SynopsisThe common assumption in the literature is that amine complexes of Lewis acids (particularly boron trifluoride) cure epoxy resins by virtue of their dissociation to liberate the Lewis acid which then functions as the "true" catalyst. Evidence is presented to indicate that this picture of aminelewis acid complex curing may be incorrect. Another mechanism is proposed which is consistent with ( a ) lack of correlation between adduct stability and reactivity, ( b ) a similarity in reactivity of BF, adducts and the corresponding fluoroborate salts, ( e ) the relationship between reactivity of BF, and PFh adducts, ( d ) a correlation between reactivity and structure of the amine in the complex, and (e) the effect of the anion on reactivity of amine salts. The relative reactivity of the catalysts has been equated with the temperature at which gelation occurs on heating at a fairly constant rate.

Fused-ring borazines are readily obtained from the reaction of trichloroor tribromoborane with aniline ortfto-substituted by amino, alkoxy, hydroxy, or mere apto groups. They are high melting compounds relatively insoluble in most solvents, and thermally stable to 300-400°.

No abstract

Septa 5, 1958 REACTION OF B-T RICHLORO-N-TRIMETHYLBORAZENE WITH GRIGNARD RE.4GENTS 4515 boron in the chains. Characterizationz3 of the phase as a solid solution of carbon in boron is consistent with the experimental absencez4 of extrema in physical properties a t any stoichiometric composition throughout the entire range. (Although it is not a system in which one imagines that complete (23) F . W. Glaser, D. Moskowitz and B . Post, J . A p p . Phys., 24, (24) Cf., P . J. Bray, Atomic Energy Commission Report N Y O -731 (1953). 7624.thermodynamic equilibrium, involving the reconstitution and sorting of chains, will be readily attained.)Acknowledgments.-We wish to thank Dr. H. A. Weakliem, Jr., for expert aid in carrying out some of the computations and the Cornel1 Computing Center for allowing us the necessary time on IBM equipment. RECEIVED JANUARY 13, 1958B-Trichloro-N-trimethylborazene reacts smoothly with Grignard reagents t o give B-alkyl derivatives. B-Triethyl-N-B-Di-(n-butyl)-B-chloroand trimethyl-, B-triallyl-N-trimethyl-and B-tri-(n-butyl)-N-trimethylborazene were prepared. B-n-butyl-B-dichloro-N-trimethylborazene were prepared using a deficiency of Grignard reagents.

Phosphorus pentafluoride is shown to react with primary amines to give a variety of products depending upon the amine, solvent, and reaction conditions. An adduct was isolated from aniline, Z,Wimethylaniline, and n-propylamine but could not be isolated from 2,6diethylaniline or isopropylamine under the mildest conditions employed (0-10') in heptane or benzene. Disproportionation of the adduct from aniline and n-propylamine in the presence of phosphorus pentafluoride gave the corresponding amine hexafluorophosphate salt and diazadiphosphetidine. Disproportionation of the adduct from more hindered amines such as 2,4dimethylaniline, 2,6dimethylaniline, and 2,6diethylaniline gave the corresponding amine hexafluorophosphate salt and aminophosphorus tetrafluoride. Addition of triethylamine (or other tertiary amine) to the reaction of primary amines with phosphorus pentafluoride causes formation of diazadiphosphetidines from hindered as well as unhindered amines, with formation of the tertiary amine hexafluorophosphate salt. The infrared and ' OF nmr spectra are discussed for certain of the products. Dehydrofluorination of ammonia-phosphorus pentafluoride with triethylamine or diisopropylethylamine-phosphorus pentafluoride gives small yields of phosphoaitrilic fluorides but the major product is an incompletely dehydrofluorinated oil.