Instability in Ostwald ripening processes

There is a dimensionless parameter which enters into the equation for the evolution of supersaturation in Ostwald ripening processes. This parameter is typically a large number. Here it is argued that the consequent stiffness of the equation results in the evolution of the supersaturation being unstable. The instability is evident in numerical simulations of Ostwald ripening.

💡 Research Summary

This paper investigates a fundamental instability in the long-term dynamics of Ostwald ripening, a classic coarsening process that occurs after phase separation. The work challenges the completeness of the widely accepted Lifshitz-Slyozov (LS) theory, which predicts that the system evolves toward a unique, self-similar state characterized by a constant growth-rate parameter ν = 27/4.

The core of the argument revolves around a dimensionless parameter, α, which emerges from the equations governing the evolution of supersaturation. For physical systems like atmospheric aerosols, α is typically very large (≈420). The study demonstrates that this large α value renders the differential equation for the supersaturation evolution “stiff.” Stiff equations are notoriously sensitive to small perturbations, suggesting that the long-term behavior of the ripening process may be inherently unstable.

The instability mechanism is precisely quantified. The growth parameter ν (or its inverse ˜ν) is shown to be proportional to α multiplied by the deviation of the mean relative droplet size from unity, ⟨y-1⟩ (equation (22)). In the limit of large α, ⟨y⟩ must be extremely close to 1 to keep ν finite. However, ⟨y⟩ is an average over a finite number N of droplets. Due to the randomness in initial droplet sizes, ⟨y⟩ exhibits statistical fluctuations on the order of 1/√N. The key insight is that equation (22) amplifies these small fluctuations in ⟨y-1⟩ by the large factor α, leading to potentially large fluctuations in ν itself. The paper defines a dimensionless variable Ω = αx/√N to characterize when this effect becomes significant. Since the characteristic droplet size x increases and the number of surviving droplets N decreases over time, Ω inevitably becomes large for systems with α » 1, signaling a breakdown of the stable LS regime.

Numerical simulations in Section 3 confirm this theoretical prediction. The author performs direct simulations of a discrete N-droplet system with finite α and randomly sampled initial sizes, integrating the coupled equations for supersaturation and individual droplet growth. The results clearly show that ν(t) fluctuates erratically around the LS value, with the amplitude of fluctuations increasing for larger α and at later times when N is smaller—exactly as predicted by the growth of Ω. Interestingly, while the detailed dynamics of ν are unstable, the average growth of the droplet radii, governed by the harmonic mean of ν, still approximately follows the t^(1/3) scaling law predicted by LS theory. This instability, arising from the discrete nature of the droplet population and statistical fluctuations, appears to have been missed in prior numerical studies that treated the droplet size distribution as a continuous function.

In conclusion, the paper posits that for physical systems where α is large, the asymptotic long-time state of Ostwald ripening may not be the stable, self-similar solution of LS theory. Instead, it may be an unstable regime where statistical fluctuations, amplified by the system’s stiffness, prevent ν from settling to a constant value. This work provides a novel perspective that could explain subtle discrepancies between the classical theory and experimental observations, highlighting the importance of discrete noise and finite-number effects in the late stages of coarsening processes.