The formation of positronium-acceptor complexes (Ps-Ac) with nitro aromatics and quinone has attracted much attention. The dependence of such reactions on temperature and on the nature of the solvent is sometimes discussed within the ambit of the bubble model of Ps in liquids. However, the description generally adopted is that of a sharp bubble boundary which is not only unrealistic as a liquid interface but has also been shown recently, through a series of papers by the present authors, to be in contradiction with experimental observations. An explanation is sought for the observed occurrence of both Arrhenius and anti-Arrhenius temperature dependence in the reaction rate for Ps-Ac complex formation seen over different temperature regimes in terms of the Kramers' turnover in the context of the improved version of the bubble model.