) the combined organic layers were dried
(MgSO4) and the solvent evaporated to yield a mixture (0.3 g)
which was separated by column chromatography to yield
triphenylphosphole 21 (0.1 g) identical with an authentic
specimen and 1,3,4-triphenylphosphole-2-carbaldehyde 23 (0.15
2
(m, 3H, PhP), 7.2–7.1 (m, 10H, PhC), 7 (d, 2H, JHP 28.7,
2
3
CH᎐C); δ (CDCl3) 152 (d, JCP 8.5, PhP), 138.4 (d, JCP 4,
᎐
C
2
1
PhC), 134.4 (Ar), 133.9 (d, JCP 19.2, Ar), 130.5 (d, JCP 10.2,
1
Ar), 129.5 (Ar), 128.8 (Ar), 128 (Ar), 127.3 (Ar) 125.3 (d, JCP
97.2), 124.3; δP (CDCl3) 5.0; MS m/z 312 [Mϩ].
3
g), mp decomposed above 165 ЊC; δH (CDCl3) 9.6 (d, 1H, JHP
Reaction of 2,5-dibromo-1,3,4-triphenyl-2,5-dihydrophosphole
1-oxides 9 and 10 with potassium hydroxide in methanol
2
13.3, CHO), 7.9–7.2 (m, 15H, Ar), 7.0–6.9 (d, 1H, JHP 39.1,
CH᎐); δ (CDCl ) 6.8 (Found mass 340.10168, required mass
᎐
P
3
To a mixture of the dihydrophosphole oxides 9 and 10 (0.8 g) in
DCM (15 cm3) was added KOH (0.18 g) in methanol dropwise
over 20 min. The mixture was stirred at 18 ЊC for a further 30
min, water (10 cm3) was added and the aqueous layer extracted
with DCM (6 × 10 cm3). The combined organic layer and the
DCM extracts were dried (MgSO4) and the solvent evaporated
to yield a gummy white solid purified by column chromato-
graphy on silica using hexane–ethyl acetate (4:1) as eluant to
yield (i) 2-bromo-5-methoxy-1,3,4-triphenyl-2,5-dihydrophos-
phole 1-oxide 19 (0.6 g) as a gum (Found: C, 62.6; H, 4.4%.
(C23H17OP) 349.10170).
Preparation of 5-bromo-1,3,4-triphenylphosphole-2-
carbaldehyde 24
The lithium reagent 22 from the dibromophosphole 18 (0.97 g,
2.06 mmol) was treated with N-methylformanilide (0.25 cm3,
2.06 mmol) at Ϫ100 ЊC and the mixture stirred for 7 h at this
temperature. Work up as above afforded a pale yellow solid
(0.35 g) which was purified by column chromatography under
1526
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528

View Article Online
nitrogen using ethyl acetate–hexane (1:6) as eluant to yield
5-bromo-1,3,4-triphenylphosphole-2-carbaldehyde 24 (0.28 g),
mp decomposed above 170 ЊC (Found C, 65.6; H, 3.7%.
C23H16BrOP requires C, 65.9; H, 3.9%); δH (CDCl3) 9.5 (d, 1H,
3JHP 12.5, CHO), 7.6–7.5 (m, 2H, Ar), 7.5–7.4 (m, 3H, Ar), 7.3–
solution. The resulting suspension was continuously extracted
with ether for 40 h. The ether extract was dried (MgSO4) and the
solvent removed to leave an orange solid. The solid was washed
several time with ice cold ether to yield a mixture of dia-
stereoisomers of (E)-1,2-bis(5-bromo-1,3,4-triphenylphos-
phol-2-yl)ethene 29 (0.41 g) mp >300 ЊC (Found: C, 68.3;
H, 4.2%. C46H32Br2P2 requires C, 68.6; H, 4.00%); δH (CDCl3)
7.6–7.2 (m, 30H ϩ 30H major and minor isomers, Ar), 6.8 (d,
2
7.1, (m, 10H, Ar); δC (CDCl3) 188.2 (d, 1H, JHP 13.0, CHO),
2
2
3
162.5 (d, JHP 10.2, Ar), 151.1 (d, JHP 6.8, Ar), 146.1 (d, JHP
6.7, Ar), 140.8 (d, 1JCP 15.8, C-CHO), 137 (d, 1JCP 17, CBr), 130
(Ar), 128 (Ar), 127.5 (Ar); δP (CDCl3) 16.4 (Found mass
418.01252, required mass (C23H16BrOP) 418.01222).
3
3
2H, minor JHP 6.6, CH᎐CH), 6.6 (d, 2H, J 7, CH᎐CH); δ
᎐
᎐
HP
P
(CDCl3) 14.3 (major isomer), 14.2 (minor isomer); MS m/z
804/806/808 (1:2:1) [M]ϩ (Found mass 804.03328. Required
mass (C46H32Br2P2) 804.034600.
Reaction of 22 with propanal
The lithium reagent 22 from the dibromophosphole 18 (0.5 g,
1.05 mmol) was reacted as above with propanal (0.08 cm3, 1.05
mmol) in ether (5 cm3) and the mixture stirred for 5 h at
Ϫ100 ЊC. After work up as above a mixture of an oil (0.3 g) of
the two diastereoisomeric pairs of 1-(5-bromo-1,3,4-triphenyl-
phosphol-2-yl)propan-1-ol 25 in the ratio 62:38 was obtained.
Attempts to purify this mixture failed due to the ready dehydra-
tion to the alkene 26. Spectral details of 25: δH (CDCl3) 7.7–7.6
(m, 2H, Ar), 7.5–7.4 (m, 3H, Ar), 7.0 (m, 10H, Ar), 4.5 (m, 1H,
CHOH) 2.0 (m, 2H, CH2 minor isomer), 1.8 (2H, CH2 major
isomer), 0.9 (t, 3H, 3JHH 7.4, CH3 major isomer), 0.6 (t, 3H, 3JHH
7.4, CH3 minor isomer); δP (CDCl3) 12.4 (62%), 12.7 (38%); MS
m/z 449 [M]ϩ, 431 [M Ϫ H2O]ϩ.
Preparation of 1,2-bis(5-bromo-3,4-dimethyl-1-phenylphosphol-
2-yl)ethene 30
In the same way as above 2,3-dibromo-3,4-dimethyl-1-phenyl-
phosphole-2-carbaldehyde (1.20 g, 6.10 mmol)16 was reacted
with the McMurry reagent to yield 1,2-bis(5-bromo-3,4-
dimethyl-1-phenylphosphol-2-yl)ethene 30 (0.8 g) as a mixture
of diastereoisomers, mp decomposed 200 ЊC (Found: C, 55.6;
H, 4.1%. C26H24Br2P2 requires C, 55.9; H, 4.3%); δH (CDCl3)
3
7.8–7.2 (m, 10H, Ar) 6.8 (d, 2H JHP 12.1, CH᎐CH minor iso-
᎐
mer), 6.7 (d, 2H, 3JHP 13.2, CH᎐CH major isomer); δ (CDCl )
᎐
P
3
Ϫ2.4 (major), Ϫ2.6 (minor); MS m/z (1:2:1) 558/556/554 [M] ϩ.
The alcohol 25 (0.2 g) was passed down a silica column using
ethyl acetate–hexane (1:6) as eluant to afford 1-(5-bromo-1,3,4-
triphenylphosphol-2-yl)propene 26 (0.1 g) as an oil which
polymerised on heating, δH (CDCl3) 7.6–7.5 (m, 2H, Ar), 7.4–7.3
Preparation of bi(5-bromo-1,3,4-triphenylphosphol-2-yl) 31
Copper() chloride was added to a stirred solution of 22 from
the dibromophosphole 18 in THF at Ϫ100 ЊC. The mixture was
stirred for 5 h and worked up as above to yield a mixture of
three compounds (0.6 g), one of which was starting material.
Separation by column chromatography using ethyl acetate–
hexane 1:1 as eluant afforded (i) bi(5-bromo-1,3,4-triphenyl-
phosphol-2-yl) 31 (0.25 g), mp >250 ЊC (Found: C, 67.5; H,
3.6%. C44H30Br2P2 requires C, 67.7; H, 3.9%); δH (CDCl3) 7.9–
7.2 (m, Ar); δP (CDCl3) 26.9; MS m/z 782/780/778 (1:2:1) [M]ϩ,
and (ii) bi(1,3,4-triphenylphosphol-2-yl) 32 (0.15 g), mp
>250 ЊC; δH (CDCl3) 7.9–7.0 (m, 36H, Ar), 6.9 (d, 2H, 2JHP 38.7,
3
2
(m, 3H, Ar), 7.2–7.0 (m, 10H, Ar), 6.3 (m, 1H, JHP 15.1, JHH
1.8, 3JHH 0.7, CH), 5.9 (m, 2JHH 1.8, 2JHH 6.6, CH), 1.6 (dq, 2JHH
3
0.7, JHH 6.6, CH3); δP (CDCl3) 14.4 (Found mass 430.04854.
Required mass (C25H20BrP) 430.04860).
Reaction of 22 with iodine
A solution of 22 from the dibromophosphole 18 (1.0 g, 2.13
mmol) was treated at Ϫ100 ЊC with iodine (0.28 g, 2.2 mmol) in
ether (5 cm3) for 30 min. The mixture was worked up as above
to afford a pale yellow solid (0.95 g) which was shown to be an
inseparable mixture by GC and GC/MS of 2-bromo-5-iodo-
1,3,4-triphenylphosphole 27 (905) and 2-bromo-1,3,4-triphenyl-
phosphole (10%). Found for 27: MS m/z 518/516 [M]ϩ, 390/392
(1:1) [M Ϫ I]ϩ; δP (CDCl3) 32.8. This mixture could be used for
copper assisted reactions (see below).
CH᎐); δ (CDCl ) 20.9 (Found mass 622.19688. Required mass
᎐
P
3
(C44H32P2) 622.19792).
Reaction of phosphole 18 with copper(I) cyanide
A mixture of the dibromophosphole 18 (1.0 g, 2.13 mmol) and
copper() cyanide (0.42 g, 4.69 mmol) in dry DMF (35 cm3)
was heated at reflux for 6 h. The mixture was poured into ferric
chloride solution in 3 M HCl (10 cm3) and the mixture warmed
for 30 min. The aqueous layer was extracted with toluene
(3 × 30 cm3), the combined extracts were dried (MgSO4) and
the toluene evaporated to yield a pasty solid (0.4 g) shown by
TLC to be three components. The mixture was separated by
column chromatography to yield (i) 31 (0.15 g), (ii) 32 (0.07 g),
and (iii) 5-bromo-1,3,4-triphenylphosphole-2-nitrile 33 (0.02 g);
δH (CDCl3) 7.8–7.1 (m, Ar); δP (CDCl3) 26.7; MS m/z 417/415
(1:1) [M]ϩ.
Reaction of 22 with dichlorophenylphosphine
The lithiocompound 22 from the dibromophosphole 18 (1.0 g,
2.13 mmol) was reacted with dichlorophenylphosphine (0.14
cm3, 1.07 mmol) in ether and the mixture stirred at Ϫ100 ЊC for
6 h. Work up as above yielded a yellow solid (0.45 g) shown
by TLC to be a mixture of a major and minor components.
Separation of the mixture by column chromatography ethyl
acetate–hexane (1:1) as eluant afforded (i) bromo(5-bromo-
1,3,5-triphenylphosphol-2-yl)phenylphosphine 28 (0.38 g), mp
decomposed at 175 ЊC; δH (CDCl3) 7.6–7.0 (m, 20H, Ar);
δP (CDCl3) 27.5 (d, JPP 61), 25.1 (d, JPP 61); MS m/z 580/578/576
1:2:1 [M]ϩ (Found mass 575.94068. Required mass (C28H20Br2P)
575.94070) and (ii) an unidentified product (0.08 g).
References
1 F. Mathey, Chem. Rev., 1988, 88, 429.
2 F. C. Leavitt, T. A. Manuel and F. Johnson, J. Am. Chem. Soc., 1959,
81, 3163; E. H. Braye and W. Hubel, Chem. Ind. (London), 1959,
1250.
3 F. Mathey, Phosphorus, Sulfur, 1994, 87, 139.
4 F. B. Clarke and F. H. Westheimer, J. Am. Chem. Soc., 1971, 93,
4591.
5 L. D. Quin, E. Middlemas, N. S. Rao, R. W. Miller and A. J.
McPhail, J. Am. Chem. Soc., 1982, 104, 2912.
6 S. Nakayama, M. Yoshifugi, R. Okazaki and N. Inamoto, Bull.
Chem. Soc. Jpn., 1975, 48, 546.
7 K. Alder and J. Hayden, Liebigs Ann. Chem., 1950, 570, 201.
8 R. M. Dodson, V. Srinivasan, K. S. Sharma and R. F. Sauers,
J. Org. Chem., 1972, 37, 2367.
Preparation of 1,2-bis(5-bromo-1,3,4-triphenylphosphol-2-yl)-
ethene 29
Titanium tetrachloride (0.24 cm3, 2.15 mmol) was added drop-
wise over 20 min to a suspension of zinc powder (0.28 g, 4.29
mmol) in THF (7 cm3) at 0 ЊC. When the addition was complete
the mixture was heated at reflux for 2 h, cooled to rt and the
aldehyde 24 (0.60 g, 1.43 mmol) in THF (25 cm3) was added
dropwise. The mixture was then refluxed for a further 3 h then
cooled and poured into 10% aqueous potassium carbonate
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528
1527

View Article Online
9 R. Brettle and Sa’ad M. Shibil, J. Chem. Soc., Perkin Trans. 1, 1981,
2917.
15 E. Deschamps and F. Mathey, Bull. Soc. Chim. Fr., 1992, 129,
486.
10 A. De Groot, B. Evenhuis and H. Wynberg, J. Org. Chem., 1968, 33,
16 E. Deschamps, L. Ricard and F. Mathey, Angew. Chem., Int. Ed.
Engl., 1994, 33, 1158.
2214.
11 R. Vanderesse, Y. Fort, S, Becker and P. Caubere, Tetrahedron Lett.,
1986, 27, 3517.
12 F. T. Luo, S.-L. Fwa and W.-S. Wang, Tetrahedron Lett., 1992, 33,
3345.
17 G. Markel, J. Steigler, P. Kreitmeier, T. Burgermeister, F. Kastner
and S. Dove, Helv. Chim. Acta., 1997, 80, 14.
18 R. M. Kellogg, M. B. Green and H. Wynberg, J. Org. Chem., 1967,
32, 3993.
13 Y. Kondo, M. Takaki and R. Asami, Kobunshi Ronbunshi, 1989, 46,
769; Chem. Abstr., 1990, 112, 139866m. (We thank Professor Kondo
for an English translation of this reference.)
14 L. Brandsma and H. Verkruijsse, Preparative Polar Organometallic
Chem., Springer-Verlag, Berlin, 1981, 1.
19 (a) J. M. Klunder and G. H. Posner, in Comprehensive Organic
Synthesis, eds. B. M. Trost and I. Fleming, Pergamon Press, Oxford,
1991, vol. 3, p. 207; (b) D. W. Knight, in Comprehensive Organic
Synthesis, eds. B. M. Trost and I. Fleming, Pergamon Press, Oxford,
1991, vol. 3, p. 481.
1528
J. Chem. Soc., Perkin Trans. 1, 2000, 1519–1528