July 5, 1961
DIALKYLAMINO DERIVATIVES OF
.4LUMINUM
2835
about 9% of the entropy expected for a doublydegenerate ground state.
observed in the 400 cycle susceptibility indicates that any relaxation time present is substantially below a millisecond even below 1°K. This is in Discussion marked contrast to the behavior of such salts The heat capacity anomaly and the correspond- as the paramagnetic alums, whose relaxation ing decrease in xmT below 2’K. suggest that a co- times are typically above 0.01 second a t liquid operative or “exchange” interaction between the helium temperatures. l7 The existence of comagnetic ions becomes important a t these tem- operative interactions might indeed be expected peratures. Such an interaction had previously been to reduce the relaxation times. Our susceptibilities above 4’K. agree, within deduced from the paramagnetic resonance spectrum of KaFe(CN)s a t low temperatures.” The heat experimental error, with the equation given by capacity anomaly is much too small to account McKim and Wolfg for their experimental data on for the R In 2 magnetic entropy of the spin doublet potassium ferricyanide. Below 3’K. our results ground state a t high temperatures. The high are from 3 to 7% higher than theirs. The difheat capacity observed by Stephenson and Morrow ference is reflected mainly in a less rapid drop in between 50’ and 160’K. was attributed to mag- XmT for our measurements. (McKim and Wolf netic interactions, who estimated the entropy as- use an equation of different form from ours, but sociated with this anomaly as 0.09 & 0.02 e.u.12 with a correction term corresponding approximately It does not appear that their anomaly is due to to a Weiss A of 0.200, compared with our 0.140.) removal of the spin degeneracy, since even below The differences between our results and theirs 20’K. K3Fe(CN), still possesses a magnetic are of the same order of magnitude as the differences susceptibility nearly the same as that expected they observed between the susceptibility of the for a single unpaired electron. The “high-tem- powdered specimen and values calculated for the perature” anomaly may be associated with a powder from the principal susceptibilities of “quenching” of the orbital magnetic moment, crystalline specimens. McKim and Wolfg estimated the magnetic which contributes significantly to the susceptibility a t higher temperatures and may be appre- heat capacity in zero field as 0.08/T2(*20%) on ciable even a t 20°, according to the calculations of the basis of a measurement of the adiabatic suscepHoward.8 IVhatever the cause, i t is evident that tibility a t 1,000 gauss, 1.05’K. The heat capacity even the sum of the entropy decreases associated a t 1.05’K. is in fact slightly over 0.11 (their with the two regions of high heat capacity is estimate 0.07). However, this temperature is insufficient to account for all the magnetic entropy. below the maximum in the heat capacity, and the I t seems likely that there is a peak in the heat actual magnetic heat capacity in the 1/T2 region capacity below 0.6’, the beginning of which may is more than twice their estimate. Acknowledgments.-We wish to thank Profesbe the observed rise in the heat capacity curve a t sor J. G. Aston for helpful discussions, Mr. L. 0.75’. Both the heat capacity and the magnetic suscep- Sfiultz for production of the refrigerants used in tibility indicate that there are pronounced inter- this investigation and PvLr. R. G. Taylor for prepaactions between the magnetic ion and its surround- ration of the specimen. We are grateful to the ings. In the first place, the observed structure Research Corporation and Westinghouse Electric of the heat capacity curve and the hint of addi- Corporation for provision of funds for construction tional structure below 0.7’K. indicate the pres- of the iron-free solenoid magnet. ence of fairly complicated cooperative effects. (17) C. J. Gorter, “Paramagnetic Relaxation,” Elsevier Publishmg Second, the fact that no dispersion effects were C o ,New York, N. Y., 1947, pp. 77 ff.
[CONTRIBUTION FROM THE
ROHM& HAASCOMPANY, REDSTONE ARSENALRESEARCH DIVISION,HUNTSVILLE, ALABAMA]
The Preparation and Reactions of Dialkylamino Derivatives of Aluminum BY JOHN K . RUFF RECEIVED DECEMBER 19, 1960 Two different types of dialkylamino derivatives of aluminum, (R&),AlX and (R*N)BAl, were prepared. Apparel] t molecular weights of the products were determined and their structures discussed. Exchange reactions with boron trichloride w-ere interpreted in terms of reactivity of the different types of dialkylamino groups.
Dialkylamino derivatives of aluminum have been discussed only brieffy in the literature. Davidson and Brown’ first reported the preparation of dimethylamino dimethyl aluminum from dimethylamine and trimethyl aluminum. I n 1955 trisdimethylamino alane, bis-dimethylamino alane and dimethylamino a1ane2 were reported by (1) N. Davidson and H. C. Brown, J . Am. Chcm. SOL., 64, 316 (1942).
Wiberg. a Recently several other aluminum derivatives containing one dimethylamino group were ~repared.~ This paper presents further results on the preparation and reactions of aluminum derivatives containing two or more dialkylamino groups. (2) posed (3) (4)
The nomenclature used in this paper is an extension of t h a t propreviously; i b i d . , 811, 2141 (1960). E. Wiberg and A. M a y , 2 . Nolurforschg., lab, 234 (19551. J. K. RUB, J . A m . Chcm. SOL.,83, 1798 (1961).
Experimental Since the products and starting materials react with moisture and oxygen, all handling of samples for analysis was performed in a nitrogen filled dry box. Analyses.-The preparation of samples for aluminum, chlorine and active hydrogen analyses was described previously.6 Nitrogen was determined by a modified Kjeldahl procedure. The Kjeldahl apparatus was first flushed with nitrogen gas and then the weighed sample contained in a small round-bottomed flask was attached. The sample was hydrolyzed with water and then a 20% sodium hydroxide solution was added. The amine liberated was determined in the usual manner. Materials.-Trimethylamine alane and dimethylamino alane were prepared as described previously.3 Lithium aluminum hydride obtained from Metal Hydrides, Inc., was used after extraction with diethyl ether.4 Di-isopropylamine was distilled from sodium hydroxide before use and dimethylamine obtained from Matheson Co. was used without further purification. Boron trichloride was purified by passing it through two mercury-filled bubblers t o remove chlorine. Tris-Dimethylamino Alane: Method A.--A 5.5 g. sample of trimethylamine alane was treated with 100 ml. of dried benzene or diethyl ether. Dimethylamine was passed through the solution until hydrogen evolution ceased. The solvent was removed under reduced pressure and the solid residue was loaded into a cold finger sublimer. Trisdimethylamino alane sublimed readily a t 90". The product weighed 8.7 g., m.p. 87-89" (lit.3 87-88'). The yield was 90%. Anal. Calcd. for Al[N(CH3)2]3: Al, 16.98; K , 26.4. Found: Al, 16.76; N, 26.6. Method B.-A 5.0 g. sample of lithium aluminum hydride was dissolved in 150 ml. of tetrahydrofuran. Dimethylamine was passed into the mixture until hydrogen evolution ceased. A solution of 4.5 g. of aluminum chloride, in 50 ml. of tetrahydrofuran, was added slowly t o the flask. The reaction mixture was refluxed for 1 hr., and then the tetrahydrofuran was displaced with benzene by distillation. The mixture was filtered and the filtrate mas treated as described in Method -4. A yield of 16.1 g. or 7670 was obtained, m.p. 86-87'. Lithium Tetrakis Dimethylamino Aluminate .-.4 solution of 4.0 E. of lithium aluminum hydride in 200 ml. of tetrahydrofiran was allowed t o react with dimethylamine as described above. The solution was filtered and the solvenot removed in vucuo. T h e solid residue was dried a t 60 overnight under high vacuum. Anal. Calcd. for LiA1[N(CH3)2I4: Al, 12.85; Li, 3.31. Found: Al, 11.81; Li, 2.94; Li/Al = 1.04. Bis-dimethylamino A1ane.-Trimethylamine alaiie, 3.6 g., was dissolved in 100 ml. of ether. The flask was evacuated on the vacuum line. Dimethylamine, 8.1 X mole, was condensed into the flask and the mixture was allowed t o warm slowly t o ambient temperature. Hydrogen was evolved. When gas evolution had ceased, the hydrogen was removed through two liquid nitrogen traps. Any unreacted dimethylamine was then re-condensed into the reaction vessel. After no further hydrogen evolution was observed, the solvent was removed under vacuum and the solid residue was sublimed a t 40". A yield of 4 . 2 g. or 91% T m s obtained, m.p. 62' (lit.3 63"). Anal. Calcd. for HAl[N(CH3)2]2: iz1, 23.28; active hydrogen, 0.862. Found: A l , 23.39; active hydrogen, 0.856. Tris-di-isopropylamino Alane . - -Approximately 50 nil. of tlried di-isopropylamine was co:idensed into 3.6 g. of tri~nethylaminealane with liquid nitrogen. Upon warming, evolution of hydrogen occurred. The mixture was refluxed under nitrogen for 20 hr. It is important that no oxygen be allowed t o enter the system during the reflux period or dark yellow t o red oils will result when the excess amine is removed ilz vucuo. The solid residue usually obtained was sublimed a t 70". The product weighed 8.9 g., 68y0 yield, m.p. 58-59', .1nal. Calcd. for [(CH3)2CHj?X3Al:'$1, 8.26; S ,13.2. Found: Al, 8.42; S,18.1. (.5) J. K. Ruff and hl. F. H a w t h o r n e , J . A m . C h e m .Cor., 83, 31: \1961).
Dimethylamino Di-isopropylamino A1ane.-The above procedure was repeated with 3.6 g. of dimethylamino alaue and 40 ml. of di-isopropylarnine. After refluxing the mixture for 36 hr. the excess amine was removed and the solid residue was sublimed a t 95'. Li z yield of 7.2 g. of the product was obtained. Anal. Calcd. for [ ( C H ~ ) Z C H ] ~ S A ~ H N ( C Al, H ~ )15.58; Z: active hydrogen, 0.581. Found: A!, 15.50; active hydrogen, 0.562. Di-isopropylamino Alane .-Di-isopropylaniino alane was prepared from 4 .O g. of di-isopropylamnionium chloride and 1.5 g. of lithium aluminum hydride as described previously.z The crude product was purified by sublimation a t 90". The product weighed 2.4 g., m.p. 130-131°. Anal. Calcd. for H2A1X[CH(CH3),]?: 41, 20.93; active hydrogen, 1.55. Found: Al, 21.24; active hydrogen, 1.55. Reaction of Aluminum Chloride with Tris-dimethylamino A1ane.--A. A 3.2 g. sample of tris-ditnethylairiino alane and 1.33 g. of aluminum chloride were placed in a sublimer. The solid mixture was heated for 1 hr. under a nitrogeii atmosphere a t 90". The sublimer was evacuated and tlic product sublimed a t 55'. A yield of 3.2 g. of the product was obtained, m.p. 55-57'. Anal. Calcd. for C1Al[N(CHB)2],:Al, 17.95; Cl, 23.60; Al/Cl = 1.01. Found: Al, 17.96; C1, 23.48. B. The same procedure was repeated using 2.7 g. of aluminum chloride and 1.6 g. of iris-dimethylamino alane. The product was sublimed a t 90". T h e product weighed 3.6 g., m.p. 151". Anal. Calcd. for C12AlN(CH3)2: Al, 19.03; C1, 50.01. Found: Al, 19.11; Cl, 50.22. Reaction of Trimethylamine Alane with Tris-dimethylamino A1ane.--A 0.8 g. sample of tris-dimethylamino alanc and 0.9 g. of trimethylamine alane were heated together under nitrogen a t 80" for 2 hr. Trimethylamine was evolved (9.8 X 10-3 moles). The dimethylamino alane sublimed a t 40", m.p. 89" (lit.289-90'). Anal. Calcd. for .41H2N(CH3)2: Al, 37.0; active hydrogen, 2.74. Found: -41, 36.3; active hydrogen, 2.66. Reaction of Dimethylamine with Tris-di-isopropylamino A1ane.-A. A tensiometric titration was performed in an apparatus previously described.6 A solution of 0.201 g. of tris-di-isopropylamino alane in 2 ml. of n-decane was prepared in the dry box. The reaction flask was attached to the monometer system and thoroughly evacuated. Known amounts of trimethylamine were then condensed into the flask which was warmed t o room temperature and a 25" water bath was placed around it. The pressure was read when no further change was detected. T h e endpoint occurred a t a ratio of dimethylamine t o the amino alane of 2.05. B. ,4 1.40 g. sample of tris-di-isopropylamino alane was placed in a flask, and 1.02 X 10-2 mole of dimethylamine were condensed onto the sample. Upon warming to arnbient temperature a liquid phase formed. Fractionation of the volatile components yielded 1.51 X 10-3 mole of dimethylamine and a clear liquid. The vapor pressure OF this liquid was 83 mm. a t 25.3' and that of a pure sample of di-isopropylamine was 81 mm. a t the same temperature. The infrared spectrum of the liquid was identical t o t h a t of pure di-isopropylamine. Reaction of Boron Trichloride with Tris-dimethylamino Alane -A. A 0.3621 g. sample of tris-dimethylamino alane in 2 nil. of n-decane was titrated tensiometrically with boron trichloride. Apparent equilibrium was attained approximately fifteen minutes after removal of the cold bath. The observed pressure was recorded. The end-point occurred a t a mole ratio of boron trichloride to tris-dimethylamino alane of 2.10. B. Tris-dimethylamino alane, 1.6 g., in 25 i d . of pentane was treated with 1.99 X 10-2 mole of boron trichloride. Upon warming, a white solid formed.-vThe solid was isolated by filtration and sublimed a t i a , m.p. 149-150'. Analysis showed it contained no boron. The solvent was removed from the filtrate. A clear liquid remained. I t was purified by fractionation through a -63" bath, m.p. -46". Upon standing a t room temperature for several days the liquid turned into a white crystalline solid, n1.p. 41'. The melting points and the phase transition identify the liquid as dimethylamino boron dichloride.6 ((i)15. LVibPrg and 1; S c h u s t e r , Z aiiorg. Cheiw., 213, i 7 tI$l:3:31.
July 5 , 1961
DIALKYLAMINO DERIVATIVES OF ALUMINUM
C. Tris-dimethylamino alane 0.163 g. was placed in a seal-off bulb with 3 ml. of n-decane. Boron trichloride, 5.47 X 10-8 mole, was added and the flask was sealed off. After standing 24 hr. a t ambient temperature, the bulb was opened and the volatile components were fractionated. A 2.34 X 10-3 mole sample of boron trichloride was recovered. The ratio of boron trichloride consumed to trisdimethylamino alane was 3.05. Reaction of Boron Trichloride with Bis-dimethylaminochloro Mane.-A. A tensiometric titration of 0.2885 g. of bis-dimethylaminochloro alane in 2 ml. of n-decane with boron trichloride was performed as described above. The end-point occurred a t a mole ratio of reactants (B/Al0) of 1.02. B. A solution of 2.0 g. of bis-dimethylaminochloro alane in 70 ml. of pentane was allowed to react with 1.35 X mole of boron trichloride. Upon warming, a white solid formed. After purification by sublimation the solid had a m.p. of 150". A clear liquid, m.p. - 44", was isolated from the filtrate as described above. It crystallized upon standing several days. Tensiometric Titration of Tris-di-isopropylamino Alane with Boron Trichloride.-A solution of 0.696 g. of tris-diisopropylamino alane in 2 ml. of n-decane was titrated with boron trichloride. The end-point occurred a t a mole ratio of boron to aluminum of 2.90. Molecular Weight Determination.-The apparent molecular weights were determined cryoscopically in benzene. in an apparatus previously d e ~ c r i b e d . ~A sample of benzene and the cell were taken into the drybox where a solution of the desired compound was prepared. The solution contained one to three grams of the compound per thirty-five grams of benzene. The cell was loaded and the freezing point of the solution was determined. The freezing point of the solvent was determined both before and after that of the solFtion. The values for benzene were reproducible to &0.01 . Table I presents the apparent molecular weights of the compounds prepared. Included in the table are some values obtained by Wiberg2 using the ebullioscopic method in diethyl ether. The large difference in the two sets of values may be due in part to the difference in the solvent employed.
available starting materials 3LiA1[T\'(CHl),Ic
+
AlC13 --+ 3LiC1
2837
+ 4A1[X(CHJ)J3
Lithium tetrakis dimethylamino aluminate was readily prepared from dimethylamine and lithium aluminum hydride. If desired, this compound can be obtained in a fair state of purity by evaporation of the solvent. However, in the preparation of tris-dimethylamino alane, isolation of the lithium salt was not necessary as aluminum chloride could be added directly to the reaction flask containing the freshly prepared solution. X more general method of preparation of trisdialkylamino alanes was found in the reaction of secondary amines with trimethylamine alane AlHa.N(CH3)a
+ 3HNR2 --+ .4l(iYR~)7+ 3H2 + K(CH3)3
This method is an adaptation of one suggested by Wiberg.2 The use of trimethylamine alane in place of aluminum hydride permits a wider choice of solvent and reaction temperature. When dimethylamine was employed, the reaction proceeded rapidly in ether or benzene a t room temperature. I t was possible, as Wiberg observed, to halt the reaction after either one or two dimethylamino groups were substituted by limiting the amount of amine used. The reaction of two moles of dimethylamine with one mole of trimethylamine alane constitutes one of the best methods of preparation of bis-dimethylamino alane. AlH3.S(CH3)3 4 2(CHj)?SH + 2H2
+ HAI[N(CH3)2]2+ S(CH1)3
When other dialkylamines were used in the reaction, complete substitution of the hydride reTABLE I quired more drastic conditions. For example, in APPARENT MOLECULAR WEIGHTOF SOME DIALKYLAMINOthe preparation of tris-di-isopropylamino alane, O F ALUMIKUM DERIVATIVES the trimethylamine alane was refluxed for 20 hr. Mol. wt. with a large excess of di-isopropylamine before (obsd.) n n (Wibergz) complete substitution of the hydride by di-iso.41[N(CH3)::]:j 353 2.22 1.02 propylamino groups was achieved. The infrared HA1 [N(CH3)2lr 289 2.49 1.71 spectrum of the product confirmed the completeHzAlN( CH3)2 219 3.00 2.08 ness of the substitution, and the proton n.m.r. A1(N [CH(CH3)z1 1 3 347 1.06 .. spectrum indicated that no isomerization of the iso[ ( C H S ) ~lzA1C1" N 294 1.95 ... propyl groups had occurred. (CH3)zNAlClz ' 301 2.12 .. I t has been observed by several workers1,*s4 [(CH3)2CH]2NAlHz 274 2.16 .. that substitution of a dimethylamino group for an [(CH3)nCH] ~ N A l H N ( C H ~ )346 ~ 1.98 ... alkyl, aryl, halide or hydride on aluminum results a By boiling point elevation in benzene. in the formation of dimers or trimers. AssociaN .m.r. Spectrum.-The proton nuclear magnetic reso- tion is generally believed to occur through the nance spectrum of tris-di-isopropylamino alane was taken on a benzene solution with a Varian Model V4300 B spec- formation of nitrogen bridges. Both tris-dimethyltrometer operating a t forty megacycles. The spectrum con- amino alane and bis-dimethylamino alane were sisted of a large doublet and a much smaller septuplet. I t found to be associated in benzene. Tris-di-isowas not possible to compare the relative areas of the two, propylamino alane, however, was monomeric in however the observed spin-spin coupling is consistent with the same solvent. The inability of this com;it1 isopropyl group. pound to polymerize is believed to be due to steric Results and Discussion hindrance. Two possible factors are assumed to Dialkylamino derivatives of aluminum may be be important in preventing dimer formation. prepared by the reaction of a secondary amine Either the di-isopropylamino group is not capable with a suitable aluminum hydride derivative. of forming a stable bridge due to steric interaction Sodium aluminum hydride, for example, was re- of the two isopropylamino groups with the adported to yield the completely substituted sodium jacent aluminum atoms, or the shielding of the tetrakis dialkylamino aluminate when allowed to aluminum by the six isopropyl groups surrounding react with a dialkylamine.' This reaction was it prevents bridge formation. Di-isopropylamino utilized in this study as a part of a convenient alane synthesized from lithium aluminum hydride preparation of tris-dimethylamino alane from and di-isopropylammonium chloride was found to be dimeric in benzene. This indicates that the ( 7 ) A. E. Finholt, el nl., J. Inorg. nnd Nuclear Chem., 1, 317 (1955).
3s3s
JOHN
K. RUFF
Vol. s3
+
[( C H ~ ) I C H ] ~ N H (CHa)zNAlH2 +
H (CHl),XdiH[CH(CHS)r!2 -t112
Dimethylamino di-isopropylamino alane is dimeric in benzene. If association occurs through nitrogen bridges, three structures of the dimer are possible. (CIIi)?
BC ‘,/A,
Fig. 1.-Tensiometric titration of some dialkylainino aluiuinum compounds with boron trichloride.
shielding of the aluminum by the six isopropyl groups is the most important factor. The great difference in the conditions under which tris-dimethylamino alane and tris-di-isopropylamino alane are formed might in part be due to a depolymerization process. Since di-isopropylamino alane is dimeric, the rupture of a nitrogen bridge must occur in the preparation of tris-ai-isopropylamino alane. The stability of the nitrogen bridge in the dimethylamino derivatives of aluminum was dernonstrated by the quantitative displacement of trimethylamine from trimethylamine alane by a single dime thylamino group, in the preparation of dimethylamino a1aiie.j S o interaction has been observed between trimethylamine and other mono dimethylamino derivatives of aluminum. However, the apparent stability of a monomeric form of tris-dimethylamino alane in ether3 and a dimeric form in benzene indicates that the tris-dimethylamino alane may form the etherate, A1 [N(CH&]s. O(C2H5)2. The possible existence of such a species suggested treatment of tris-dimethylamino alane with trimethylamine. KOindication of interaction was observed at ambient temperature when the amino alane was titrated tensiometrically with trimethylamine. Tris-di-isopropylamino alane is monomeric and might therefore be expected to form a mono amine complex, if the amine meets the steric requirements of the amino alane. KOinteraction was observed in a tensiometric titration with trimethylamine, however. Because of the lower steric requirements of dimethylamine a reaction might be expected. After the addition of two iiioles of dimethylamine per mole of tris-diisopropylamino alane, a sharp linear pressure increase was noted. Di-isopropylamine was isolated in good yield from the reaction. I t is apparent that an exchange reaction occurred instead of complex formation
I11
Structure I is favored for steric reasons; also, if the di-isopropylamino group should be the bridging group, it is difiicult to explain why reaction with the remaining hydride does not occur, since dissociation of a diisopropylamino bridge occurs in the preparation of tris-di-isopropylamino alane. I t is felt that the increased stability of a “dimethylamino bridge” (as opposed to a “di-isopropylamino bridge”) prevents further substitution. The stability of a bridging dialkylamino group might make it less reactive towards exchange with other groups than a non-bridging dialkylamino group. In order to determine whether exchange reactions would occur, tris-dimethylamino alane was allowed to react with either aluminum chloride or trimethylamine alane. An exothermic reaction occurred between the amino alane and aluminum chloride in the absence of solvent to give aminochloro alanes in good yield. Either bisdimethylaminochloroalane or diniethylaminodicliloro alane was obtained depending upon the mole ratio of the reactants. .4lCli
+ 2[(CH3):K]3A1+3[(CHa)2S;2AICI + [(CI-13)2N],X1+3C1,AIN(CH,)2
2A1Cl3
The products are white crystalline solids which are readily hydrolyzed by traces of water. Both are dimeric in benzene. Similarly, a reaction between trimethylamine alane and tris-dimethylamino alane 2\CII,j)~XII+ ([(CI-I,),C€1]2S)3Al-+ 2[(C€Ii)rCH]zSH + [(CrI,)*N]:AlS[CH~CH~)1].produced dimethylamino alane. The trimethylAlthough the dialkylamino aluminum derivative amine was displaced during the reaction and was iormed in the reaction above was not isolated, an recovered quantitatively. effort was made to prepare bis-di-isopropylamino 2AlHn.S( C1In)a f A1 [ S ( CH:
