173
PzF4 appears to be stronger as a Lewis base toward BH3 than is PF3; slow decomposition of P2F4.BH3 by fluoride shift and cleavage of the P-P bond gives F3PBH3 which then decomposes to give significant concentrations of F3P and B2H6. On the other hand, FzPPFt.BH8 +F3PBH3(g) + PF(S) no P2F4 was ever detected in the system during deThe solid yellow (PF)* is presumably polymeric. No composition, and B2Hs appeared only after the appearevidence for decomposition to P2F4(g) and B&(g) ance of F3PBH3. These observations suggest that the was observed. It is also significant that FZPPF2, in B-P interaction in the adduct may be enhanced by sharp contrast to the related fluorides of n i t r ~ g e n , ~ partial double-bond interaction between phosphorus reacts smoothly and nonexplosively with diborane. atoms in the parent PzF4. Such interaction would A 1.03-mmole sample of B2H6 (Callery Chemical enhance the basicity of one phosphorus atom at the Co.) was frozen into a 50-ml bulb on the vacuum line. expense of the other and in a manner which contrasts A 1.04-mmole sample of P2F,' was frozen into a sharply with the basicity toward BH, of other bidentate second evacuated 250-ml bulb, connected to the first ligands such as hydrazine. bulb through a stopcock. With the stopcock closed Acknowledgment. Support of this work through between the two, the P2F4and B2H6 were allowed to a NASA traineeship held by K. W. M. is gratefully vaporize and warm to room temperature. The vapors acknowledged. were then allowed to mix by opening the stopcock Karen W. Morse, Robert W. Parry between the units. After 3-4 hr at 25", the gaseous Department of Chemistry, Unicersity of Michigan mixture was passed through traps held at - 85, - 100, Ant2 Arbor, Michigan 48104 and -196". The PzF4.BH3 retained in the -100" Receioed Nocember 4, 1966 trap was further purified by opening the -100" trap to a -196" trap for 2 6 3 6 hr. The product, a white crystalline solid at - loo", represented a 30% conCrystalline Catalysts Containing AI-0-A1 or version of P2F4 to P2F4.BH,. The molecular weight Zn-N-Zn Groups for Stereospecific as determined by vapor density was 155 (calculated Polymerization of Propylene Oxide for F4P2.BH3: 152); H- (by hydrolysis in acid soluSir : tion): 3.1H-/PzF4.BH3. Vapor pressure data were Since the pioneering work by Price on the stereonot obtained because of decomposition in the liquid specific polymerization of propylene oxide, ' a large phase to PF3BH3 and its decomposition products, number of papers dealing with the mechanism of the but the product is sufficiently volatile to pass through a polymerization as well as with the nature of catalysts trap at -85". have been published. But, unfortunately, the illMajor mass spectrometer peaks given as mass numdefined nature of catalysts and the rather low stereober (tentative assignment) and relative abundance are: specificity of polymerization make the detailed in152 (PzF4*BH3+), 1.67; 138 *(PzF4+), 8.15; 119 *(P2terpretation of the stereoregulating mechanism diffiF3+), 5.71; 83 (F?PBH3+), 8.05; 82 (FzPBHz+), 11.0; 81 (F,PIOBHp+ and F2P1'BH+), 3.43; 80 (F2Pl0BH+), cult. We now wish to report two types of crystalline catalyst for stereospecific polymerization of propylene 1.55: 69 *(F2P+), 100; 50 *(PF+), 24.5; and 31 *(P+), oxide. These catalysts are the organometallic com25.6. Much of the pattern (marked by a n asterisk) is pounds, one containing aluminum and oxygen and identical with that for PzF4. The over-all pattern is another zinc and nitrogen. consistent with the formula assigned. Infrared abBis(diethyla1uminum) oxide (Et2A10A1Et2, 11) has sorptions are listed as [frequency, cm-' (probable been mentioned as one of the reaction products of assignment), and intensity]: 2432 (vB-H), m br; 1102 triethylaluminum with carbon dioxide4 or with water,j ( ~ B - H ) , W; 1042 (6,-~), m; 902 (vas, p-F), vvs br; but has not been isolated in a pure form. It was 850 (vSym,p-p), S ; 727 ( ~ B H J m ; 670-680 (?), W, br; found, by a new route according to eq 2, that exactly 598 (YP-B?), W ; 442 ( ~ F - P F ) , VW, 395 ( T B - ~ ) , W ; 370 equimolar amounts of lithium diethylaluminate (Et,(~F-P-F), wm. The spectrum of the solid shows AlOLi, I)6 reacted in toluene at low temperature with resolution of the B-H stretching region near 2400 cm-' diethylaluminum chloride to give bis(diethyla1umiinto two distinct peaks as expected. (The symbols num) oxide (EtzAlOAlEt2, 11) accompanied by the used above are defined as follows: v = stretch, 6 = precipitation of lithium chloride in a quantitative deformation, p = rocking, T = torsional.) The amount.' This oily product disproportionated to I I B nmr taken at -80" on the neat liquid shows a quartet with JB--H = 101 cps and 6 (relative to B(1) C. C. Price and M. Osgan, J. Am. Chem. Soc., 78, 690, 4787 ( 1956). (OCH,),) equal to 60 ppm. (2) J. Furukawa and T. Saegusa, "Polymerization of Aldehydes No splitting of the 'lB signal by phosphorus is oband Oxides," Interscience Publishers, Inc., New York, N. Y., 1963, served, whereas other compounds containing a B-P pp 125-208; A. E. Gurgiolo, Reu. Mucromol. Chem., 1,76 (1966). (3) All experiments described in this paper were done under an bond of comparable stability show a definite doublet atmosphere of nitrogen or argon. patterns2 Broadening of the "B signal suggests that (4) I
