Spirocyclic and bicyclic phosphazenes

Abstract

The experimental results obtained in the present investigation clearly indicate that it is easy to replace two chlorine atoms in the reactions of, trimeric chloride, N P Cl^, with the bifunctional reagents, ethylenediamine 3 3 6 and ethanolamine. The derivatives obtained correspond to the formula, H 3P5(xaH2CH2NH)Cl4, (where X = NH or 0). Further replacement of chlorine atoms to obtain crystalline compounds, seems to be difficult, all such attempts made under a variety of experimental conditions lead to the formation of sticky resinous materials. However, in the ethanolamine system only, two crystalline compounds, N 3P3(O0H2CH2NH)2Cl2 (13 and 14), have been isolated in small quantities (< 5$). Whereas the geminal derivatives, N ^ G l ^ X g , (X = NHBu* or Ph) react readily with ethylenediamine or ethanolamine with the replacement of two chlorine atoms, the reactions of nongeminal 2 -trans-4 -N,P,Cl.(NMe2)2 (5) with these amines proceed much less rapidly and yield only non-crystalline resinous materials. A tris-chloro-derivative, N^P^EHCHgCHgBTH) (NHBu’b) Gl^ ( 9 ) has been isolated from the reaction of N^P^CNHCHgCH^NlOCl^ (6) with four equivalents of _t-butylamine in benzene. This compound appears to be the first geminal cyclotriphosphazene derivative containing three chlorine atoms and three alkylamino- substituents. The ethanolamino-analogue of the above tris derivative (9) could not be prepared. The and ^ P NMR spectroscopic data for (ethylenediamino)- and (ethanolamino)cyclotriphosphazene derivatives indicate that they possess spirocyclic structures (replacement of chlorine atoms from the same phosphorus atom). The spirocyclic structures of the compounds, N3P,(XCH2CH2HH)C14, (X = HH or 0), are further substantiated by the % l.«R spectroscopic data for their dimethylamino-derivatives, M,P,(MS|),(X0H20H2HH) (7 and 15). The results obtained in the present investigation show that an ansa-type structure for the ethylenediamino-derivative, N 5P3(KHCH2CH2HH)C14 (6) proposed by BecKe-Goehring and Bopped is untenable. The formation of considerable amount of resinous material(s) is a prominent feature of the reactions of ethylenediamine and ethanolamine »ith 1 ^ 0 1 , ana its derivatives, particularly when the stoichiometry of the amine * phosphazene substrate exceeds 2.1 (for ethylenediamine) or 31 (for ethanolamine) The insolubility of the resinous material(s) in common organic solvents precluded the detailed study necessary to characterise the resin(s) fully. However, infrared evidence suggests that these resin(s) probably contains cross linked cyclotriphosphazene units. A tentative mechanism has been proposed for the observed chlorine atom replacement and the formation of resins. The aminolysis reactions of bis(primary amino)hexa- chlorocyclotetraphosphazenes, N^P^NHRjjClg, leading to the formation of bicyclic derivatives have been studied in detail and the results obtained constitute the most significant highlight of the present work.. The reaction of 2 -trans-6 -»4P4 (NHEt)2 Cl6 (22) with dimethylamine in various solvents have been investigated. The yield of the bicyclic compound, H4P4 (NMe2 )5 (HEt)(NHEt) (2 5 ) relative to that of the fully substituted cyclo- N P (HMe„)c(NHEt) 9 tetraphosphazene derivative, ^ 4 ^ 2 '6 V ' 2 (2-4) increases in the order, O H , ® « 0H2 C12 < CHCl3. There is no evidence for the formation of bicyclic compound in diethyl ether and carbon tetrachloride, only the fully substituted cyclotetraphosphazene derivative (24) is obtained in high yields (ca. 80$). The use of stoichiometric quantities of dimethyl- amine and, an excess of triethylamine in the reaction of 2 - trans - 6 - N / F , ( N H E t ) 2 g ^ results in an enhanced yield of the bicyclic compound (25)* The reaction of bis(ethylamino) -derivative (22) with an excess of triethylamine in chloroform does not yield a bicyclic compound. These results throw some light on the mechanism of the formation of bicyclic compounds. The reactions of 2,6-bis(primary amino)hexachloro- cyclotetraphosphazenes, N^P^NHR) = M6’ 'Pr ’ Pri , Bun , Bu11, Ph and CH2Ph) with an excess of dimethylamine have been studied. The formation of a bicyclic compound could not be detected when R = Bu*, Pr1 or Ph. For the remaining substituents, the yield of the bicyclic compound, N4P4(HMe2)5(HR)(NHR), relative to that of the fully aminolysed cyclotetraphosphazene derivative, N4P4(NHR)2(WMe2)g, increases in the order, Bu11 < Be < 0E,,Ph < Et $ Pr11. In some cases, pure products could not be isolated and product analyses have been carried out by 31P NMR spectroscopy. The 31P resonances of bicyclic phosphazene derivatives occur in the range 24 to 156 as against 10 to 56 for fully aminolysed cyclotetraphosphazenes. Thus 31P NMR spectroscopy has proved to be a powerful analytical and structural tool in the study of aminolysis reactions leading to the formation of bicyclic phosphazenes. Bicyclic phosphazenes can also be identified by their smaller TLC Rf values (silica gel, ethyl acetate eluent) compared to those of the fully aminolysed cyclotetraphosphazene derivatives. In separate experiments, fully aminolysed cyclotetraphosphazene derivatives, N4P4(NMe2)g(NHR) 2, (R = Me, Et, Pr11, Pri , Bun , Bu "fc and Ph) have been prepared by the reactions of bis(primary amino)hexachlorocyclotetraphosphazenes, N 4P4(NHR)2C16 , with an excess of dimethylamine in methyl cyanide. An interesting feature of the reactions in methyl cyanide is the formation of hydrochloride adducts of fully aminolysed cyclotetraphosphazenes in good yields. The fully substituted cyclotetraphosphazene derivatives mentioned above provide spectroscopic and TLO data for comparison with the corresponding bicyclic derivatives. The spectroscopic data for the fully substituted cyclotetraphosphazene derivatives are also helpful in deducing the structures of the bis(primary amino)hexachloro-derivatives from which they are derived. The reaction of mono(ethylamino)heptachlorocyclotetraphosphazene, ^ ( N H E t ) 0 1 ? , with dimethylamine even in the presence of an excess of triethylamine yields only the fully aminolysed cyclotetraphosphazene derivative, N 4P4(NMe2)7 (NHEt) (5 1 .) and there is no evidence for the formation of a bicyclic compound. On the other hand, the mixed amino-derivative, H^tSHStJtSHBu^Olg (g2) reacts with an excess of dimethylamine in chloroform to give a bicyclic compound, H ^ H M e . , ) 5(mst)0 ™ “*) (J4) in which the ethyl group (and not the t-butyl group) is attached to the bridgehead nitrogen atom. Tetrameric chloride (18_) reacts with an excess of methylamine in chloroform to give three products: (a) the octakis(methylamino) cyclotetraphosphazene, N^P^NHMeJg (55), (b) the bicyclic compound, N4P4(NHMe)g(NMe) ( % ) and (c) the hydrochloride adduct of the bicyclic compound, N 4P4(NHMe)g(Me) .HOI (57). Besides these three crystalline compounds, a small quantity (ca. 20$) of a non-crystalline resinous material is also obtained. When the above reaction is carried out in diethyl ether, only the octakis(methylamino)- derivative (55) is obtained. The ring P=N stretching, "0 (P=ST), frequency for the bicyclic phosphazene and the fully aminolysed cyclotetraphosphazene derivatives occur at ca. 1200 and 1250 to 1270 cm-1 respectively thus permitting a distinction between the two. The bands observed at 800-845 cm-1 in the infrared spectra of bicyclic compounds have been assigned to the vibrations of the -P-N-P- bridging unit. The v (P=N) for the hydrochloride adducts of fully aminolysed cyclotetraphosphazenes show an upward shift of 30-40 cm compared to the free bases and suggest protonation at a ring nitrogen atom. The infrared spectrum of the hydrochloride adduct of the bicyclic derivative, N4P4(NHMe)g(NMe) .HOI (57) shows no appreciable shift of v (P=N) compared to its value in the free base (56); on the other hand, the vibrations attributable to the -P-N-P- bridging unit shift to higher frequencies by 20-30 cm-1. These observations indicate that in the hydrochloride adduct (57), the protonation occurs at the bridgehead nitrogen atom. The H and P NMR spectra of the bicyclic phosphazene derivatives besides confirming their structures have many interesting features. The protons of the groups attached to the junction phosphorus atoms are deshielded compared to the remaining protons and this deshielding is observed even for the y-protons. The 31P NMR spectra of bicyclic phosphazenes provide examples of AgBg, A^BC and AgBX type patterns. It has been established that the reactions of N ^ C l g with primary amines as well as the reactions of bis(primary amino)hexachlorocyclotetraphosphazenes, N ^ N H R ) 2Clg, with dimethylamine involve three competing processes: (a) normal stepwise replacement of chlorine atoms to yield fully aminolysed cyclotetraphosphazene derivatives, (b) intramolecular nucleophilic reaction leading to the formation of bicyclic derivatives, and (c) intermolecular condensation process resulting in the formation of resinous materials. The relative yields of the three types of products depend on the reaction medium, the nature of the substituent present on the phosphazene substrate and the reacting nucleophile. The presence of two primary amino-substituents on the phosphazene substrate seems to be essential for the formation of bicyclic derivatives and a proton abstraction mechanism has been proposed. In all, more than fifty phosphazene compounds (most of them for the first time) have been prepared and characterised. Structures for many of the derivatives have been proposed on the basis of 1 H and 31P NMR data.