Different thermal reactivity of a 1,2-thiaphospholo[a]phosphirane in free and metal carbonyl complexed form

Article abstract of DOI:10.1039/b310256d

The W(CO)₅ and Fe(CO)₄ complexes of the bicyclic phosphirane 3,5,6,6-tetraphenyl-1-phospha-2-thiabicyclo-[3.1.0]hex-3-ene undergo a thermal 2-phenylphosphirane → dihydrophosphaisoindole ring expansion, while the free phosphirane suffers both a [2 + 1] cycloreversion and a fragmentation yielding a butadienyl sulfide.

Full text of DOI:10.1039/b310256d

Different thermal reactivity of a 1,2-thiaphospholo[a]phosphirane in free
and metal carbonyl complexed form†
Tamaki Jikyo and Gerhard Maas*
Division of Organic Chemistry I, University of Ulm, Albert-Einstein-Allee 11, D-89081 Ulm, Germany.
E-mail: gerhard.maas@chemie.uni-ulm.de; Fax: +49 (0)731 50 22803; Tel: +49 (0)731 50 22790
Received (in Cambridge, UK) 25th August 2003, Accepted 23rd September 2003
First published as an Advance Article on the web 14th October 2003
The W(CO)5 and Fe(CO)4 complexes of the bicyclic phos-
phirane 3,5,6,6-tetraphenyl-1-phospha-2-thiabicyclo-
[3.1.0]hex-3-ene undergo a thermal 2-phenylphosphirane ?
dihydrophosphaisoindole ring expansion, while the free
phosphirane suffers both a [2 + 1] cycloreversion and a
fragmentation yielding a butadienyl sulfide.
The thermal rearrangement 3?5 can be described as a
2-phenylphosphirane ? dihydroisophosphindole ring expan-
sion and requires the C–C bond of the phosphirane ring to be
broken. This is unusual because, as far as we know, most
thermally induced ring-opening reactions of phosphiranes
described so far occur at one of the P–C bonds.1,7 (The
The general reactivity patterns of the strained three-membered
phosphirane ring system are well established.1 Recent interest
in these cyclic phosphines has focused on their ligand properties
in catalytically active transition-metal complexes.2,3 For this
application, however, phosphiranes incorporated in polycyclic
frameworks appear to be better suited due to their lower
reactivity compared to many monocyclic phosphiranes.3
In this context, we became interested in bicyclic phosphir-
anes which are [a]-annelated with the P–C bond of a
heterophosphole. Only a few such systems were known before
we found, both experimentally4 and computationally,5 that the
PNC double bond of 3,5-diphenyl-1,2-thiaphosphole6 (1) is an
excellent acceptor for the diazo dipole and cyclopropanation of
this bond occurs easily. In fact, thiaphosphole 1 reacts with
diazodiphenylmethane to give the (1,2-thiaphospholo)phos-
phirane 2‡ in good yield (Scheme 1).† The progress of the
reaction is easily monitored by 31P NMR: dp(1) = 204.2 ppm,
dP(2) = 281.4 ppm.
Phosphirane 2 readily formed the pentacarbonyltungsten
complex 3 (Scheme 2) which was characterised by single-
crystal X-ray diffraction analysis (Fig. 1).ठThis complex (dP
= 255.4 ppm) has a surprisingly low thermal stability and
undergoes a rearrangement already at 80 °C to form the
thienoisophosphindole derivative 5‡ (dP = 125.1 ppm) which
was identified by XRD analysis (Fig. 2).§ The tetracarbonyliron
complex of phosphirane 2 is obviously even less stable
thermally, because treatment of 2 with diiron nonacarbonyl at
80 °C/30 min directly yielded the ironcarbonyl-complexed
rearrangement product 6 (dP = 195.2 ppm). Complex 6 is a
rather unstable red oil which after column chromatography was
obtained in only 17% yield.‡
Scheme 2 Reactions and conditions: (i) toluene, 80 °C, 3 h, 76% yield.
Scheme 1 Reactions and conditions: (i) dry toluene, 70 °C, 24 h, 74% yield;
(ii) W(CO)5(THF), dry THF, 20 °C, 24 h, 71% yield.
Fig. 1 Solid-state structure of 3. Selected bond distances (Å) and angles (°):
P1–C3 1.878(3), P1–C4 1.872(3), C3–C4 1.576(4), P1–S1 2.097(1), P1–W
2.480(1), W–C29 2.001(4); C3–P1–C4 49.7(1), S1–P1–C3 96.5(1), S1–P1–
C4 106.1(1).
† Electronic Supplementary Information (ESI) available: full experimental,
spectroscopic and analytical data. See http://www.rsc.org/suppdata/cc/b3/
b310256d/
2794
CHEM. COMMUN., 2003, 2794–2795
This journal is © The Royal Society of Chemistry 2003

T. J. thanks the Alexander von Humboldt Foundation for a
fellowship.
Notes and references
‡ Selected physical and spectroscopic data: Compound 2: colourless
3
needles, mp 146 °C; 1H NMR: d 6.50 (d, JP,H = 4.8 Hz, 1H, NCH),
6.78–7.50 (m, 20H, HPh); 13C{1H} NMR: d 42.1 (d, 1JP,C = 48.6 Hz), 73.4
(d, 1JP,C = 40.6 Hz), 125.6, 126.4, 126.6, 127.0, 127.7, 128.2, 129.2, 129.4,
130.9, 131.5, 133.5, 137.6, 142.2 (2 C). Compound 3: colourless prisms,
m.p. 127 °C; 1H NMR: d 6.18 (d, 3JP,H = 17.2 Hz, 1H, NCH), 6.48 (d, J =
8.3 Hz, 2HPh), 6.90–7.50 (m, 18HPh); 13C{1H} NMR: d = 39.9 (d, 1JP,C
=
1
23.5 Hz), 67.1 (d, JP,C = 26.2 Hz), 126.0–135.9 (16 C), 142.9, 194.2,
195.8; 31P NMR: d 255.4 (t, 1J(31P,183W) = 266 Hz). Compound 5:
colourless prisms, mp173–174 °C; 1H NMR: d 5.28 (d, JP,H = 14.0 Hz, 1H,
PCH), 6.30 (d, JP,H = 18.3 Hz, 1H, NCH), 7.2–7.4 (m, 17Haryl), 7.66 (dd, J
= 8.1 and 1.9 Hz, 2HPh); 13C{1H} NMR: d 62.8 (d, 1JP,C = 3.8 Hz), 75.3
(d, 1JP,C = 11.0 Hz), 125.4–134.3 (14 C), 141.7, 142.9, 194.3, 194.4, 197.7;
Fig. 2 Solid-state structure of 5. Selected bond distances (Å) and angles (°):
P1–S1 2.090(3), P1–C3 1.911(6), P1–C10 1.879(7), P1–W1 2.5103(16),
W1–C15 1.965(7); S1–P1–C3 96.1(2), S1–P1–C10 104.9(2), C3–P1–C10
92.9(3), C3–P1–W1 127.0(2).
1
31P NMR: d 125.1 (t, 1J(31P,183W) = 262 Hz). Compound 6: red oil; H
3
NMR: d 5.24 (d, JP,H = 14.2 Hz, 1H), 6.36 (d, JP,H = 20.9 Hz, 1H),
6.9–7.3 (17Haryl), 7.60 (dd, J = 8.1 and 1.8 Hz, 2HPh); 13C{1H} NMR: d
60.4 (d, 1JP,C = 9.1 Hz), 72.7 (d, 1JP,C = 16.3 Hz), 122.3–137.4 (9 C) and
142.1, 151.1, 210.7, 210.9; 31P NMR: d 195.2. Compound 7: colourless
solid, mp 129–130 °C; IR (KBr): n = 2550 cm21 (w, SH); 1H NMR: d 3.09
(s, 1H, SH), 6.57 (s, 1H, 2-H), 7.01–7.44 (m, 20HPh). All NMR spectra were
taken from CDCl3 solutions at 400.13 (1H), 100.62 (13C) or 161.98 Hz
electrocyclic ring-opening of a 9-phosphabicyclo[6.1.0]nona-
2,4,6-triene P-oxide is an exceptional case.8) Two reasons may
account for the ring opening at the C–C bond: (a) this bond (C3–
C4 in Fig. 1, 1.576(4) Å) is rather long; (b) homolytic cleavage
of this bond yields a 1,3-diradical 4 (Scheme 2) that is
exceptionally well resonance-stabilised by the adjacent p
systems.
(
31P).
¯
§ Crystal data for 3: C33H21O5PSW, M = 744.38, triclinic, space group P1
(no. 2), a = 11.159(3), b = 11.268(3), c = 13.387(4) Å, a = 69.34(3), b
= 79.86(3), g = 69.31(3)°, V = 1470.9(7) Å3, Z = 2, Dc = 1.681 g cm23
,
Surprisingly, the free bicyclic phosphirane 2 shows a
different thermal behaviour compared with its metal complexes.
The thermal stability of the free phosphirane is higher, and
thermal impact at 120 °C yields a mixture of butadienyl sulfide
7‡ (30%), thiaphosphole 1 (31%), and tetraphenylethene (38%)
(Scheme 3), but no ring expansion product. Obviously, the
products result from two processes: (a) a [2 + 1] cycloreversion
yielding 1 and diphenylcarbene which then dimerises, and (b) a
fragmentation with loss of the phosphorus atom the fate of
which is not known. Perhaps, the latter process begins with a [2
+ 1] cycloreversion that generates a phosphinidene R–S–P
which after cleavage of the sulfur–phosphorus bond and H
abstraction from the solvent yields sulfide 7.
m(Mo–Ka) = 4.09 mm21, T = 193 K; 15 692 measured reflections, 5 352
independent reflections (Rint = 0.0504). Refinement of 370 variables
converged at R1 = 0.0273, wR2 = 0.0552 for all independent reflections
and R1 = 0.0240, wR2 = 0.0542 for 4910 reflections with I > 2s(I). For
5: C33H21O5PSW, M = 744.38, monoclinic, space group P2/n, a =
8.875(2), b = 10.701(2), c = 30.832(5) Å, a = 90, b = 92.81(2), g = 90°,
V = 2924.7(9) Å3, Z = 4, Dc = 1.691 g cm23, m(Mo–Ka) = 4.12 mm21
,
T = 193 K; 20 349 measured reflections, 4 431 independent reflections (Rint
= 0.0510). Refinement of 378 variables converged at R1 = 0.0312, wR2 =
0.0772 for all independent reflections and R1 = 0.0516, wR2 = 0.0816 for
3209 reflections with I > 2s(I). For both structures, data collection was
done on a Stoe IPDS diffractometer. The structures were solved using
SHELXS and refined on F2 values using SHELX-97. Hydrogen atoms were
included at calculated positions and treated as riding on their bond
neighbours; H2 and H10 in 5 were refined freely. CCDC 218743 (5) and
218744 (3). See http://www.rsc.org/suppdata/cc/b3/b310256d/ for crys-
tallographic data in .cif or other electronic format.
1 F. Mathey and M. Regitz, in Comprehensive Heterocyclic Chemistry II,
eds. A. R. Katritsky, C. W. Rees and E. F. V. Scriven, Vol. 1A, ed. A.
Padwa, Pergamon, Oxford, 1996, pp 277–304; F. Mathey and M.
Regitz, in Phosphorus-Carbon Heterocyclic Chemistry, ed. F. Mathey,
Elsevier, Oxford, 2001, pp 17–55.
Scheme 3 Conditions: (i) mesitylene, 120 °C, 6 h.
2 A. Marinetti, F. Mathey and L. Ricard, Organometallics, 1993, 12,
1207–1212.
The reason for the different thermal behaviour of free
phosphirane 2 and its P-complexed counterpart 3 (as well as the
analogous iron complex not shown in Scheme 2) is not clear at
present. In fact, if one accepts a different behaviour, one might
have expected the opposite, since it is known that some
phosphirane–W(CO)5 complexes easily undergo cyclorever-
sion with elimination of a P–W(CO)5 fragment.1,9 We suggest
that the elongation of the C–C bond in the metal-complexed
phosphirane ring is a crucial factor: although the length of this
bond in phosphirane 2 is unknown, it is generally expected that
participation of the phosphorus lone pair in bonding leads to an
elongation of the C–C ring bond and a shortening of the two P–
C bonds.3 However, while this effect has been documented for
several Rh2 and Pt3 complexes of phosphiranes, comparisons
between free phosphiranes and their associated W(CO)5
complexes appear not to be available yet. On the other hand, the
C–C bond in the calculated structure of Cr(CO)5(phosphirane)
complex10 is indeed longer by ca. 0.02–0.03 Å than the
experimental and calculated values for the free ligand.
3 J. Liedtke, S. Loss, C. Widauer and H. Grützmacher, Tetrahedron, 2000,
56, 143–156; J. Liedtke, H. Rüegger, S. Loss and H. Grützmacher,
Angew. Chem., Int. Ed., 2000, 39, 2478–2481; J. Liedtke, S. Loss, G.
Alcaraz, V. Gramlich and H. Grützmacher, Angew. Chem., Int. Ed.,
1999, 38, 1623–1626.
4 J. Kerth, T. Jikyo and G. Maas, Eur. J. Org. Chem., 2003,
1894–1903.
5 T. Jikyo, J. Schatz and G. Maas, J. Phys. Org. Chem., 2003, 16,
504–512.
6 I. Shinoda, A. Takahashi, T. Saito and T. Uchida, Bull. Chem. Soc. Jpn.,
1994, 67, 2785–2794.
7 W. J. Richter, Chem. Ber., 1983, 116, 3293–3300; K. Lammertsma, J.-
T. Hung, P. Chand and G. M. Gray, J. Org. Chem., 1992, 57,
6557–6560; B. Wang, C. H. Lake and K. Lammertsma, J. Am. Chem.
Soc., 1996, 118, 1690–1695.
8 N. S. Rao and L. D. Quin, J. Am. Chem. Soc., 1983, 105, 5960–5961.
9 F. Mathey, Chem. Rev., 1990, 90, 997–1025.
10 C. Laporte, G. Frison, H. Grützmacher, A. C. Hillier, W. Sommer and
S. P. Nolan, Organometallics, 2003, 22, 2202–2208; M. T. Nguyen, E.
Van Praet and L. G. Vanquickenborne, Inorg. Chem., 1994, 33,
1153–1157.
In addition to the mode of formation, butadienyl sulfide 7 is
an interesting compound per se, because it is a novel example of
stable thioenols (vinyl sulfides) which is obviously not in
equilibrium with the thioketone tautomer.‡ In fact, it is
vinylogous to another stable thioenol, Ph2CNC(SH)Ph.11
11 T. Selzer and Z. Rappoport, J. Org. Chem., 1996, 61, 5462–5467; Y.
Chiang, A. J. Kresge, N. P. Schepp, V. V. Popik, Z. Rappoport and T.
Selzer, Can. J. Chem., 1998, 76, 657–661; See also: S. Sklenak, Y.
Apeloig and Z. Rappoport, J. Chem. Soc., Perkin Trans. 2, 2000,
2269–2279.
CHEM. COMMUN., 2003, 2794–2795
2795