586
Organometallics 1987, 6, 586-591
Generation and Chemical Trapping of ( 1,2-EthanedOphosphinidene)tetracarbonyltungsten Claude Charrier, Nicole Maigrot, and Franqois Mathey" Laboratoire CNRS-SNPE, BP No. 28, 94320 Thiais, France Received July 22, 1986
A series of l,n-bis(3,4-dimethylphospholyl)alkaneshas been synthesized by reaction of (3,4-dimethylphospholy1)lithium with 1,n-dibromoalkanes. The cleavage of these bis(phospholy1)alkanesby lithium in THF generally affords pure (3,4-dimethylphospholyl)lithiumthat has been characterized for the first time by 'H and 13CNMR spectroscopy. However, when n = 3, the cleavage of the 1,3-bis(phospholyl)propane leads to a phosphabicyclo[3.3.0]octene that results from the intramolecular cyclization of a transient l-(3-lithiopropyl)phospholeby addition of the carbanion onto the phosphole dienic system. The 1,2bis(3,4-dimethylphospholyl)ethanereacts with W(CO)6at 130 O C to give the P,F"-W(CO)4chelate complex, This complex in turn reacts with dimethyl acetylenedicarboxylate to give cleanly the corresponding [ 1,2-bis-(7-phosphanorbornadien-7-yl)ethane]tetracarbonyltungsten complex through [4 + 21 cycloaddition with the dienic systems of the two phosphole rings. This new complex is an efficient generator of (1,2, n ethanediphosphinidene)tetracarbonyltungsten,P-CH2-CH2-P=W (C0I4,at 130 "C in toluene. This diphosphinidene complex has been trapped by methanol, diethylamine, tolan, and cyclooctene to give a series of new W(CO)4complexes including the fiist known 1,2-bis(phosphirenyl)-and bis(phosphirany1)ethane derivatives. In a series of recent papers, we have demonstrated that transient phosphinidene complexes [R-P=M(CO),] (M = Cr, Mo, W) as generated by thermal decomposition of the appropriate 7-phosphanorbornadiene complexes have a rich and versatile chemistry. Indeed, they readily insert into O-H,l N-H,' and some activated C-H bonds,2 and they form three-membered rings with C=C3 and C=C4 multiple bonds. The experiments were conducted with some simple R substituents at phosphorus such as methyl and phenyl. In a subsequent step, we started to investigate the chemistry of terminal phosphinidene complexes with functional alky15v6and alkoxy groups7 and noted some interesting variations in the reactivity of these species7and some original rearrangements.6i6 As a logical further step in this program, we present here our work on the generation and chemical properties of the first reported chelating I
.1
diphosphinidene P-CH2-CH2-P=W(C0)4.
Results and Discussion Our general aim was to study the properties of P(CH2),-P vl(P),vl(P')-complexes. According to our previously established route to terminal phosphinidene com~ l e x e s , l -the ~ necessary starting points for building generators of such species were the still unknown 1,n-bis(3,4-dimethylphospholyl)alkanes.These compounds were prepared according to eq 1. The yields of 2-5 from 1 are calculated for the chromatographed products. The lower yield of 3 is due to a higher sensitivity of this phosphole toward oxidation during the purification procedure and not to the appearance of side products during the synthesis in this particular case. Similar l,n-bis(2,3,4,5-tetraphenylphospholy1)alkanes had been previously reported by Braye? but these phospholes are useless for our purpose since their dienic systems are deactivated by the phenyl substitution. Having at hand these new 1,n-bis(phos(1)Marinetti, A.;Mathey, F. Organometallics 1982,1 , 1488. (2)Svara, J.; Mathey, F. Organometallics 1986,5,1159. (3) Marinetti, A.;Mathey, F. Organometallics 1984,3, 456. (4)Marinetti, A.;Mathey, F.; Fischer, J.; Mitschler, A. J. Am. Chem. SOC.1982,104,4484. (5)Deschamps, B.;Mathey, F. J. Chem. SOC.,Chem. Commun. 1985, 1010. (6)Svara, J.; Marinetti, A.; Mathey, F. Organometallics 1986,5,1161. (7)Alcaraz, J. M.;Svara, J.; Mathey, F. Nouueau J. Chim. 1986,10. (8)Braye, E.H.; Caplier, I.; Saussez, R. Tetrahedron 1971,27,5523.
Me
Me
Me
Me
I
Ph 1 Me
Me
2,n.l (overall yield 3 , n = 2 (overall y i e l d 4 , n = 3 (overall y i e l d 5 , n = 4 (overall y i e l d
53%)
38%) 63%)
60%)
pholyl)alkanes, we decided to investigate their properties a little bit further before going on with our initial program. Through the cleavage of their two exocyclic P-C bonds by lithium, these species might be used as convenient starting points for the synthesis of 1,n-dilithioalkanes (eq 2). We thus attempted to cleave these bonds, and, contrary to our expectations, we observed the consumption of only 2 equiv of lithium/rmol of bis(phospholy1)alkane instead of 4. When n = 1, 2, and 4, the end product was pure (3,4-dimethylphospholy1)lithium (6) (eq 3). Obviously, the weakly aromatic transient l-(w-lithioalkyl)-3,4-dimethylphosphole 7 decomposed very rapidly to give the highly aromatic phospholyl aniong and volatile hydrocarbon by products. This result offered us the opportunity to characterize more fully the phospholyl anion 6. Indeed up to now, such species have been characterized only by their 31PNMR spectra.lOJ1 Undoubtedly this is due to the fact that they had never been obtained in the pure state previously. A pure solution of 6 in perdeuteriotetrahydrofuran was analyzed by lH, 13C,and 31PNMR (9)For a discussion on the aromaticity of phospholes and phospholyl anions see: Mathey, F. Top. Phosphorus Chem. 1980,10,1. (10)Quin, L. D.; Orton, W. L. J . Chem. Soc., Chem. Commun. 1979, 401. (11)Charrier, C.; Bonnard, H.; de Lauzon, G.; Mathey, F. J . Am. Chem. SOC.1983,105,6871.
0276-7333/87/2306-0586$01.50/0Q 1987 American Chemical Society
Organometallics, Vol. 6,No.3, 1987 587
(1,2-Ethanediphosphinidene)tetracarbonyltungsten Me
Me
of an excess of methanol (eq 5). The yield of this trapping reaction appeared to be very poor. Thus it was quite obvious that this classical scheme was of no practical use.
Me
room temp
Me
Me
Me
[V] 6
Me
Me
Me
(CHZ),LI
7 C71 t ~ L isI 6
+
Li(CH2)nLI
-!
(3)
6 t C(CH2),3
C71
(2)
n.1.2.4
Me 7 n: 3
11 (30%)
M eo,
A
(4)
Hf
/P
8
0
spectroscopy. The 'H NMR data are surprisingly similar to those of a 3,4-dimethyl-substitutedphosphole: 6 (C4DsO), 6 1.98 (s br, 6 H, Me), 6.22 (d br, 2J(H-P) = 39.5 Hz, 2 H, =CH); 1 (CDCl,), 6 1.91 (dd, 6 H, Me), 6.36 (dd, V(H-P) = 38 Hz, 2 H, =CH). On the contrary, the 13C NMR spectrum of 6 shows one very characteristic feature. The intracyclic lJ(C-P) coupling constant is huge (45 Hz), whereas it is low in tervalent phospholes12(e.g., 7.3 Hz for 1). There is here an obvious similarity with phosphaalkenes in which the doubly bonded phosphorus and carbon are strongly coupled13 [e.g., 43.5 Hz for mesityl(diphenylmethy1ene)phosphinel. This result is a consequence of the high electronic cyclic delocalization within 6 that imparts some double-bond character to the P-C bonds of the phospholyl ring. When n = 3, the corresponding 1(lithioalkyl)-3,4-dimethylphosphole7 decomposes via a different route. The carbanionic side chain attacks the phosphole dienic system to give an original bicyclic phosphine 8 that has been characterized as its P-sulfide 9 (eq 4). The formula of 9 has been established by elemental analysis and by exact mass measurement (calcd 186.0631, found 186.06356). The 'H NMR spectrum indicates the absence of vinylic proton, and the 13C NMR spectrum shows two inequivalent methyls, four methylene groups, one CH group, and two inequivalent fully substituted vinylic carbons. A similar addition of n-butyllithium onto the dienic system of l-n-butyl-3,4-dimethylphosphole has already been described in the literature.14 Coming back to our initial programm, we then turned our attention toward the synthesis of l,n-bis(7-phosphanorbornadien-7-y1)alkanes.Following the general scheme, we first allowed 3 to react with an excess of W(C0)6THF and obtained the expected complex 10 in fair yield. Then 10 was converted into the corresponding 7-phosphanorbornadiene complex 11 by reaction with dimethyl acetylenedicarboxylate. Unfortunately, the yield of this conversion was poor. Finally, in order to check the efficiency of 11 as a'generator of [ (OC)5W=P-CH2-CH2-P= W(CO),], we performed its decomposition in the presence (12) Bundgaard, T.; Jakobsen, H. J. Tetrahedron Lett. 1972, 3353. (13) Klebach, Th. C.;Lourens, R.; Bickelhaupt, F. J.Am. Chem. SOC. 1978,100,4886. (14) Mathey, F.;Mankowski-Favelier,R. Org. Magn. Reson. 1972,4, 171.
C02Me
-CH2-
cocgw
CH2 -P
(5)
I" W(CO),
C02Me
12 (18%)
In our initial work on the synthesis of 7-phosphanorbornadiene complexes,15we showed that the steric bulk of the substituent a t phosphorus had a strong adverse effect on the yield of the Diels-Alder cycloaddition between the phosphole complex and dimethyl acetylenedicarboxylate. Thus, we thought that the main reason for the low yield of 11 lay in the poor accessibility of the two sides of the dienic systems of 10. Consequently, we decided to replace the P,P complex 10 by the chelate complex 13 for which one side of both dienic systems is well exposed to external attack. This proved to be a good choice (eq 6). That the dienic systems of 13 are quite reactive was immediatelyobvious when we performed the complexation of 3 by an excess of W(CO)6: in this case, besides 13, we obtained the bimetallic complex 14 in which one of the dienes is ?r-complexed. The reaction of 13 with dimethyl acetylenedicarboxylateproceeded much more rapidly than the reaction of 10 and gave a much better yield of the 7-phosphanorbornadiene chelate 15. According to its 31P NMR spectrum, 15 is symmetrical with two equivalent phosphorus atoms (6(31P)+244.4 in C6D6, 1J(31P-183w) = 232 Hz). Moreover, the attack of the alkyne has taken place on the sides of the dienes that are a t the opposite of the W(CO)4moiety as evidenced by the very low V(CP) couplings of the vinylic carbons bearing the C02Me groups (G(CC0,Me) 146.37 in C6D6, 2J(c-P)E 2.4 HZ). These low couplings are correlated with the =C-CP-W dihedral angles and are characteristic of this stereochemi~try.~ 15 is practically the only isomer formed in this reaction according to the 31PNMR spectrum of the crude reaction mixture. The thermal decomposition of 15 a t ca. 120-130 OC proved to be an excellent method for generating the unstable (1,2-ethanediphosphinidene)tetracarbonyltungsten (16) (eq 7). This diphosphinidene complex has the same varied chemistry as the more classical terminal monophosphinidene complexes (eq 8-1 1). Apparently, the reaction with methanol gives a single isomer of 17 according to the 31PNMR spectrum (G(31P) +138.35 in C6D6,'J(P-H) = 350 Hz). However, the 31Pdecoupled 'H NMR spectrum shows two singlets in a 1:l ratio at 3.02 and 3.08 ppm in C6D6for the methoxy groups (15) Marinetti, A.;Mathey, F.; Fischer, J.; Mitschler, A. J. Chem. SOC., Chem. Commun. 1982,667.
588 Organometallics, Vol. 6, No. 3, 1987
Charrier et al.
Me W(C0)g. toluene 130 OC
Me
Me
Me
13
1
14
Me OzC-C C-COZMe 80 *C. 3 h
Me
Cl(H)P-CH,-CH,-P(H)Cl
16
ligand (eq 12). On the other
21 (85%)
hand, iodine converts 19 into the tungsten(2+) complex 22 (eq 13). The 31PNMR spectrum demonstrates that
I
Ph
178. b (40%)
I2
l9
C H ~ C I room ~ . temp, 4 5 m:i
Ph
Ph 22(54%)
[16'
1
18 a, b (55%)
:p