Minimum size for a nanoscale temperature discriminator based on a thermochemical system

What are the limits of size reduction for information processing devices based on chemical reactions? In this paper, we partially answer this question. We show that a thermochemical system can be used to design a discriminator of the parameters associated with oscillations of the ambient temperature. Depending on the amplitude and frequency of the oscillations, the system exhibits sharp transitions between different types of its time evolutions. This phenomenon can be used to discriminate between different parameter values describing the oscillating environment. We investigate the reliability of the thermochemical discriminator as a function of the number of molecules involved in the reactions. A stochastic model of chemical reactions and heat exchange with the neighborhood, in which the number of molecules explicitly appears, is introduced. For the selected values of the parameters, thermochemical discriminators operating with less than 10(5) molecules appear to be unreliable.