The interaction of phosphavinyl Grignard reagents with group 15 halides: Synthesis and structural characterisation of novel heterocyclic and heterocage compounds

Article abstract of DOI:10.1039/b204663f

The reactions of a phosphavinyl Grignard reagent, [CyP=C(Bu^t)MgCl(OEt₂)], Cy = cyclohexyl, with a variety of group 15 halide compounds in a number of stoichiometries have been investigated. When the Grignard reagent is reacted with P

Full text of DOI:10.1039/b204663f

View Article Online / Journal Homepage / Table of Contents for this issue
The interaction of phosphavinyl Grignard reagents with group
15 halides: synthesis and structural characterisation of novel
heterocyclic and heterocage compounds
Cameron Jones,*a Peter C. Junk,b Anne F. Richardsa and Mark Waugha
a
Department of Chemistry, University of Wales, Cardiff, P.O. Box 912, Park Place, Cardiff,
UK CF10 3TB
School of Chemistry, Box 23, Monash University, Clayton, Vic. 3800, Australia
b
Received (in London, UK) 13th May 2002, Accepted 7th June 2002
First published as an Advance Article on the web 11th July 2002
t
=
The reactions of a phosphavinyl Grignard reagent, [CyP C(Bu )MgCl(OEt2)], Cy ¼ cyclohexyl, with a
variety of group 15 halide compounds in a number of stoichiometries have been investigated. When the
Grignard reagent is reacted with Ph2PCl or CyPCl2 , Cy ¼ cyclohexyl, in a 1:1 and 2:1 stoichiometry
t
respectively, the 1,3-diphosphapropene, Ph2PC(Bu ) PCy, and triphosphabicyclo[2.1.0]pentane compound,
=
CyP{C2(But)2P2Cy2}, are formed. The 2:1 and 3:1 reactions of the Grignard reagent with either PCl3 or AsCl3
lead to the strained triphospha- and arsadiphosphabicyclo[1.1.1]pentanes, ButC{m-P(Cy)}2{m-ECl}CBut,
E ¼ P or As in moderate yield. The related 1:1 reaction of the Grignard reagent with PCl3 affords two products,
t
a phosphino-substituted phosphorus ylide, Cl2PC(Bu ) P(Cy)(Cl)2 , and a tetraphosphabicyclo[2.1.1]hexane,
=
ButC{m-P(Cl)P(Cy)}{m-P(Cy)}{m-P(Cl)}CBut, the mechanism of formation of which is discussed. An
analogous 1:1 reaction of the Grignard reagent with SbCl3 led to an unusual heterocyclic compound,
t
t
=
(Cl2Sb)(Bu )CSb(Cl)C(Bu ) P(Cl)(Cy)P(Cy), which quantitatively decomposes in solution to yield the known
1,2-dihydro-1,2-diphosphete, P2(Cy)2C2(But)2 . All prepared compounds have been crystallographically
characterised.
Introduction
Several years ago we developed a high yielding, stereo- and
regiospecific synthetic route to a range of phosphavinyl
t
=
Grignard reagents, Z-[RP C(Bu )MgCl(OEt2)], R ¼ alkyl or
aryl.1 Since that time we have been systematically investigating
the utility of these compounds as transfer reagents in reactions
with main group and transition metal halide complexes. We
have found that phosphavinyl Grignard reagents generally
behave significantly differently to their widely exploited vinyl
counterparts. These differences include the ease with which
the phosphavinyl fragment undergoes coupling reactions at
the metal centre to give a variety of novel metallocage, hetero-
cyclic and other coupled products, e.g. 12 and 2.3 In addition,
Results and discussion
oxidative coupling of the phosphavinyl fragment can occur
in these reactions to give strained heterocyclic systems such
as 34 and 4.5 Despite the facility of such coupling reactions,
complexes containing 1 or 2 terminal phosphavinyl ligands
In the initial stages of this work attempts were made to prepare
heteroleptic group 15–phosphavinyl complexes via the reaction
t
of [CyP C(Bu )MgCl(OEt2)] 5 with either Ph2PCl or in situ
=
t
6
=
can be prepared, e.g. [(o-C6H4O2)B–C(Bu ) PCy] and
generated CyPCl2 (Scheme 1). In the former reaction the
expected 1,3-diphosphapropene, 6, was formed in high yield
(75%) with retention of the stereochemistry of the phosphavi-
nyl fragment, whilst in the latter reaction a phosphavinyl cou-
t
2
=
[Me2Sn{–C(Bu ) PCy}2], and their utility as ligands has been
explored.
Given the wide applicability of vinyl group 15 compounds to
several areas of synthesis, e.g. organo–group 15 synthesis,7
macrocycle formation8 etc., and the ever increasing import-
ance of phosphorus containing heterocycle and cage com-
pounds,9 we wished to extend our study to the preparation
of phosphavinyl–group 15 compounds. This has led to the
formation of a number of novel and often unexpected hetero-
cyclic and heterocage compounds as described herein.
pling occurred to give
a
high yield (62%) of the
triphosphabicyclo[2.1.0]pentane compound, 7. The intermedi-
ate in this reaction is presumably 8 though this was not
observed when the reaction was followed by 31P NMR spectro-
scopy as 7 rapidly formed, even at ꢀ50 ꢁC.
The spectroscopic data for both compounds are consistent
with their proposed structures. In the case of 6 the 31P{1H}
DOI: 10.1039/b204663f
New J. Chem., 2002, 26, 1209–1215
1209
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
Scheme 1 Reagents and conditions: i, Ph2PCl, Et2O; ii, 1/2 CyPCl2 ,
Et2O.
NMR spectrum displays a low field signal (350.1 ppm) for the
phosphaalkene P-centre and a higher field signal (23.6 ppm)
for the PPh2 fragment with a large mutual two bond coupling
of 140 Hz. The 31P{1H} NMR spectrum of 7 displays 3 signals,
one at ꢀ104.3 ppm in the region normally associated with 3-
membered phosphirane rings and the other two at lower field
(ꢀ16.3 and 11.2 ppm), corresponding to the phosphorus cen-
tres in the 4-membered ring. The one bond coupling constant
(86 Hz) between these two centres is low but can be explained
by the likely high degree of p-character in the bond between
these two centres.
˚
Fig. 2 Molecular structure of compound 7. Selected bond lengths (A)
and angles (ꢁ): P(1)–C(1) 1.902(4), P(1)–P(2) 2.1986(17), P(2)–C(2)
1.860(4), P(3)–C(1) 1.854(4), P(3)–C(2) 1.864(4), C(1)–C(2) 1.576(5),
C(1)–P(1)–P(2) 78.17(13), C(2)–P(2)–P(1) 81.79(13), C(1)–P(3)–C(2)
50.17(17), C(2)–C(1)–P(3) 65.25(19), C(2)–C(1)–P(1) 99.8(2), P(3)–
C(1)–P(1) 109.06(19), C(1)–C(2)–P(2) 97.8(2), C(1)–C(2)–P(3)
64.59(19), P(2)–C(2)–P(3) 120.6(2).
X-Ray crystal structure analyses of 6 and 7 were carried out
to confirm their proposed structures (Figs. 1 and 2, Table 1).
Compound 6 was found to be monomeric and to crystallise
in the Z-isomeric form. The bond lengths P(1)–C(1) [1.678(3)
˚
˚
A] and P(2)–C(1) [1.855(3) A] are significantly different but
normal for localised double and single bonds respectively.10
These bond lengths are close to those reported recently in
substituents at P(1) and P(2) are trans to each other and the
P(3) substituent is in the endo-position. It is noteworthy that
a similar bicyclic framework was observed for the stannadi-
phosphabicyclopentane, 1,2 and in that case the phosphirane
cyclohexyl substituent is also endo.
the first crystallographically characterised 1,3-diphosphapro-
1
=
pene complexes, [W(CO)5{Z -Z-P(Ph)2C(Cl) PMes*}] and
[W(CO)4{Z -P,P-Z-P(Ph)2C(Cl) PMes*}],
2
=
Mes* ¼ C6H2-
But3-2,4,6.11 Compound 7 is the first structurally characterised
example of
a 2,3,5-triphosphabicyclo[2.1.0]pentane. It is
As an extension of this work attempts were made to prepare
homoleptic phosphavinyl–group 15 complexes by reacting 3
equivalents of 5 with ECl3 , E ¼ P, As, Sb and Bi. The reac-
tions with PCl3 and AsCl3 led to the formation of the tripho-
spha- and arsadiphosphabicyclo[1.1.1]pentanes, 9 and 10
respectively, in moderate yields after crystallisation from hex-
ane (Scheme 2). Presumably these are formed via the bispho-
sphavinyl phosphorus or arsenic intermediates, 11, which
undergo facile phosphavinyl coupling reactions to give the
bicyclic product. The intermediates were not observed when
the reactions were followed by 31P NMR spectroscopy and
even at ꢀ50 ꢁC 9 and 10 rapidly formed. It is worth noting that
the proposed intermediate, 11, is closely related to 8 but the
framework of the phosphavinyl coupled product derived from
the former is significantly different to that of the latter. It is
also interesting that only two equivalents of the phosphavinyl
Grignard react with the element halides and the remaining
chloride ligands in 9 and 10 are not open to substitution. This
contrasts with the formation of 2 from the reaction of 3
equivalents of 5 with AlCl3 . It is thought that in that case
an intermediate analogous to 9, i.e. E ¼ Al, is involved though
the likely trigonal planar Al centre in this intermediate is more
open to nucleophilic attack than the pyramidal P or As centres
in 9 and 10. When the reactions of 5 with PCl3 or AsCl3 were
carried out in a 2:1 stoichiometry, 9 and 10 were again formed
in similar yields. In addition, it should be mentioned that
monomeric and all bond lengths within the bicyclic framework
are normal for single bonded interactions. The cyclohexyl
related
trihalotriphosphabicyclo[1.1.1]pentanes,
ButC{m-
P(X)}3CBut, X ¼ Cl, Br or I, have been reported to be formed
by a very different route involving treatment of the 1,3-dipho-
sphabicyclobutanediyl complex, [Cp2Zr{Z2-C2(But)2P2}], with
PX3 .12
˚
Fig. 1 Molecular structure of compound 6. Selected bond lengths (A)
and angles (ꢁ): P(1)–C(1) 1.678(3), P(2)–C(1) 1.855(3), P(1)–C(6)
1.856(3), P(2)–C(12) 1.835(3), P(2)–C(18) 1.835(3), P(1)–C(1)–P(2)
129.80(16), C(1)–P(1)–C(6) 111.18(13), C(2)–C(1)–P(1) 118.1(2),
C(2)–C(1)–P(2) 112.0(2).
The spectroscopic data for 9 and 10 are as expected in that
the 31P{1H} NMR spectra of both compounds display signals
1210
New J. Chem., 2002, 26, 1209–1215

View Article Online
Table 1 Crystal data for compounds 6, 7, 9, 10, 12, 13, 14 and 16
6
7
9
10
12
13
14
16
Chemical formula
FW
Crystal system
C23H30P2
368.41
triclinic
C28H51P3
480.60
triclinic
C22H40ClP3 C22H40AsClP2 C27H40O5P2W C22H40Cl2P4 C11H20Cl4P2 C22H40Cl4P2Sb2
432.90
orthorhombic orthorhombic monoclinic
476.90
690.38
499.32
monoclinic
P21/n
9.4270(19)
16.555(3)
16.512(3)
90
356.01
monoclinic
P21/c
751.78
monoclinic
P21/c
¯
P1
¯
P1
Space group
˚
Pbcn
10.124(2)
23.941(5)
9.791(2)
90
Pbcn
P21/n
10.441(2)
18.702(4)
15.226(3)
90
a/A
˚
9.7667(8)
9.8576(5)
12.5778(9)
87.871(5)
69.191(6)
69.303(6)
10.012(4)
10.877(4)
15.047(7)
72.31(3)
89.79(3)
66.25(3)
10.1511(11)
24.438(3)
9.8440(11)
90
8.1380(16)
12.136(2)
16.208(3)
90
10.367(2)
23.930(6)
10.889(2)
90
b/A
˚
c/A
a/ꢁ
b/ꢁ
g/ꢁ
90
90
90
90
103.99(3)
90
93.34(3)
90
95.58(3)
90
114.71(3)
90
3
˚
V/A
1053.47(13) 1415.7(10)
2
2373.1(8)
4
2442.1(5)
4
2885.0(10)
4
2572.6(9)
4
1593.2(5)
4
2864.2(10)
4
Z
T/K
2
293(2)
0.21
4017
150(2)
0.22
5977
150(2)
0.37
4176
223(2)
1.64
10121
150(2)
4.15
5653
150(2)
0.51
48132
150(2)
0.92
27552
150(2)
2.38
30376
m(Mo-Ka)/mmꢀ1
Reflections collected
Unique reflections (Rint) 3777 (0.0223) 5751 (0.0835) 2135 (0.0841) 1725 (0.0771) 5199 (0.0478) 5884 (0.0361) 3647 (0.0499) 5560 (0.0800)
0.0256
0.0643
R1 (I > 2s(I))
wR02 (all data)
0.0425
0.1176
0.0574
0.1449
0.0670
0.1877
0.0923
0.2470
0.0394
0.1366
0.0286
0.0700
0.0488
0.1207
Scheme 2 Reagents and conditions: i, 1/3 ECl3 , Et2O.
for 3 or 2 chemically inequivalent phosphorus centres with no
measurable two bond P–P couplings, again most likely because
of the acute nature of the C–P–C angles in the strained cages.
˚
Fig. 3 Molecular structure of compound 9. Selected bond lengths (A)
and angles (ꢁ): Cl(1)–P(1) 2.095(3), P(1)–C(7) 1.848(5), P(2)–C(7)
1.926(5), P(20A)–C(7) 1.931(5), C(7A)–P(1) 1.816(5), C(7A)–P(2)
1.898(5), C(7)–P(1)–C(7A) 76.5(3), C(7)–P(2)–C(7A) 72.8(3), C(7)–
P(20A)–C(7A) 72.2(2), Cl(1)–P(1)–C(7) 110.7(2), P(1)–C(7)–P(2)
84.2(2), P(1)–C(7)–P(20A) 89.0(2), P(2)–C(7)–P(20A) 87.6(2). Symmetry
transformation A: ꢀx, y, ꢀz + 1/2.
1
As expected, only one But signal is seen in the H NMR spec-
tra and both compounds exhibit molecular ion peaks in their
APCI mass spectra.
The X-ray crystal structures of 9 and 10 are isomorphous
and show the hetero-atoms of each to be disordered over
two sites. In addition, there is a disorder within the cyclohexyl
substituents. Each disordered molecule lies on a 2-fold axis
containing Cl(1) and the mid point between C(7) and C(7A).
Only one disordered set is shown in Figs. 3 and 4 (see Table
1). The strained nature of these bicyclic compounds becomes
clear when the angles within the cages are examined. In the
case of 9 the average framework C–P–C angle is 74.7ꢁ whilst
the average P–C–P angle is 86.9ꢁ. Similar angles are seen in
2 and the closely related compound, ButC{m-P(Cl)}3CBut, in
which the C–P–C and P–C–P angles are 74.1(3)ꢁ and
87.5(3)ꢁ respectively.12 All the group 15 centres in 9 and 10
have distorted pyramidal geometries and the bond lengths to
the neighbouring C or Cl centres are in the normal range.10
The 3:1 reactions of 5 with SbCl3 or BiCl3 were not clean
and led to oily mixtures which 31P{1H} NMR spectroscopy
suggested comprised a multitude of phosphorus containing
products. One of the more prominent peaks seen in these mix-
tures occurred at ꢀ52 ppm which was assigned to the known
1,2-dihydro-1,2-diphosphete, 3, which most likely forms via
an oxidative coupling of two phosphavinyl fragments. Evi-
dence for such a coupling comes from the fact that elemental
antimony or bismuth were deposited in these reactions. More-
over, when the product mixture from the antimony reaction
was treated with [W(CO)5(THF)] the W(CO)5 complex of 3,
namely [W(CO)5{Z1-P2(Cy)2C2(But)2}] 12, was formed, albeit
in very low yield ( < 1%). The low yield of 12 prevented its
spectroscopic characterisation but its X-ray crystal structure
was obtained and is included here. Its molecular structure
(Fig. 5, Table 1) shows it to be monomeric with one P-centre
coordinated to a W(CO)5 fragment through its lone pair.
The geometry of the 1,2-dihydro-1,2-diphosphete ligand is
very close to that seen in the structure of the uncoordinated
ligand4 and other related compounds, e.g. Ph2P2C2But2
13
.
Considering the unexpected formation of 9 and 10 from the
2:1 and 3:1 reactions of 5 with PCl3 and AsCl3 it was thought
of particular interest to attempt analogous 1:1 reactions of 5
with group 15 halides. Again the outcome of these reactions
was unpredictable and led to novel results (Scheme 3). When
the PCl3 reaction was followed by 31P{1H} NMR spectroscopy
signals due to two major products appeared in the reaction
mixture at temperatures above ꢀ50 ꢁC. Both products were
isolated from the mixture as crystalline solids, though not in
as high yields as the NMR spectrum of the reaction mixture
New J. Chem., 2002, 26, 1209–1215
1211

View Article Online
Scheme 3 Reagents and conditions: i, 1/3 ECl3 , Et2O; ii, SO2Cl2 ,
Et2O; iii, 2 days, ꢀSb(s) , ꢀSbCl3 .
Fig. 4 Molecular structure of compound 10. Selected bond lengths
ꢁ
˚
(A) and angles ( ): Cl(1)–As(1) 2.201(3), As(1)–C(7) 1.980(7), P(1)–
C(7) 1.907(8), P(10A)–C(7) 1.923(9), C(7A)–As(1) 1.959(8), C(7A)–
P(1) 1.878(8), C(7)–As(1)–C(7A) 71.0(4), C(7)–P(1)–C(7A) 74.4(5),
C(7)–P(10A)–C(7A) 73.5(4), Cl(1)–As(1)–C(7) 110.7(2), As(1)–C(7)–
P(1) 85.6(3), As(1)–C(7)–P(10A) 89.4(3), P(1)–C(7)–P(10A) 89.1(4).
Symmetry transformation A: ꢀx, y, ꢀz + 1/2.
transfer reaction could be mimicked by treating in-situ gener-
ated 15 with a suitable chlorinating agent. To this end an ethe-
real solution of 5 was reacted with one equivalent of PCl3 at
ꢀ90 ꢁC. One equivalent of SO2Cl2 was quickly added to the
mixture which was then warmed to room temperature. The
31P NMR spectrum of the reaction mixture showed that 14
was the only phosphorus containing product which was subse-
quently isolated in moderate yield.
The spectroscopic data for 13 are suggestive of its proposed
structure as its H NMR spectrum displays two signals due to
1
chemically inequivalent tert-butyl groups. In addition, its
31P{1H} NMR spectrum exhibits two low field signals arising
from the chlorine substituted P-centres and two higher field
signals due to the cyclohexyl substituted P-centres. There is a
characteristic one bond P–P coupling (296 Hz) between P(2)
and P(4) (see Fig. 6 for numbering scheme) and a very large
Fig. 5 Molecular structure of compound 12. Selected bond lengths
ꢁ
˚
(A) and angles ( ): W(1)–P(1) 2.5845(14), P(1)–C(1) 1.847(5), P(1)–
P(2) 2.2063(19), P(2)–C(2) 1.849(5), C(1)–C(2) 1.366(8), C(1)–P(1)–
W(1) 126.68(17), P(2)–P(1)–W(1) 126.82(7), C(2)–P(2)–P(1) 75.64(18),
C(2)–C(1)–P(1) 101.3(4), C(1)–C(2)–P(2) 104.5(4).
suggested. The first of these, 13 (15% isolated yield), is a tetra-
phosphabicyclo[2.1.1]hexane in which two P-centres have Cl
substituents and two have cyclohexyl substituents. The second
product is the phosphino-substituted phosphorus ylide, 14 (5%
isolated yield).
Although no intermediates were observed in this reaction it
seems likely that the expected product, a dichlorophosphino-
phosphaalkene 15, E ¼ P, initially forms. A reductive coupling
reaction then occurs in which two molecules of 15 transfer a
chlorine atom to another molecule of 15 by an unknown pro-
Fig. 6 Molecular structure of compound 13. Selected bond lengths
ꢁ
˚
(A) and angles ( ): Cl(1)–P(1) 2.0848(6), Cl(2)–P(2) 2.0943(7), P(1)–
C(1) 1.8660(14), P(1)–C(2) 1.8873(14), P(2)–C(1) 1.8500(14), P(2)–
P(4) 2.2334(6), P(3)–C(2) 1.9146(14), P(3)–C(1) 1.9411(14), P(4)–C(2)
1.8989(14), C(1)–P(1)–C(2) 83.19(6), C(1)–P(1)–Cl(1) 109.60(5), C(2)–
P(1)–Cl(1) 106.77(5), C(1)–P(2)–Cl(2) 105.81(5), C(1)–P(2)–P(4)
94.42(5), Cl(2)–P(2)–P(4) 101.45(3), C(2)–P(3)–C(1) 80.51(6), C(2)–
P(4)–P(2) 91.35(5), P(2)–C(1)–P(1) 114.57(7), P(2)–C(1)–P(3)
100.91(6), P(1)–C(1)–P(3) 82.34(6), P(1)–C(2)–P(4) 103.78(7), P(1)–
C(2)–P(3) 82.50(6), P(4)–C(2)–P(3) 113.08(7).
cess to yield the observed ylide, 14. Tꢂhis would allow the two
t
reduced fragments, ‘‘CyP C(Bu )PCl ’’, to couple to give the
bicyclic compound, 13. It was postulated that the chlorine
=
1212
New J. Chem., 2002, 26, 1209–1215

View Article Online
two bond coupling (269 Hz) between P(1) and P(3). The mag-
nitude of this coupling could be due to a cross ring interaction
between the two P-centres which both have endo-substituents
with the P-lone pairs, presumably in close proximity to each
other. There is precedent for such large couplings between P-
centres in the 1- and 3-positions of puckered 4-membered rings
contained within organophosphorus cage compounds.14 The
31P{1H} NMR spectrum of the ylidic compound, 14, exhibits
an AB spin system, the lower field signal of which is derived
from the P(III) centre. The very large two bond coupling (377
Hz) between the P-centres is not unexpected as similar cou-
plings have been observed in closely related ylidic systems,
=
e.g. Ph3P C(Me)PCl2 .
15
The molecular structure of 13 (Fig. 6, Table 1) confirms its
bicyclic formulation and shows the substituents at P(2) and
P(4) to be trans to each other whilst the P(1) and P(3) substi-
tuents are in endo-positions. It is clear that the
P(1)C(1)P(3)C(2) 4-membered ring is puckered and has a short
˚
cross ring P(1)ꢃ ꢃ ꢃP(3) distance of 2.507 A which is well within
16
the sum of the van der Waals radii for two phosphorus centres
˚
(3.8 A). This close interaction likely gives rise to the large
mutual coupling between these P-centres observed in the
31P{1H} NMR spectrum of this compound. The molecular
structure of 14 (Fig. 7, Table 1) shows it to be monomeric
and to contain one phosphorus centre, P(1), in the +5 oxida-
tion state and one, P(2), in the +3 oxidation state. The C(1)–
Fig. 8 Molecular structure of compound 16. Selected bond lengths
ꢁ
˚
(A) and angles ( ): Sb(1)–C(1) 2.128(6), Sb(1)–C(2) 2.225(6), Sb(1)–
Cl(2) 2.4467(19), Sb(2)–C(2) 2.215(6), Sb(2)–Cl(3) 2.4041(19), Sb(2)–
Cl(4) 2.484(2), Cl(1)–P(1) 2.069(2), P(1)–C(1) 1.714(6), P(1)–P(2)
2.217(2), P(2)–C(2) 1.878(6), C(1)–Sb(1)–C(2) 86.3(2), C(1)–P(1)–P(2)
106.9(2), C(2)–P(2)–P(1) 96.3(2), P(1)–C(1)–Sb(1) 109.4(3), P(2)–
C(2)–Sb(1) 101.8(3), Sb(2)–C(2)–Sb(1) 94.9(2), Sb(2)–C(2)–P(2)
107.0(3), Cl(3)–Sb(2)–Cl(4) 89.99(8), Cl(3)–Sb(2)–C(2) 98.82(17),
Cl(4)–Sb(2)–C(2) 100.29(17), Cl(2)–Sb(1)–C(2) 100.04(16), Cl(2)–
Sb(1)-C(1) 101.14(17).
˚
P(1) distance of 1.6925(14) A supports this proposal and is
normal for an ylidic P–C bond.10 Although the P(2)–C(1) dis-
˚
tance at 1.7460(15) A is short for a single bond it can be
explained by the fact that C(1) is sp2-hybridised (S angles
359.8ꢁ).
The 1:1 reactions of 5 with either AsCl3 or BiCl3 were not
clean and gave inseparable mixtures of products which in the
latter reaction included significant amounts of 3. Interestingly,
the corresponding 1:1 reaction with SbCl3 was very clean and
gave a completely different result (Scheme 3). The 31P NMR
spectrum of the reaction mixture showed only one phosphorus
containing product, 16, that was isolated in good yield (66%)
after recrystallisation from hexane. This unusual heterocycle
contains an antimony and a phosphorus centre in the +3 oxi-
dation state and another phosphorus(V) centre that forms part
of an ylidic system. In addition, there is a terminal, exocyclic
SbCl2 fragment. Despite many attempts, we have not been able
to shed light on the mechanism of formation of 16 but it is
thought that it probably forms via a coupling and subsequent
rearrangement of two molecules of the supposed intermediate,
15, E ¼ Sb. One of the most interesting features of this com-
pound is that although it is moderately thermally stable in
the solid state, in toluene solutions it quantitatively decom-
poses to the diphosphete, 3, over 2 days, depositing elemental
antimony and presumably SbCl3 . Therefore compound 16 can
be considered as a trapped intermediate in a phosphavinyl cou-
pling reaction.
The 31P{1H} NMR spectrum of 16 displays an AB pattern
with a large one bond coupling of 314 Hz between the low field
ylidic phosphorus signal at 97.7 ppm and the resonance for the
P(III) centre at 16.0 ppm. Its molecular structure (Fig. 8, Table
1) confirms the heterocyclic nature of the compound and dis-
plays a distorted tetrahedral geometry for P(1) with the
˚
P(1)–C(1) distance of 1.714(6) A being in the expected region
for an ylidic bond (cf. compound 14).10 All other bond lengths
within the compound are normal for single bonded interac-
tions. The geometry about P(2) is distorted pyramidal as are
the geometries for Sb(1) and Sb(2). The angles about the anti-
mony centres are, however, more acute than those about the
phosphorus centre and are suggestive of considerable p-char-
acter to the bonding to their neighbouring atoms, as is com-
mon in tertiary antimony(III) compounds.
Conclusions
The reactions of a phosphavinyl Grignard reagent, 5, with a
variety of group 15 halide compounds in a number of stoichio-
metries have been carried out. These have led to the formation
of a series of novel organo–group 15 compounds which include
heterocycles, cages and acyclic unsaturated systems. The use of
several of these compounds as ligands is currently being
explored.
Fig. 7 Molecular structure of compound 14. Selected bond lengths
ꢁ
˚
(A) and angles ( ): Cl(1)–P(1) 2.0343(6), Cl(2)–P(1) 2.0378(6), Cl(3)–
P(2) 2.0934(6), Cl(4)–P(2) 2.1179(6), P(1)–C(1) 1.6925(14), P(2)–C(1)
1.7460(15), C(1)–C(2) 1.5659(19), P(1)–C(6) 1.8269(14), P(1)–C(1)–
P(2)
130.31(10), Cl(3)–P(2)–Cl(4) 96.18(3), Cl(1)–P(1)–Cl(2) 99.20(3).
106.76(8),
P(1)–C(1)–C(2)
122.74(10),
P(2)–C(1)–C(2)
New J. Chem., 2002, 26, 1209–1215
1213

View Article Online
APCI: m/z (%) 433 [M+, 72], 397 [M+ ꢀ Cl, 100%]; IR (Nujol)
n/cmꢀ1 1260(s), 1140(m), 1093(m), 970(m).
Experimental details
General remarks
ButC{m-P(Cy)}2{m-AsCl}CBut 10
All manipulations were carried out using standard Schlenk and
glove box techniques under an atmosphere of high purity
argon or dinitrogen. The solvents diethyl ether and hexane
were distilled over either potassium or Na/K alloy then
To a solution of AsCl3 (45ml, 0.54 mmol) in diethyl ether (20
ml) at ꢀ78 ꢁC was added a solution of 5 (0.34 g, 1.10 mmol)
in diethyl ether (10 ml) over 5 min. The resulting solution
was warmed to room temperature overnight, filtered and vola-
tiles removed in vacuo. The residue was extracted into hexane,
filtered and the filtrate placed at ꢀ30 ꢁC overnight to yield col-
1
freeze/thaw degassed prior to use. H and 31P NMR spectra
were recorded on Bruker DPX400, Bruker AMX 360 or Jeol
Eclipse 300 spectrometers in deuterated solvents and were
referenced to the residual 1H resonances of the solvent used
(1H NMR) or to external 85% H3PO4 , 0.0 ppm (31P NMR).
Mass spectra were recorded using a VG Fisons Platform II
instrument under APCI conditions. Melting points were deter-
mined in sealed glass capillaries under argon, and are uncor-
rected. Microanalyses were obtained from the Warwick
Microanalytical Service. Reproducible microanalyses of 6, 7,
9 and 10 could not be obtained due to small amounts of impu-
rities in the crystalline samples. However, the NMR spectra of
these samples suggested their purity was greater than 95%. The
starting material 5 was prepared by a literature procedure.1 All
other reagents were used as received.
1
ourless crystals of 10. (41% yield) m.p. 182–184 ꢁC; H NMR
(400 MHz, C6D6 , 300 K) d 1.10–2.10 (m, 20H, CH2), 1.21
(s, 18H, But), 2.95–3.06 (m, 2H, CH); 31P{1H} NMR (145.8
MHz, C6D6) d 13.5 (s, PCy), 73.8 (s, PCy); MS APCI: m/z
(%) 477 [M+, 100], 442 [M+ ꢀ Cl, 52%]; IR (Nujol) n/cmꢀ1
1376(m), 1260(s), 1020(m), 800(m).
ButC{m-P(Cl)P(Cy)}{m-P(Cy)}{m-P(Cl)}CBut 13
To a solution of PCl3 (140ml, 1.58 mmol) in diethyl ether (20
ml) at –78 ꢁC was added a solution of 5 (0.50 g, 1.60 mmol)
in diethyl ether (10 ml) over 5 min. The resulting orange solu-
tion was warmed to room temperature overnight, filtered and
volatiles removed in vacuo. The residue was extracted into hex-
ane, filtered and the filtrate placed at ꢀ30 ꢁC overnight to yield
colourless crystals of 13. (15% yield) m.p. 119–121 ꢁC; 1H
NMR (400 MHz, C6D6 , 300 K) d 1.30 (s, 9H, But), 1.54 (s,
9H, But), 0.98–2.72 (m, 22H, Cy); 31P{1H} NMR (121.7
t
Ph2PC(Bu ) PCy 6
=
A solution of 5 (0.54 g, 1.75 mmol) in diethyl ether (10 ml) was
added over 10 min to a solution of Ph2PCl (0.32 ml, 1.75
mmol) in diethyl ether (30 ml) at ꢀ78 ꢁC. The resulting solu-
tion was warmed to room temperature and stirred overnight.
Volatiles were removed in vacuo and the yellow residue
extracted into hexane (5 ml), whereupon the extract was slowly
cooled to ꢀ30 ꢁC to afford 6 as yellow prisms (75% yield) m.p.
2
1
MHz, C6D6) d 25.8 (d, P(3), JPP 269 Hz), 66.6 (d, P(4), JPP
1
2
296 Hz), 120.3 (d, P(2), JPP 296 Hz), 140.0 (d, P(1), JPP
269 Hz); MS APCI: m/z (%) 499 [M+, 100]; IR (Nujol) n/
cmꢀ1 1644(m), 1263(m), 1207(m), 1102(m), 997(w), 881(w);
found C 52.55, H 8.17%; calc. for C22H40Cl2P4 : C 52.92, H
8.07%.
1
98–103 ꢁC; H NMR (400 MHz, C6D6 , 300 K) d 1.20 (s, 9H,
But), 0.92–1.82 (m, 11H, Cy), 6.86–7.70 (m, 10H, Ph);
31
1
2
=
P{ H} NMR (145.8 MHz, C6D6) d 350.1 (d, P C, JPP
2
140 Hz), 23.6 (d, PPh2 , JPP 140 Hz); MS APCI: m/z (%)
368 [M+, 100], 285 [M+ ꢀ Cy, 5%]; IR (Nujol) n/cmꢀ1
1280(s), 1105(m), 980(s).
t
Cl2PC(Bu ) P(Cy)(Cl)2 14
=
To a solution of PCl3 (84 ml, 0.95 mmol) in diethyl ether (20
ml) at ꢀ78 ꢁC was added a solution of 5 (0.30 g, 0.95 mmol)
in diethyl ether (10 ml) over 5 min. To the resulting orange
solution was added SO2Cl2 (76 ml, 0.95 mmol). The reaction
mixture was then warmed to room temperature overnight, fil-
tered and volatiles removed in vacuo. The residue was
extracted into hexane (10 ml), filtered and the filtrate placed
at ꢀ30 ꢁC overnight to yield colourless crystals of 14. (26%
yield) m.p. 90–92 ꢁC; 1H NMR (300 MHz, C6D6 , 300 K) d
CyP{C2(But)2P2Cy2} 7
A 1 M solution of CyMgCl (1.10 ml, 1.1 mmol) was added to a
solution of PCl3 (0.10 ml, 1.1 mmol) in diethyl ether (20 ml) at
0 ꢁC and the solution stirred for 15 min. This was then cooled
to ꢀ78 ꢁC and a solution of 5 (0.70g, 2.2 mmol) in diethyl ether
(15 ml) added to it over 10 min. The resulting suspension was
warmed to room temperature and stirred overnight, filtered
and volatiles removed in vacuo. The residue was extracted into
hexane (10 ml), filtered and the filtrate placed at ꢀ30 ꢁC over-
night to afford colourless crystals of 7. (62% yield) m.p. 114–
116 ꢁC; 1H NMR (400 MHz, C6D6 , 300 K) d 1.15–2.73 (m,
33H, Cy), 1.46 (s, 9H, But), 1.52 (s, 9H, But); 31P{1H} NMR
1.71 (s, 9H, But), 0.66–2.82 (m, 11H, Cy); 31P{1H} NMR
2
=
(121.7 MHz, C6D6) d 77.0 (d, P C, JPP 377 Hz), 151.1 (d,
2
PCl2 , JPP 377 Hz); MS APCI: m/z (%) 102 [PCl2+, 100], 83
[Cy+, 62]; IR (Nujol) n/cmꢀ1 1286 (m), 1243 (s), 1208 (m),
1162 (m), 1002 (m), 849(w); found C 37.91, H 5.81%; calc.
for C11H20Cl4P2 : C 37.11, H 5.66%.
1
(145.8 MHz, C6D6) d 104.3 (s, CP(Cy) C), ꢀ16.3 (d, PP, JPP
1
86 Hz), 11.2 (d, PP, JPP 86 Hz); MS APCI: m/z (%) 481
[M+, 100], 398 [M+ ꢀ Cy, 12%]; IR (Nujol) n/cmꢀ1 1280(s),
t
(Cl2Sb)(Bu )CSb(Cl)C(Bu ) P(Cl)(Cy)P(Cy) 16
t
=
1110(m), 985(s).
To a solution of SbCl3 (0.36 g, 1.58 mmol) in diethyl ether (20
ml) at ꢀ78 ꢁC was added a solution of 5 (0.50 g, 1.60 mmol) in
diethyl ether (10 ml) over 5 min. The resulting orange solution
was warmed to room temperature overnight, filtered and vola-
tiles removed in vacuo. The residue was extracted into hexane,
filtered and the filtrate placed at ꢀ30 ꢁC overnight to yield yel-
ButC{m-P(Cy)}2{m-PCl}CBut 9
To a solution of PCl3 (71 ml, 0.8 mmol) in diethyl ether (20 ml)
at ꢀ78 ꢁC was added a solution of 5 (0.50 g, 1.62 mmol) in
diethyl ether (10 ml) over 5 min. The resulting solution was
warmed to room temperature overnight, filtered and volatiles
removed in vacuo. The residue was extracted into hexane, fil-
tered and the filtrate placed at ꢀ30 ꢁC overnight to yield col-
1
low crystals of 16. (66% yield) m.p. 83–85 ꢁC (dec.); H NMR
(400 MHz, C6D6 , 300 K) d 1.38 (s, 9H, But), 1.72 (s, 9H, But),
0.88–2.49 (m, 22H, Cy); 31P{1H} NMR (121.7 MHz, C6D6) d
1
1
1
ourless crystals of 9. (36% yield) m.p. 184–186 ꢁC; H NMR
16.0 (d, PCy, JPP 314 Hz), 97.7 (d, P C, JPP 314 Hz); MS
APCI: m/z (%) 559 [M+ ꢀ SbCl2 , 100]; IR (Nujol) n/cmꢀ1
1263 (s), 1207 (m), 1112 (m), 916 (w), 816 (m); found C
35.46, H 5.53%; calc. for C22H40Cl4P2Sb2 : C 35.14, H 5.36%.
=
(400 MHz, C6D6 , 300 K) d 1.05–1.95 (m, 20H, CH2), 1.36
(s, 18H, But), 2.85–2.95 (m, 2H, CH); 31P{1H} NMR (145.8
MHz, C6D6) d ꢀ0.4 (s, PCy), 74.4 (s, PCy), 95.8(s, PCl); MS
1214
New J. Chem., 2002, 26, 1209–1215

View Article Online
2
3
4
5
C. Jones and A. F. Richards, J. Chem. Soc., Dalton Trans., 2000,
3233.
C. Jones and A. F. Richards, J. Organomet. Chem., 2001, 629,
109.
C. Jones, A. F. Richards, S. Fritzsche and E. Hey-Hawkins, Orga-
nometallics, 2002, 21, 438.
C. Jones, J. A. Platts and A. F. Richards, Chem. Commun., 2001,
663.
C. Jones, M. Waugh and S. Aldridge, unpublished results.
E.g. (a) V. Maraval, R. Laurent, B. Donnadieu, M. Mauzac,
A. M. Caminade and J. P. Majoral, J. Am. Chem. Soc., 2000,
122, 2499; (b) L. Lassalle, S. Legoupy and J. C. Guillemin, Orga-
nometallics, 1996, 15, 3466; (c) S. Legoupy, L. Lassalle, J. C. Guil-
lemin, V. Metail, A. Senio and G. Pfister-Guillouzo, Inorg. Chem.,
1995, 34, 1466; (d ) J. C. Guillemin, Recent Res. Dev. Inorg. Chem.,
1998, 53, and references therein.
P. G. Edwards, P. D. Newman and K. M. A. Malik, Angew.
Chem., Int. Ed., 2000, 39, 2922 and references therein.
(a) K. B. Dillon, F. Mathey and J. F. Nixon, in Phosphorus:
The Carbon Copy, Wiley, Chichester, 1998; (b) Phosphorus–
Carbon Heterocyclic Chemistry: The Rise of a New Domain, ed.
F. Mathey, Pergamon, Amsterdam, 2001, and references therein.
Structure determinations
Crystals of 6, 7, 9, 10, 12, 13, 14 and 16 suitable for X-ray
structure determination were mounted in silicone oil or epoxy
resin. Crystallographic measurements were made using either
Nonius CAD4, Nonius Kappa CCD or Siemens SMART
CCD diffractometers. The structures were solved by direct
methods and refined on F2 by full matrix least squares
(SHELX97)17 using all unique data. All non-hydrogen atoms
are anisotropic with H-atoms included in calculated positions
(riding model). The quality of the X-ray data for compound
10 was poor but the fact that this compound is isomorphous
to 9 leaves no doubt about its gross molecular framework.
Crystal data, details of data collections and refinements are
given in Table 1. The molecular structures of the complexes
are depicted in Figs. 1–8 and show ellipsoids at the 30% prob-
ability level.
6
7
8
9
CCDC reference numbers 187286–187293. See http://
www.rsc.org/suppdata/nj/b2/b204663f/ for crystallographic
data in CIF or other electronic format.
10 As determined from a survey of the Cambridge Crystallographic
Database.
11 S. Ito and M. Yoshifuji, Chem. Commun., 2001, 1208.
12 P. Binger, T. Wettling, R. Schneider, F. Zurmuhlen, U. Bergstra¨s-
¨
Acknowledgements
ser, J. Hoffman, G. Maas and M. Regitz, Angew. Chem., Int. Ed.
Engl., 1991, 30, 207.
13 C. Charrier, N. Maigrot, F. Mathey, F. Robert and Y. Jeannin,
Organometallics, 1986, 5, 623.
14 D. E. Hibbs, M. B. Hursthouse, C. Jones and R. C. Thomas,
J. Chem. Soc., Dalton Trans., 1999, 2627.
We gratefully acknowledge financial support from EPSRC
(studentships for A. F. R. and postdoctoral fellowship for
M. W.).
15 A. Schmidpeter and G. Jochem, Tetrahedron Lett., 1992, 33,
471.
16 J. Emsley, The Elements, 2nd edn., Oxford University Press,
1991.
References
1
D. E. Hibbs, C. Jones and A. F. Richards, J. Chem. Soc., Dalton
Trans., 1999, 3531.
17 G. M. Sheldrick, SHELXL-97, University of Go¨ttingen, 1997.
New J. Chem., 2002, 26, 1209–1215
1215