TABLE V HEATAND ENTROPY O F ACTIVATION AS*
+
OCBHa 4 BHa CO CO --c OCBHI BHa -c BHa BHa BIH~ BHs + B I H ~ BHa
+ +
+
[CONTRIBUTION
357
REACTIONS OF DIMETHYLAMINOPHOSPHINES
Feb. 5, 1962
AH*
Temp. range
Ref.
10.2 28.4 327 to 337°K. This paper -23.0 9.6 327 to 337'K. This paper Bauer: . . . . 28.4 , , .. 0 ,, . , .. . Assumed by Bauer'
..
FROM THE
. . . . . .. . . ..
DEPARTMENT
calculated to be -23.0 e.u. Table V summarizes the values of heat and entropy of activation. A rough calculation indicates, a t 60°, that the equilibrium constant for OCBH8 = BH3 CO is of the order of lod6 ( K p Jatm.) and that ka is approximately 1.6 x loa atm.-' rnin.-'. It is obvious that kz is much larger than kl and the decomposition of carbon monoxide borane to borane and carbon monoxide (eq. 1) is the slow step of the process.
+
OF CHEMISTRY, CARNEGIE INSTITUTE OF PITTSBURGH 13, PENNSYLVANIA]
TECHNOLOGY, SCHENLEY PARK,
Phosphorus Nitrogen Chemistry. IV. The Reactions of Dimethylaminophosphines with Boron Trihalides and Trialkyls1*2 BY ROBERTR. HOLMESAND RAYMOND P. WAGNER^ RECEIVED JUNE 19, 1961 The reactions of P ( K (CH3)l]ar CH3P[h'(CHa)2]2and ( C H ~ ) ~ P N ( C Hwith ~ ) Z B(CHzCHs)l, B(CH3)r, BF3 and BC1, have been investigated. Evidence indicates that B(CH2CHs)s forms liquid I: 1 adducts a t room temperature. B(CHs)a forms The reactions with BF', BC1, and excess B(CH8)a were complex and slow. Analysis of the solid 1:l adducts a t -46'. H ~ ) ~ The data indicate products yields (CHa)zNBY2, where Y is CH3, F or C1, in all cases except the P [ N ( C H ~ ) Q ] ~ - B Ccase. that the reaction ( C H D ' I , P [ N ( C H ~ ) ~(]3~~ x)BYp -+ (3 - x)(CH3)lNBYs (CI-i3)xPY~-xwhere 3c is 0. 1 or 2, proceeds CHs)a - B ( cases. t o some extent in all cases, and over i 5 % in the P[N(CHa)2]3-BFa, P[N(CHa)2&-BCls and ( C H ~ ) Z P N ( C H ~ ) ~ Secondary reactions between ( CHs)xPYa-x and the reactants led to an investigation of (CH3)PClz and (CH8)ZPCl with BCL. Solid 1:l adducts were found a t 0" which were slightly dissociated a t room temperature. A possible reaction mechanism indicated by the data is considered.
+
+
Previously a study4 of the thermal decomposi- were performed with the aid of a dry box. Samples of the phosphines, after purification in the vacuum line, were tion of borine adducts of (CH&PN(C&)Z led to weighed without exposure to air directly into ampoules the production of N-B and P-B products. Work- fitted with break tips and sealed and stored until ready ing with trimethylborane and triethylborane, 1: 1 for use. A special "reaction section" previously described5 adducts were formed in this study with the di- was used to study pressure-composition diagrams and to methylaminophosphines, P [M(CH3)2]3, CHIP [N- carry out the reactions in the systems described below. In all tables reporting pressure-composition data, the (CH&]2 and ( C H B ) Z P N ( C H ~ )however, ~; in the mole fractions refer to condensed phase values. All ratios case of the trimethylborane adducts and excess of quantities referred to in reaction mixtures are given in trimethylborane, further reaction ensued. Pre- tenns of moles. Materials.-The purification of tank boron trifluoride liminary experiments showed that such reactions and trichloride (Matheson) was described previously.6 proceed much faster with boron trichloride in the Vapor pressures of boron trifluoride samples a t -111.7" initial stages as evidenced by the rapid take up of were 297 mm. and for boron trichloride, 477 mm. a t 0'. the gaseous trichloride by tris-dimethylaminoTrimethylborane was prepared by a method described phosphine. Fractionation of the product mixture by Brown.' After fractionation the vapor pressure a t in this case led to the isolation of dimethylamino- -78.5" was 30.6 mm. (literature value,*32 mm. a t - 78.4'). Triethylborane (Callery Chemical Co.) was fractionated boron dichloride, phosphorus trichloride and small just prior to its use since increases in pressure were noted amounts of an oil. on storing.8 The fractionated material exhibited a vapor To ascertain the variety and extent of the re- pressure of 12.2 mm. a t 0' (literature value7J 12.4 mm., actions occurring as well as correlate the behavior 12.5 mm. a t 0"). The pressure would rise to about 16 mm. a t 0' after 24 hr. and to as high as 40 mm. a t 0' after a involved as far as possible, a more thorough in- month of storage at room temperature.* vestigation was conducted in which the reactions Tris-dimethylaminophosphine, P[N( CH,)21s, was preof the dimethylaminophosphine series with tri- pared by the dimethylamine reaction with phosphorus methylborane, boron trichloride and boron tri- trichloride in ether solution, similar to that described by Burg and S10ta.~ After distillation and fractionation in fluoride were studied systematically. the line using - 8 " , -78" and -196' traps, the -8' trap contained the product having a vapor pressure of 2.2 Experimental and Results Apparatus.-High vacuum systems equipped with mercury-float valves were used for part of this study. In general transfer operations outside the litle and weighings
--
(1) Previous paper in the series: R. R . Holmes, J. A m . Chem. Soc., 85, 1334 (1961). Presented before the Inorganic Division at the 140th Meeting of the American Chemical Society, Chicago, Illinois, September, 1961. (2) This paper represents part of the work submitted by Raymond P. Wagner in partial fulfillment of the requirements for the degree of Doctor of Philosophy a t t h e Csrnegie Institute of Technology. (3) Research assistant (1959-1961),National Science Foundation Grant. (4) A . B. Burg and P. 1. Slota, Jr., J. Am. Chtm. Soc., 82, 2145
(lam).
mrn. a t 20" ( l i t e r a t ~ r e2.8 , ~ mm. a t 20') and a boiling point of 161' at 736.4 mm. (literature,g 163.5 a t 760 mm.). Dimethylaminodirnethylphosphine, ( CH3)2PN(CHI)2, was prepared and fractionated according to the procedure of Burg and Slota.9 Vapor pressure a t 0" was 12.2 rnm. (literature,O 12.53 rnm. a t 0"). ( 5 ) H. C. Brown, L. P. Eddy and R. Wong, ibid., 76, 6275 (1963). (6) H. C. Brown and R. R. Holmes, ibid., 78, 2173 (1956). (7) H.C. Brown, ibid., 67,374 (1945). ( 8 ) A. Stock and F Zeidler, Ber., 64. 531 (1921). These authors observed similar behavior and showed that small amounts of CtH' and
HI were produced. (9) A. B. Burg and P. J. Slota, Jr., J . Am. Chrm. Soc., SO, 1107 (1968).
3 58
ROBERTR. HOLMES A N D RAYMOND P. W A G N E R
Methyldichlorophosphine was generously donated by the Food, Machinery and Chemicals Corporation. About 50 ml. was distilled in a simple apparatus under a dry nitrogen atmosphere. It was further fractionated in the line by allowing the vapors t o escape from a -42" trap and pass through traps cooled t o -60", -78' and -196'. The sample mostly in the -60" trap was tensiometrically homogeneous and had a vapor pressure of 25.5 mm. a t 0". However, after as little as one day the vapor pressure would rise a few mm. Hence, the compound was fractionated just prior to use. 4
1
i E
h 2 w
5
u)
Lo
w n e
VOl. 54
Experimental and Results
Dimethylaminophosphie Adducts with Triethylborane .The results (Fie. 1) of Dressure-comnosition studieq of thr systems, P [N(6H3);],-B( C H Z C H ~af) ~ 0", C H P [ S (CH3)2]2CHIB(CH2CH3h a t 0' and the (CH3)2PP\'(CH3)2-B( C H ~ ) aLt 0" and Z O O , indicate the formation of liquid 1 1 adducts a t least in the latter two systems. The P[K(CHa)z]a-B(CHnCH3)a system shows extensive dissociation. The vapor pressure of B ( C H Z C H ~a)t~ 0.0" was 12.6 mm. and a t 20.0", 38.7 mm. Values of the vapor pressures of t h e 1: 1 adducts (except for the P[S(CH3)2]3 adduct) were obtained on independent samples prepared by mixing equivalent amounts of tile components. They are 0.2 mm. a t ,O.O" for CH3P[Na t 0.0 and 2.2 mm. a t ( C H ; ) ~ ] Z . B ( C H ~ C H0.4 ~ ) ~mm. , These values 20.0 for, (CH~)~PN(CH~)~.B(CHZCH~)~. were used 111 constructing Fig. 1. The latter was necessary since the values obtained in the pressure-composition determinations may be slightly high, the determination taking several hours during which the B( C H Z C H ~pressure )~ slowly rises apparently giving small amounts of inert gaseous impurities. The latter effect was particularly noticeable in the study of the CH3P[S(CH3)2]2-B(CH2CH3)3system. Dimethylaminophosphine Adducts with Trimethylborane. Pressure-composition studies were made on all three aminophosphine systems with trimethylborane a t -46.0". Representative data is shown for the ( C H I ) ~ P XCH3)2( B(CH3)3system in Table I. The slow rise in pressure in each case to a 1: 1 mole ratio follomd by a large increase in slope with further additious serves to indicate 1: I adduct formation. The phase present a t this point was a slightly On warming to room wet white solid in each case a t -46". temperature, the 1: 1 adducts became liquid; however, the diagrams here may be complicated by further slow reaction with excess trimethylborane as discussed below.
TABLE I THE
DI~[ETIIYI,AMINODIMETHYLPHOSPHINE-TRIMETHYI
I
ANE
Mole fraction, B (CHs)a condensed phase
0
VOLE FRACTION 8 (CH2CH3)31N CONDENSED PHASE
0.066 ,155 ,276 ,344 ,402 .462 ,480
SYSTEM AT -46.0"
Press., mm.
1.1 2.1 3.9 6.3 9.8 9.3
Mole fraction, B(CHs)a condensed phase
Press, mm.
0.512 ,538 .584 ,639 ,703 , 784
78.4 132.2 187.5 210. 8 202.9 209.4
32.7 Fig. 1.-Pressure-composition studies of dimethylaminoIn any event efforts t o obtain equilibrium pressurernethylphosphines and triethylborane: 0 , P [ N ( C H ~ ) Z ] ~composition diagrams were not pursued a t this time. Instead the slow reactions were studied in sealed ampoules a t 0'; B(CH2CH3)3 a t 0'; 0 , CH3P[S(CH3)2]2-B(CH2CH3)3 0,(CH3)2PN(CH3)2-B(CH2CH3)3 a t 0'; 0 ,(CH&PN( CH8)2- which were allowed to remain a t room temperature for varying lengths of time, after which the ampoules were B(CH2CH3)3 a t 20.0". opened in the vacuum line and the contents fractionated and analyzed. Preparation of Bis-dimethylaminomethy1phosphine.The data on the sealed tube reactions are summarized i n CHaP[N(CH3)2]2,absent from the previous literature, was Table 11. The number of mmoles of reactants in the sealed prepared by the reaction of methyldichlorophosphine, ampoules is shown (column two) as well as the rn~nolesof CH3PC12,and dimethylamine in the vacuum line. (CH3)2- products formed where determined (column five). The NH, generated from the hydrochloride, passed over BaO per cent. yield is based on the amount of product formed and condensed on sodium chips (v.P., 566 mm. a t O"), according to the equation, was measured out in excess (44.26 mnioles) and condensed ( C H ~ ) , P [ X ( C H ~ ) Z I ~(-3~ - x)BYs = on 9.46 nimoles of CHaPCl? at -196". The reaction was (3 - X ) ( C H ~ ) Z X B ~ ' Z(CH3)xP\'sx allowed to proceed a t -23". A Dry Ice cone was used on the upper part of the reaction ampoule to cause the amine where x is 0, 1 or 2 and Y is a halogen atom or a methyl to reflux. After 2 hr. the contents were cooled t o -78' and group. Column four shows the moles of BY3 consumed stored for 24 hr. The aminophosphine product and excess per mole of phosphine, the latter being determined by arnine were transferred and fractionated (O', -78", - 196" fractionating the unreacted BY3 from the products and establishing its identity by vapor pressure measurements. traps), leaving the hydrochloride behind. The excess Pertinent details for each s!-stem are discussed below. amine separated showed an over-all reaction mole ratio of Aminophosphine Reactions with Trimethy1borane.amine to CH3PC1, of 3.9: 1 (4.0:1 expected for 100%,yield). The reaction of ( CH3)zPP\'(CH3)2with B( CHI)3was straight The clear liquid in the -78" trap was further fractionated CH3)2 and (CH3)3PB(CH3)3 ( O O , -23", -78", -196" traps). The -23" trap caught forward yielding ( CH3)2NB( as the only isolatable products, the former identified by the bulk of the product and was tensiometrically homogeneous. Vapor pressures were 1.2 mm. a t 0.1", 1.8 mm. a t vapor pressure, molecular weight and analysis. 5.0", 2.7 mm. a t 10.Oo, 3.7 mm. a t 15.0°, 5.5 min. a t 20.0"; T e m p . , ' C . -45.2 -21.8 -15.0 -10.0 0 15.0 19.8 b.p., 141'at 755mm.; m.p., -52'(floatmethod). 3.1 16.2 23.3 31.1 5 3 . 6 1 1 4 . 8 139.2 Press., mm. Anal. Calcd. for (CHs)P[N(CH3)2],: C, 44.76; H, 11.27: Press., mm. (calcd.)lo 2.8 14.7 22.5 3 0 . 2 5 2 . 6 1 1 1 . 1 185.8 Found: C, 45.03; H, 11.15.
+
+
REACTIONS OF DIMETHYLAMINOPHOSPHINES
Feb. 5 , 1962
Expt.a no.
359
Yield,
70
1
2
12.5
3
78.3
4
91.0
5
27.8
6
7
92.7 90.1
8
9 a
All experiments were conducted a t room temperature except no. 7 (studied a t 0 " ) .
I,
Product indicated.
Anal. Calcd. for (CH3)&B(CH3)2: C, 56.54; H, 14.24; its vapor pressure of 370.1 mm. a t -111.7" and 187.6 mm. N, 16.49. Found: C, 56.33; H , 14.02; N, 16.24. Moleca t -119" and molecular weight, 88.4 (calcd., 87.97). ular weight (vapor density), 85.5 (calcd., 84.97). The solid present was sublimed to a region on the stem The ( CHI)IPB(CH3)3 was identified by vapor pressure, of the ampoule which was cooled with a Dry Ice cone by heating the ampoule t o 40'. Analysis of the solid showed 1.94 mm. a t 26.5" (Sujishill reports 1.86 mm. a t 30.0"), it to be (CH3)2NBFz, first described by Brownla and later and analysis. Anal. Calcd. for (CH3)3PB(CH3)3: C, 54.59; H , 13.74; more fully by Elurg and Banus.lO A yellow oil remained. Anal. Calcd. for (CH3)2NBF2: C, 25.86; H, 6.51; P , 23.47; B, 8.20. Found: C, 54.25; H, 13.35; P , 23.28; N, 15.08; B, 11.65; F , 40.91. Found: C,25.61; H , 6.69; B, 8.49. The data illustrate the following slow reaction occurring N, 15.07; B, 11.83; F,40.76. In the CH3P[N(CH3)42 reaction the solid ( CH3)?NBFZ (CHa)aPN(CHa)z 2B(CH3)3 + was isolated by sublimation from the reaction mixture, the (CHa)aPB(CHa)a (CHs)zNB(CHs)z latter being heated toward 100' to complete the separation. The reactions of B(CH3)3 with CH3P[,N(CH3)2]z and The solid was trapped by a Dry Ice cone placed on the upper P[N(CH3)2]3 were more complex yielding in each case a stem of the ampoule. Excess BF3 passed through and was colorless liquid and an uncharacterized, non-volatile white fractionated and measured. The water insoluble S B F Zwas cut off, weighed and analyzed. solid. In addition a small amount of (CH3!2NB(CHa)2 Anal. Calcd. for (CHa)zNBF2: N, 15.08; F , 40.91. was isolated from the CHaP[N(CH3)2]2 reaction (vapor Found: N, 15.31; F, 40.58. pressure, 3.5 mm. a t -45.0' and 33 mm. a t -9.0D), but it was not searched for properly among the products The remaining liquid was caused t o reflux a t 140'; of the P[N(CH3)2I3reaction with B(CHa)3. however, it was so viscous that it remained on the walls Analysis of the liquid from the CHaP[N(CH3)2]2 reaction of the ampoule behaving more like a glassy solid. Congave theseresults: Anal. C, 42.23; H , 11.91; N, 21.86; P , tinuing the heating to 200" caused small additional amounts 14.00; B, 10.32. The total is 100.32% and gives a calcuof BF3 and (CH3)2NBF2t o be driven off. A similar lated empirical formula, C~.~BHZB.~~NS.~~P~.OOBZ.~~. The reaction of BF3with (CHs)zPN(CHI)*behaved analoanalysis was obtained for the liquid from the P[N(CH3)2]3 gously to the CH3P[N(CH3)2]z reaction, producing a reaction. Anal. C, 48.05; H , 11.57; N, 18.71; P , 12.42; viscous liquid and a solid. The latter was separated as B, 8.64, totaling 99.3970. The calculated formula is C10.00 before by sublimation a t 100" but only a small amount (0.36 H28,soN3.33Pl.ooB~.oo. The formulas, however, do not lend g . ) isolated. It was insoluble in water and soluble in themselves toward facile interpretation. fluorine boiling water, characteristic of ( C H ~ ) Z N B F ~A. ' ~ Aminophosphine Reactions with Boron Trifluoride analysis was low (Calcd.: F, 40.91. Found: F , 37.79); Table I1 shows that (CHa)&BF2 was isolated from each of however, considerable difficulty was experienced in separathe aminophosphine reactions with BF3; however, in the tion. case of CHIP[?;( CHa)2]2 and ( CH3)*PN(CH3)2 further A yellow solid remaining was separated from the viscous analysis was complicated by the presence of viscous liquid liquid by inverting the ampoule after resealing and placing products. The pertinent features of these reactions are dis- its stub in a -196" bath. The liquid was transferred and cussed. the solid removed and analyzed. In the P[N(CH3)2]3 reaction the gases pumped off conAnal. C, 20.76; H , 6.06; N, 3.57; P, 18.83: B, 3.66; sisted of a mixture of excess BF3and PF3. Pyridine proved F , 46.50. The analyses totaled 100.38y0 and corresponded effective in separating the gases by forming a stable complex to the empirical formula, (25.3 HI^.^ N1.o PI.$B1.o F7.5. with BFa6, the PF3 being unaffected. The gas mixture Along with the formation of (CHs)zNBFz, CHsPF2 in was bubbled through pyridine in a small U-trap cooled t o some form is expected as a product of the reaction of CHIP-30". A Dry Ice cone on the low pressure arm of the trap [N(CH3)2]~ with BF3. The compound C H 3 P F ~ has been prevented pyridine from escaping. PF3 was identified by reported however, recentlyla and shows a b.p. of -28';
+
+
(10) Calculated from the equation of A. B. Burg and J. Banus, J. A m . Chcm. SOL,. 76, 3903 (1954). (11) S . Sujishi. P h . D . Thesis, Purdue University, 1949.
(12) J. F. Brown, J . A m Chcm. Soc., 74, 1219 (1952). (131 V. N. Kulakova, V. M. Zinov'yev and L. 2 . Soborovsikiy, Zhur. Obshchei Khim., 29, 3957 (1959).
360
ROBERTR. HOLMESAND RAYMOND P. WAGNER
a search for it proved fruitless. Indications are then that
the viscous liquid produced in the CHaP[N( CH3)2]2reaction may contain some complex or complexes of CH3PFz such as CHIPFYBFIor CHaPFz~CH3P[N(CHa)z]z. Similarly the solid, Cs.,Hls 4Nl.&'1.DBl.oF7,6, and viscous liquid from the (CH~)~PN(CHI)I reaction may contain (CHS)ZPF in some form. The latter compound has not yet been described in the literature although F. Seel and cow o r k e r ~claim ~ ~ it can be prepared from (CH3)2PCl using KSOZF; however, they give no indication of having actually prepared it. Possibly by using stronger Lewis bases such as (CH,),P, the fluoromethylphosphines could be displaced. Aminophosphine Reactions with Boron Trichloride.The initial reaction of P[S(CH8)2]aand BCls was much more rapid than the corresponding reaction with B( as shown by preliminary pressure-composition studies. At 0" BCla is taken up rapidly forming a solid with a slow rise in pressure to a 3 . 1 mole ratio. Afterwards, the pressure increases sharply with each further addition and the system becomes liquid-solid. At -46' the change in slope a t 3 : 1 is much more distinct and the pressure rises, thereafter, to the vapor pressure of pure BCk. In one run carried out a t 0" (Table 11) the contents of the anipoule were analyzed. The quantitative separation proved tedious and the details are given elsewhere.16 All volatile products were transferred out of the ampoule a t room temperature leaving a white solid and a small amount of a green oil. The unreacted BCL was separated by collecting it in a -196" trap. All other volatile products were trapped a t -78". On standing solid formed in the -78' trap which proved to be a result of liquid (CH3)ZNBCln dimerizing. The latter combined with the white solid remaining in the ampoule (also dimeric (CH3)aNBClz which was separated from the small amount of oil with CCl,) was analyzed: m.p. 138-141' (literature value,18 142'). Brown and Osthoffln previously established the dimer formulation for the solid on the basis of molecular weight and dipole moment measurements. Anal. Calcd. for [(CHs)zNBC12]2: C, 19.09; H, 4.81; N, 11.13. Found: C, 19.08; H,4.84; N, 11.08. The liquid remaining in the -78' trap, after refractionation, was tensiometrically homogeneous, and its vapor pressure curve closely followed that of PCL. The molecular weight was determined (vapor density), 137.7 (calcd. The for PCk, 137.3), normal b.p., 75-76" (PCla, 74.2'). over-all reaction corresponded closely to P[N(CHa)2]3
+ 3BCL +PCla + 53 [(CH#)&"CLIZ
Vol. 84
The data obtained indicate the reaction CHilP[N(CHa)2lz 4- 3BC4 + CHsPClz.BCls [(CHs)zNBCI,]r After recovery of unreacted BC13 in the (CH8,)?PN(CH3)2 reaction there remained a mixture of solids in the ampoule. Treating the mixture with 5% NaOH solution produced some solution. The insoluble fraction was filtered, washed and dried. Analysis showed it to be [( CH3)zNBClzlz. Anal. Calcd. for (CH3)2X'BClz:C, 19.09; H , 4.81; N, 11.13; C1, 56.37; B, 8.60. Found: C, 19.90; H, 4.98, N, 10.82; C1,55.67; B,8.73. The solid portion, soluble in 5y0 NaOH, was suspected A pressure-composition study of to be ( CHZ)ZPCI.BC~~. the latter system a t 0' showed the existence of a 1: 1 adduct similar to that described for CHaPC12.BC13. The adduct was soluble in 5% NaOH. The (CHI)~PCIfor the latter study was prepared according to the piocedure of Burg and S1ota.O The latter authors have found (CH3)2PCl to react vigorously with (CH3)2PN(CH3)z to form a nonvolatile adduct, (CH3)2NP(CH3)2(CH3)2PCl.* This fact shows that either the initial reaction between (CH3)zPN(CH!)2 and BCla is sufficiently fast forming some mtermediate before the production of (CH3):PC1 so as not to CH3)2PCl or permit the formation of ( CH3)oNP(CH3)2( that the (CH3)zPCl adduct with BCL i s more stable than (CHa)zNP(CH3)2(CH3)zPC1. Another possibility is the displacement of ( CH3)2PCI from (CHa)aNP(CHI)P(CII3)zPCI by BCla.
+
Discussion In general the dimethylaminophosphines can be said to react according to the equation
+
(CHs)xP[N(CH,)21a.x (3-x)BYa ----) (CH&NBYn (CH~),PYI.~ where Y is -CH3, F or C1
+
Except for the P[N(CH3)2]a-BFI, P[N(CH3)2]8BC13 and (CH&PN(CH~)Z-B(CH~)~ cases, there are many and varied side and competing reactions. Such reactions as those between BY3 and (CH3)2NBYz, B P I and (CH3).PY~3..,, (CH&NBYz and (CH3).PY(3-,), and (CH3).PY(z-.) with ('2Hd.P[N(CH&]3-. will have to be characterized in separate studies before the dimethylarninophosphine reactions can be fully understood. Possible Course of the Reaction.-Burg and Banuslo have studied the formation and decornposition of (CH3)2NB(CH3)2BF3. In this adduct the most probable structure is a N-BF3 linkage since no other electron pair donors are available. The products of decomposition were almost exclusively (CH&BF and (CH3)zNBFz suggesting a fluoride-shift reaction. Fluoride shifts were also found in the [(CH&N]&H-BF3 reaction which gave HSiF3.l' The Si-H group is isoelectronic with phosphorus but has no unshared electron pairs. Here again BFI attack a t nitrogen is reasonable with fluorides migrating to SiH. The aminophosphines are seen to have two possible sites for attack, nitrogen and phosphorus. A t present insufficient data is available to decide the point of initial attack by a boron halide or alkyl;'* however, the over-all reactions proceed
The reactions of CH?P!N(CHa)2]2 and (CH*)?PN(CHa)z with BC13 proceeded similarly producing [(CHs)?NBClzlz ill each case and suspected BCL adducts with methylchlorophosphines (Table 11). The amounts of products, however, were not determined. Product identifications in each system were made as follows: In the CH3P[N(CH3)?Izreaction, in addition to unreacted BC13, the volatile fraction contained CH3PC12, vapor pressure, 25.8 mm. a t 0.0". Anol. Calcd. for CHaPCh: C, 10.27; H, 2.59; C1, 60.66. Found: C. 10.57; H, 2.75; C1,60.36. I t was suspected that CH~PCIZ and BCla were associating as noticed by a drop in pressure when the separated CH3PC12 and BCll were added together. The over-all ratio of 3.23 to 1 is higher than expected indicating some BC& to be tied up with CHsPC1z. The existence of CHaPC1yBCl3 was substantiated by studying the system a t 0'. A 1: 1 white solid was clearly shown (a solid-liquid plateau present, 24 mm., followed by a sharp rise in pressure a t 1: I). The solid remaining in the ampoule was washed out with distilled water, filtered, dried and analyzed to be [(CH3)2KBClzlz. (17) A. B . Burg, "Fluorine Chemistry," J. H . Simons, ed., AcaAqzal. Calcd. for [(CH3)2NBC12]z: N, 11.13; C1, 56.37. demic Press, Inc., New York, N. Y., 1950, p. 109. Found: N, 11.07; C1, 56.34. (18) However, previous work has indicated P-B bonding t o be pre(14) F . Seel, H . Jonas, I,. Riehl and J. Langer, Angaw. Chsm., 67, 32 (185.5).
(15) R . P . Wagner, P h . D . Thesis, Carnegie Institute of Technology, 1961. (16) C. A. Brown and R . C. Ostho5, J . A m . Chsm. Soc., 7dP 2340
(lesa).
ferred in diborane complexes. Thus X-ray information is consistent with the formulation (NHs)SBHa.19 Indirect evidence's has been used to infer P-B bonding in [N(CHa)r]rP.BHa. In (CHa)rPN(CHa)r (BHa)a4, thermal decomposition yields products consistent with the formation of P-B bonding at some point as well as N-B bonding. (19) C. E. Nordman, A c f o C r y s f . ,lS, 535 (1960). (20) T. Reetz and B. Katlafsky, J . Am. Chrm. Soc.. 84,6036 (1960).
Feb. 5, 1962
FACTORS AFFECTING THE TRANSMETALATION REACTION
361
analogously to those described above. At some with additional reaction proceeding very slowly; point a N-B bond is formed and either halogen however, no adducts were obtained with the boron migration or methyl migration occurs to give the halides apparently due to the much higher reactivity encountered. products, (CH3)zNBYz and (CH3)xPYa-x. Acknowledgment.-The authors are indebted to In the aminophosphine reactions with B (CH8)a 1:1 adducts may serve as initial steps. Studies the National Science Foundation for a grant which a t -46' showed that such adducts were isolatable helped this work in part.
[CONTRIBUTION FROM
THE
DEPARTMENT OF CHEMISTRY OF THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE 39, MASSACHUSETTS I
The Preparation of Organolithium Compounds by the Transmetalation Reaction. IV. Some Factors Affecting the Transmetalation Reaction1 BY DIETMARSEYFERTH AND MICHAEL A. WEINER RECEIVED JUNE 23, 1961
Evidence is presented to show that the exchange reaction occurring between organolithium compounds and allyl- and vinyltin compounds is an equilibrium reaction. The position of the C==C bond in olefinic organotin compounds is an important factor with regard t o the transmetalation reaction, since 3-butenyl- and 4-pentenyltin compounds were found not t o react with n-butyllithium in ether. The action of phenyllithium on triphenylvinyllead gave vinyllithium and tetraphenyllead; on the other hand, phenyllithium added across the C=C bond of triphenylvinylgermane.
I n previous papers of this series practical laboratory directions were given for the preparation of vinyllithium allyllithium and methallyllithiurn by the exchange reactions occurring between vinyl-, allyl- and methallyltin compounds and phenyland n-butyllithium R&n
+ 4 CsHsLi+(C&J4Sn + 4 RLi (R = CHFCH-
and C H F C H C H ~ )
preparative value. For this reason the solubility factor was examined. I n the systems n-Bu,SnCH=CHa
+ C&Li
Et20
-+
+ CHFCHLi
(a)
+ n-BuLi -+ n-BulSn + C H F C H L i
(b)
n-BuaSnCa, n-Bu&nCH=CH,
Et20
all components are soluble in the ether medium. Tri-n-butylvinyltin and phenyllithium were allowed to react in ether solution during 2 hr.; the mixture then was treated with acetone and hydrolyzed. The products isolated included dimethylvinylcarbinol (63%), tri-n-butylphenyltin (73%) and tri-nbutylvinyltin (12%). Similarly, reaction b, in which the vinyllithium produced was characterized by its reaction with triethylbromosilane, gave tetran-butyltin in 87% yi,eld, as well as triethylvinylsilane in 65a/, yield. Similar observations were made when the reaction of allyltri-n-butyltin with phenyllithium was studied; tri-n-butylphenyltin was isolated in 78% yield. Obviously the insolubility of tetraphenyltin is not the most important factor, since good conversions are obtained even when the organotin products are soluble in the reaction medium. The presence of the starting material, (CH1=CH),Sn + 4 CeH,Li + (CaH6)Sn 4 C H F C H L i tri-n-butylvinyltin, has been noted among the products of the reaction between phenyllithium and went to completion for all practical purposes, tetra- tri-n-butylvinyltin. In order to ascertain whether phenyltin being formed in virtually quantitative this represents an equilibrium situation, the reverse yield. The insolubility of tetraphenyltin in ether4 reaction, that between vinyllithium and tri-nmight possibly have been the factor which made butylphenyltin, was examined. In addition to trithis reaction proceed to near completion, and the n-butylphenyltin (75Yo), a liquid representing ca. possibility existed that in a system where both 5% yield of crude tri-n-butylvinyltin was obtained. products were soluble the equilibrium (if one existed) This material was purified and identified positively might be such as to make the transmetalation of no as tri-n-bu tylvinyl tin. To substan tiate further this evidence, the forward reaction was carried out (1) Part 111, D.Scyferth and M. A. Weiner, J . Org. Chem., $6, 4797 (1961). again, using the same reaction conditions. In this (2) R. G. Jones and H. Gilman, Chcm. Reus., 64, 863 (1054). case tri-n-butylphenyltin and tri-n-butylvinyltin (3) D. Seyferth and M. A. Weiner, J. Am. Chcm. SOC.,83, 3583 were isolated in yields of 76976 and lO%,respectively. (1961). An indirect, though substantial, piece of evidence (4) W. Strohmeier and K. Miltcnberger, Chem. Ber., 91, 1357 (1958). in favor of an equilibrium situation was observed Transmetalations of this type are of value, since they make available organolithium reagents, the preparation of which either is difficult or impossible b y the more conventional procedures. I n order to make the most efficient use possible of the transmetalation reaction for this purpose, we have studied in greater detail the factors which affect this reaction. The results of some of our studies are discussed below. Equilibria in Transmetalation Reactions.-It has been reported2 that some transmetalation reactions represent equilibrium situations. This raised the question as to whether or not the exchange reaction between vinyltin and allyltin compounds and organolithium compounds might also involve an equilibrium reaction. As was shown in Part I,3 the reaction
+
