Various acidic compounds were tested as potential initiators of the cationic polymerization of L,L-dilactide. Regardless of solvent and temperature only trifluoromethanesulfonic acid and methyl triflate proved to be useful initiators. With methyl triflate as initiator L,L-dilactide polymerizes faster than diglycolide, whereas diglycolide is more reactive when anionic or complexing initiators are used. Initiation with methyl triflate yields poly(L-lactide) with methyl ester endgroups indicating that the chain growth proceeds via cleavage of the alkyl-oxygen bond. At polymerization temperatures < 100 "C the resulting poly(L-lactide) is 100% optically active indicating that the chain growth does not proceed via free carbenium ions, but via triflate ester endgroups. All attemps to identify cyclic dioxocarbenium ions as reaction intermediates failed, because propagation is much faster than initiation. However, in nitromethane oligomers with triflate ester endgroups were detected. For comparison the reaction of y-valerolactone with methyl triflate was studied by means of IR and ' H NMR spectroscopy. Methylation of the exocyclic oxygen yielding a cyclic dioxocarbenium ion was found to be the sole reaction.

The polymerization of 2,2-dimethyltrimethylene carbonate (1) was initiated with borontrifluoride etherate, trifluoromethanaulfonic acid (triflic acid), methyl triflate and ethyl fluorosulfate. The reaction was studied at 20, 50, 80 and 120°C in 1,2-dichloroethane and nitrobenzene. Methyl triflate and triflic acid turned out to be the most reactive initiators. Yet, due to "black biting" degradation, gelds > 80Vo were never obtained. With BF3 etherate yields up to 98% and molecular weights (4) up to 5 200 g/mol were found. From H NMR and IR spectra it can be deduced that methyl triflate alkylates the exocyclic oxygen (carbony1 group) and the resulting carbenium ion reacts with its counterion by ring-opening of the alkyl-oxygen bond. The oligomers and polymers formed in this way initially possess methyl (or ethyl) carbonate groups at one end and triflate groups at the other end of the chain. Elemental analyses and IR spectra prove that no carbon dioxide is lost in the course of the cationic polymerization. DSC measurements showed that crystallinity and melting points of poly(1) largely depend on the polymerization conditions in contrast to the glass transition temperature. 1 a) Part 11: Cf.13. b,

No abstract

Homopolymerisations of B-propiolactone and c-caprolactone, initiated by means of methyl trifluorosulfonate, triethyloxonium tetrafluoroborate or acetylium perchlorate, were investigated. Both ' H and 13C NMR spectra proved that the alkylating initiators yield polyesters with alkyl ester end groups, indicating a chain growth via alkyl-oxygen cleavage of the lactone. At temperatures below 100 "C cationic polymerizations initiated by alkylating reagents were found to proceed via end groups which may cause degradation due to back-biting. When ecaprolactone was reacted with excess methyl triflate, high concentration of triflate ester end groups were formed, whereas in the case of 8-propiolactone active end groups were not detectable by ' H NMR spectroscopy. Initiation with acetylium perchlorate yielded a polyester with acetate end groups. Acetate end groups were also obtained, when "living" polymers, initiated with methyl triflate, were reacted with acetic anhydride. It could be shown that the formation of acetate end groups does not indicate an electrophilic attack at the endocyclic oxygen. Furthermore, it is discussed that any experimental evidence for a cationic chain growth via acyloxygen bond cleavage is lacking.

No abstract