Recovering phospholes from phosphacymantrenes

Article abstract of DOI:10.1016/S0022-328X(01)01101-9

Several attempts to recover phospholes from phosphacymantrenes are described. UV irradiation gives 1,1′-biphospholes when the ring carries two phenyl substituents. Lithium reduction appears to give somewhat erratic results. Exo-attack by phenyllithium lea

Full text of DOI:10.1016/S0022-328X(01)01101-9

Journal of Organometallic Chemistry 634 (2001) 131–135
www.elsevier.com/locate/jorganchem
Recovering phospholes from phosphacymantrenes
Bernard Deschamps *, Patrick Toullec, Louis Ricard, Franc¸ois Mathey *
Laboratoire ‘Hete´roe´le´ments et Coordination’, UMR 7653 CNRS, DCPH, Ecole Polytechnique, 91128 Palaiseau Cedex, France
Received 18 April 2001; accepted 5 July 2001
Abstract
Several attempts to recover phospholes from phosphacymantrenes are described. UV irradiation gives 1,1%-biphospholes when
the ring carries two phenyl substituents. Lithium reduction appears to give somewhat erratic results. Exo-attack by phenyllithium
leads to h4-anionic complexes derived from the corresponding 1-phenylphospholes. Quite surprisingly, this attack by phenyl-
lithium is compatible with a carbonyl fonctionality on the ring. These anionic complexes yield h4-phospholium derivatives upon
quaternization of phosphorus with BrCH2CH2Z (Z=CN, COOEt). One of these phospholium complexes has been caracterized
by X-ray crystal structure analysis. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Manganese; Phospholes; Phospholium salts; Phosphacymantrenes; Nucleophilic attacks
les relies on the bromine to lithium exchange in 2-bro-
mophospholes [11,12]. Alternatively, it is possible to
obtain some 2-functional phospholide ions using a
[1,5]-sigmatropic shift of the functional group from
phosphorus to the a-carbon of the ring [13,14]. Against
this background, it is obviously useful to be able to
convert a 2-functional phosphacymantrene or ferrocene
into the corresponding phospholes. This work describes
our attempts with phosphacymantrenes. The case of
phosphaferrocenes appears to be more difficult to solve.
1. Introduction
From a synthetic standpoint, the most useful charac-
teristic of phosphacymantrenes and phosphaferrocenes
is their ability to undergo electrophilic substitution
reactions [1–3]. This has allowed the synthesis of sev-
eral planar chiral 2-functional phosphaferrocenes which
have been used in asymmetric catalysis [4–7]. From a
theoretical standpoint, such chemistry is possible be-
cause of the very low energy of the sp2-lone pair at
phosphorus in both types of complexes [8,9]. By way of
contrast, the pyramidal phospholes sp3-lone pair is the
HOMO of the system. Hence, electrophilic substitution
reactions on phospholes are only possible when very
bulky P-substituents are present to force planarisation,
thus favoring the delocalisation of the lone pair over
the phosphole ring [10]. At the moment, the most
efficient technique for synthesizing functional phospho-
2. Results and discussion
Our first experiments were based upon an observa-
tion of Jaouen et al. [15] concerning the decomplexa-
tion of cymantrenes into cyclopentadienes by photolysis
in methanol. A transposition of this technique was
attempted with phosphacymantrenes (1–3).
The results were clearcut. The decomplexation was
successful for the phenyl-substituted derivative (3)
(Scheme 1) but led to a variety of decomposition prod-
ucts in the case of 1.
Scheme 1.
* Corresponding author.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)01101-9

132
B. Deschamps et al. / Journal of Organometallic Chemistry 634 (2001) 131–135
The resulting 1,1%-biphosphole (4) was characterized
lium)manganesetricarbonyl complexes were prepared
for the first time by Lindner through a completely
different approach [18]. Moreover, a similar reaction
scheme has been shown to be at work during the
reaction of tBuLi, then PhC(O)Cl with 3,3%,4,4%-te-
tramethyl-1,1%-diphosphaferrocene [19].
1
by H-NMR, 31P-NMR and mass spectrometry in com-
parison with an authentic sample. The LUMO of phos-
phacymantrenes is essentially an antibonding
combination of the pz(P) and dxz(Mn) orbitals [16]. The
promotion of an electron from the HOMO to the
LUMO very probably induces a cleavage of the P–Mn
bond. Product biradical then dimerizes and ultimately
gives the 1,1%-biphosphole. Unfortunately, whatever its
actual mechanism, this method proved to be uneffective
for functional compounds such as the 2-acetyl deriva-
tive of 1.
Without too much hope, we attempted to transpose
this type of reaction with the 2-acetylphosphacy-
In a second series of experiments, we studied the
reduction of phosphacymantrenes (1) and (3) by naph-
thalene–lithium in THF. The results observed with 3
are depicted in Scheme 2.
Longer reaction times did not affect the 5/6 ratio.
The final products 3 and 7 were identified by H-NMR,
Scheme 2.
Scheme 3.
1
31P NMR and mass spectrometry, and by comparison
with an authentic sample for 7. Both its 31P chemical
shift [17] and the nature of its alkylation product
demonstrate that 6 is the 2,5-diphenylphospholide ion.
Since compound 5 is reoxidized into the starting
product, we suspect that the actual reduction of 3
proceeds as follows (Scheme 3).
At the moment, we have been unable to get stable
derivatives of 5 and the proposed mechanism remains
speculative. Similar results have been obtained during
the reduction of 1. Unfortunately, we have been unable
to transform completely the starting phosphacymantre-
nes into the corresponding phospholides and the
method is obviously not compatible with reactive func-
tional groups.
Scheme 4.
We then decided to reinvestigate the reaction of
phosphacymantrenes with organolithium compounds.
A preliminary study [1] showed that the treatment of 1
with butyllithium, then with sulfur leads to 1-butyl-3,4-
dimethylphosphole sulfide. Similarly, the reaction of 1
with phenyllithium, then with sulfur gives 1-phenyl-3,4-
dimethyl–phosphole sulfide in 50% isolated yield
(Scheme 4). The monitoring of the reaction mixture
after the addition of PhLi showed a single 31P reso-
nance at −8 ppm in THF. We suspected the formation
of the h4-anionic species 8. This formulation was confi-
rmed by the reaction of 8 with quaternizing agents
giving the zwitterionic products 9 and 10 (Scheme 4).
In compound 9, the a-CH protons appear as a
doublet at 1.61 ppm (2JHP=18.4 Hz). For 10, the data
2
are quite similar (lH 1.77 ppm, JHP 18.7 ppm) so, the
,
Fig. 1. Molecular structure of (9). Significant bond distances (A) and
complexation of the dienic system is clearly established.
The X-ray crystal structure analysis of 9 (Fig. 1) shows
that PhLi selectively attacks the phosphorus atom of 1
on the exo side. Other noteworthy features are a rela-
angles (°): Mn(1)–C(1) 2.150(2), Mn(1)–C(2) 2.099(2), Mn(1)–C(3)
2.093(2), Mn(1)–C(4) 2.156(2), Mn(1)–P(1) 2.7651(8), P(1)–C(1)
1.740(2), P(1)–C(4) 1.748(2), P(1)–C(7) 1.814(2), P(1)–C(12)
1.819(2), C(1)–C(2) 1.447(3), C(2)–C(3) 1.414(3), C(3)–C(4) 1.449(3);
C(1)–P(1)–C(4) 90.2(1), C(7)–P(1)–C(12) 103.6(1), C(1)–P(1)–C(7)
115.3(1), C(1)–P(1)–C(12) 116.5(1), C(4)–P(1)–C(7) 116.5(1), C(4)–
P(1)–C(12) 115.2(1), C(1)–P(1)–Mn(1) 51.03(7), C(4)–P(1)–Mn(1)
51.25(7), C(7)–P(1)–Mn(1) 100.95(7), C(12)–P(1)–Mn(1) 155.45(7).
,
tively weak P–Mn interaction at 2.765 A and internal
P–C (ring) bonds which are very short at 1.740–
4
,
1.748(2) A. It must be stressed here that (h -phospho-

B. Deschamps et al. / Journal of Organometallic Chemistry 634 (2001) 131–135
133
mantrene (11). The attack of 11 by a sub-stoechiometric
amount of PhLi proved to be extremely clean. It gives
a single product 12 with a 31P resonance at −6 ppm in
THF. Thus, the attack of 11 exclusively takes place at
phosphorus, contrary to any reasonable expectation.
This dramatically demonstrates the extraordinary elec-
trophilicity of the phosphorus atom of 11 (Scheme 5).
The final product 13 displays all the expected spectral
characteristics. The a-CH proton resonates at high field
nance at l= −25 ppm. No further evolution could
then be detected. After the irradiation was stopped, the
solvents were removed under vacuum. The resultant oil
was redissolved in 5 ml of an hexane/diethyl ether (1:1)
mixture and filtered through a pad of 1 cm of silica.
The solvents were removed under vacuum to give the
2,2%,5,5%-tetraphenylbiphosphole (4) as a yellow oil (80
mg, 64% yield).
2
(lH 2.11 ppm, JHP=16.0 Hz), and the ketonic carbon
3.3. Reduction of phosphacymantrene (3) by
naphthalene–lithium
appears at 202.78 ppm. Since this kind of h4-phospho-
lium complexes can be easily decomplexed [20] and
dequaternized, we have here an unexpectedly clean and
general method to recover phospholes from
phosphacymantrenes.
2,5-Diphenylphosphacymantrene (3) (375 mg, 1×
10−3 mol) was placed in a Schlenk tube under nitrogen
atmosphere and 10 ml of distilled THF were added via
a septum. Then, 20 ml of a 10−1 M naphthalene–
lithium solution in THF were added in the same way.
The orange solution slowly turned red. 31P-NMR mon-
itoring showed the complete consumption of 3 after 15
min and the appearance of two new resonances: the
first one (l= +78 ppm) corresponding to the phos-
3. Experimental
3.1. General
pholide ion (6); the second one (l= +3.5 ppm; JP–H
=
All reactions were performed under an inert atmo-
sphere (nitrogen or argon). NMR spectra were mea-
sured on a multinuclear Bruker 300 MHz spectrometer.
Chemical shifts are expressed in ppm from internal
TMS (1H and 13C) or external 85% H3PO4 (31P); cou-
pling constants are expressed in Hz. Mass spectra (Elec-
tron Impact, unless otherwise noted) were measured at
70 eV by the direct inlet method. Elemental analysis
were performed at the service de Microanalyse du
CNRS, Gif sur Yvette, France.
34 Hz) unidentified. Addition of two equivalents of
iodomethane (284 mg; 2×10−3 mol) resulted in a color
change from red to orange; two products were detected
by 31P-NMR spectroscopy. They were isolated after
evaporation of THF under reduced pressure and chro-
matography of the crude oil with silica gel and an
hexane/diethyl ether mixture (9:1) as the eluent. The
first product obtained was the 1-methyl-2,5-
diphenylphosphole (7) (110 mg, 44% yield) and the
second one was the starting 2,5-diphenylphosphacy-
mantrene (3) (70 mg, 19% yield). Both products have
3.2. Photochemical decomplexation of
phosphacymantrene (3)
1
been characterized by 31P-, H-, 13C-NMR and mass
spectroscopy.
2,5-Diphenylphosphacymantrene (3) (200 mg, 5.3×
10−4 mol) was dissolved in 400 ml of a 1:1 mixture of
diethyl ether and ethanol. The resulting solution was
transferred into the irradiation cell and stirred with
argon bubbling for 15 min, then submitted to UV
irradiation for 75 min. The color of the solution quickly
changed from light orange to dark red. Monitoring the
reaction progress after 15 min by 31P-NMR spec-
troscopy showed the complete disappearance of the
starting material and the appearance of a single reso-
3.4. Phospholium complexes
3.4.1. (p4-1-Phenyl-1-(2-ethoxycarbonylethyl)-
3,4-dimethylphospholium)tricarbonyl manganese (9)
To a solution of 3,4-dimethylphosphacymantrene (1)
(0.5 g, 2×10−3 mol) in THF (10 ml) at −78 °C was
added dropwise phenyllithium 1.7 M (1.25 ml, 2.12×
10−3 mol). The mixture was allowed to warm to room
temperature (r.t.) and treated with ethyl 3-bromopropi-
onate (0.24 ml, 2×10−3 mol). After 30 min the solvent
was removed and the residue extracted with
dichloromethane, washed with water, dried with mag-
nesium sulfate and purified by chromatography on
silicagel with dichloromethane as the eluent; 0.6 g of 9
(70% yield) was obtained.
1
3
NMR (CDCl3) H: l 1.13 (t, JHH 7.1 Hz, CH2CH3),
2
1.61 (d, JHP 18.4 Hz, Ha), 2.10 (s, Me), 2.55 (m, CH2),
3
2.91 (m, CH2), 3.99 (q, JHH 7.1 Hz, COOCH2), 7.19–
7.38 (m, Ph).13C {1H}: l 14.63 (s, COOCH2CH3), 16.96
3
1
Scheme 5.
(d, JCP 6.2 Hz, CH3), 22.04 (d, JCP 80.9 Hz, PCH2),

134
B. Deschamps et al. / Journal of Organometallic Chemistry 634 (2001) 131–135
1
2
25.79 (d, JCP 82.5 Hz, Cb), 29.70 (d, JCP 3.3 Hz,
53.31 (d, 1JCP 87.7 Hz, Ca), 92.05 (d, 2JCP 23.1 Hz, Cb),
2 3
2
CH2COOEt), 61.83 (s, OCH2), 94.68 (d, JCP 17.0 Hz,
Cb), 128.16 (d, JCP 10.6 Hz, C ortho), 129.62 (d, JCP
9.1 Hz, C meta), 131.99 (d, 3JCP 2.5 Hz, C para), 138.01
99.86 (d, JCP 15.8 Hz, Cb), 118.03 (d, JCP 13.8 Hz,
3 2
2
2
CN), 129.89 (d, JCP 9.9 Hz, C meta), 129.93 (d, JCP
4
10.6 Hz, C ortho), 133.19 (d, JCP 2.5 Hz, C para);
1
3
1
(d, JCP 12.5 Hz, C ipso), 172.08 (d, JCP 12.5 Hz,
133.53 (d, JCP 31.0 Hz, C ipso), 202.78 (broad s,
CH2CO), 227.43 (MnCO).31P: l 36.38.
COCH3), 225.30 (s, MnCO). 31P: l 41.14.
Mass spectrum: m/z 423 ([M], 21.5%), 339 ([M−
3CO], 87.8%), 284 ([M−Mn−3CO], 60.5%).
Anal. Calc. for C20H19NO4PMn: C, 56.74; H, 4.49.
Found: C, 56.85; H, 4.61%.
Anal. Calc. for C20H22O5PMn: C, 56.07; H, 5.14.
Found: C, 55.82; H, 5.17%.
3.4.2. (p4-1-Phenyl-1-(2-cyanoethyl)-3,4-
dimethylphospholium)tricarbonylmanganese (10)
To a solution of 3,4-dimethylphosphacymantrene (1)
(0.1 g, 0.4×10−3 mol) in THF (2 ml) at −78 °C,
phenyllithium 1.7 M (0.25 ml, 0.425×10−3 mol) was
added dropwise. The mixture was allowed to warm to
r.t. and treated with 3-bromopropionitrile (0.033 ml,
0.4×10−3 mol). After 30 min the solvent was removed
and the residue extracted with dichloromethane,
washed with water, dried with magnesium sulfate and
purified by chromatography on silicagel with
dichloromethane as the eluent; 0.11 g of 10 (85% yield)
was obtained.
3.4.4. Crystallographic data for C20H22MnO5P (9)
M=428.29 g mol−1; monoclinic; space group P21/c;
,
,
,
a=14.102(5) A, b=17.354(5) A, c=16.299(5) A, i=
3
,
92.350(5)°, V=3985(2) A ; Z=8, two identical
molecules per asymmetric unit; Dcalc=1.428 g cm−3
;
v=0.770 cm−1; F(000)=1776. Crystal dimensions
0.20×0.20×0.20 mm3. Total reflections collected
18466 and 7510 with I\2|(I). Goodness-of-fit on F2
1.030; R=0.048, wR2=0.1205 (I\2|(I)); R=0.0854,
wR2=0.1377(all data); maximum/minimum residual
density 0.675(0.081)/−0.643(0.081) e A−3. Data were
,
1
2
NMR (CDCl3) H: l 1.77 (d, JHP 18.7 Hz, Ha), 2.20
collected on a KappaCCD diffractometer at 150.0(1) K
,
(s, CH3), 2.79 (broad s,CH2), 3.05 (broad s, CH2),
with Mo–Ka radiation (u=0.71073 A).
2
7.27–7.52 (m, Ph). 13C {1H}: 13.64 (d, JCP 2.9 Hz,
3
1
CH2CN), 16.80 (d, JCP 6.0 Hz, CH3), 23.56 (d, JCP
1
78.9 Hz, PCH2), 24.61 (d, JCP 80.3 Hz, Ca), 94.83 (d,
4. Supplementary material
2JCP 17.5 Hz, Cb), 118.15 (d, JCP 12.6 Hz, CN), 127.94
3
2
3
(d, JCP 10.5 Hz, C ortho), 129.90 (d, JCP 9.3 Hz, C
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC No. 161737 for compound 9.
Copies of this information may be obtained free of
charge from The Director, CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK (fax: +44-1233-336033; e-
mail: deposit@ccdc.cam.ac.uk or www: http://www.
ccdc.cam.ac.uk).
1
meta), 132.58 (s, C para), 136.99 (d, JCP 21.8 Hz, C
ipso), 226.94 (s, CO). 31P: l 36.82.
Mass spectrum: m/z 297 ([M−3CO], 19.5%), 242
([M−Mn−3CO]).
Anal. Calc. for C18H17NO3PMn: C, 56.69; H, 4.46.
Found: C, 56.93; H, 4.65%.
3.4.3. (p4-1-Phenyl-1-(2-cyanoethyl)-2-acetyl-
3,4-dimethylphospholium)tricarbonyl manganese (13)
To a solution of 2-acetyl-3,4-dimethylphosphacy-
mantrene (11) (0.2 g, 0.68×10−3 mol) in THF (4 ml)
at −78°C, phenyllithium 1.7 M (0.6 ml, 0.68×10−3
mol) was added dropwise. The mixture was allowed to
warm to r.t. and treated with 3-bromopropionitrile
(0.056 ml, 0.67×10−3 mol). After 30 min the solvent
was removed and the residue extracted with
dichloromethane, washed with water, dried with mag-
nesium sulfate and purified by chromatography on
silicagel with dichloromethane/ethyl acetate (95/5) as
the eluent; 0.11 g of 13 (59% yield) was obtained.
Acknowledgements
One of us, P. Toullec, thanks BASF for financial
support.
References
[1] F. Mathey, A. Mitschler, R. Weiss, J. Am. Chem. Soc. 100
(1978) 5748.
[2] G. de Lauzon, B. Deschamps, J. Fischer, F. Mathey, A.
Mitschler, J. Am. Chem. Soc. 102 (1980) 994.
[3] B. Deschamps, L. Ricard, F. Mathey, Organometallics 18 (1999)
5688.
[4] S. Qiao, G.C. Fu, J. Org. Chem. 63 (1998) 4168.
[5] K. Tanaka, S. Qiao, M. Tobisu, M.M.-C. Lo, G.C. Fu, J. Am.
Chem. Soc. 122 (2000) 9870.
[6] C. Ganter, C. Kaulen, U. Englert, Organometallics 18 (1999)
5444.
1
1
NMR (CDCl3) H: l 2.11 (d, JHP 16.0 Hz, HCP),
4
2.19 (d, JHP 1.1 Hz, CH3), 2.28 (s, COCH3), 2.38 (d,
4JHP 0.8 Hz, CH3), 2.57–2.92 (m, CH2), 3.05–3.30 (m,
2
CH2), 7.48–7.56 (m, Ph). 13C {1H} l 13.46 (d, JCP 2.3
3
Hz, CH2CN), 14.83 (d, JCP 3.7 Hz, CH3), 18.03 (d,
3JCP 5.5 Hz, CH3), 22.98 (d, 1JCP 77.4 Hz, Ca), 23.81 (d,
3
1JCP 78.9 Hz, PCH2), 27.93 (d, JCP 7.7 Hz, COCH3),
[7] R. Shintani, M.M.-C. Lo, G.C. Fu, Org. Lett. 2 (2000) 3695.

B. Deschamps et al. / Journal of Organometallic Chemistry 634 (2001) 131–135
135
[8] N.M. Kostic, R.F. Fenske, Organometallics 2 (1983) 1008.
[9] C. Guimon, G. Pfister-Guillouso, F. Mathey, Nouveau J. Chim.
3 (1979) 725.
[10] G. Keglevich, Z. Bo¨cskei, G.M. Keseru¨, K. Ujsza´szy, L.D. Quin,
J. Am. Chem. Soc 119 (1997) 5095.
[15] S. Top, E.B. Kabun, G. Jaouen, J. Am. Chem. Soc. 122 (2000)
736.
[16] DFT calculations have been performed on the parent complex
[Mn(h5-C4H4P)(CO)3] by P. Le Floch. Further details will be
published later.
[11] E. Deschamps, F. Mathey, Bull. Soc. Chim. Fr. 129 (1992) 486.
[12] E. Deschamps, L. Ricard, F. Mathey, Angew. Chem. Int. Ed.
Engl. 33 (1994) 1158.
[17] C. Charrier, F. Mathey, Tetrahedron Lett. 28 (1987) 5025.
[18] E. Lindner, A. Rau, S. Hoehne, J. Organomet. Chem. 218 (1981)
41.
[13] S. Holand, M. Jeanjean, F. Mathey, Angew. Chem. Int. Ed.
Engl. 36 (1997) 98.
[19] B. Deschamps, J. Fischer, F. Mathey, A. Mitschler, L. Ricard,
Organometallics 1 (1982) 312.
[14] S. Holand, N. Maigrot, C. Charrier, F. Mathey, Organometallics
17 (1998) 2996.
[20] B. Deschamps, J. Fischer, F. Mathey, A. Mitschler, Inorg.
Chem. 20 (1981) 3252.