Ostwald Ripening: Curvature-Driven Coarsening

• Ostwald ripening is a curvature-driven coarsening mechanism where differences in interfacial chemical potential induce mass transfer from smaller to larger dispersed structures.
• Classical LSW theory models this process with a t^(1/3) growth law, although experimental deviations arise due to finite volume fractions and complex geometries.
• Experimental studies across bubbles, droplets, and solid islands confirm key signatures like particle shrinkage and coarsening, while non-equilibrium effects can reverse or arrest the process.

Ostwald ripening is a curvature-driven coarsening process in dispersed systems in which smaller droplets, bubbles, particles, or islands shrink and larger ones grow through diffusive mass transfer mediated by the surrounding phase. Across the literature represented here, the common mechanism is the Gibbs–Thomson or Young–Laplace elevation of chemical potential at high curvature: small objects are less stable, their interfacial equilibrium concentration is higher, and material therefore migrates toward larger, lower-curvature objects. Classical theory is usually framed by Lifshitz–Slyozov–Wagner (LSW) mean-field kinetics, but recent work extends the concept to buoyant bubble columns, porous media, electrochemical deposition, active matter, and non-equilibrium condensates, while also identifying regimes of inhibition, reversal, arrest, and geometry-controlled deviations from dilute mean-field behavior (Watanabe et al., 2014, Inoue et al., 2023, Uematsu, 28 Mar 2025, Keister et al., 31 Jul 2025, Burlakov et al., 2014).

1. Classical mechanism and thermodynamic basis

In the standard picture, Ostwald ripening is driven by the curvature dependence of interfacial thermodynamics. For bubbles, the Young–Laplace relation raises internal pressure as radius decreases; one formulation given for a spherical bubble is

$$p = p_{\rm bubble} - p_{\rm bulk} = \frac{2}{r}\gamma,$$

so smaller bubbles have larger internal pressure and higher equilibrium dissolved-gas concentration at the interface (Cima et al., 2018). A closely related formulation for gas bubbles in a liquid is

$$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$

with $r<R$, implying $P_r>P_R$ and hence gas diffusion from the smaller to the larger bubble (Cima et al., 2017).

For droplets and particles, the same principle is often expressed as a Gibbs–Thomson or Kelvin relation. In one classical form,

$$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$

so curvature increases equilibrium solubility and destabilizes small particles (Kabalnov, 26 Apr 2026). In active or aerosol-like droplet formulations, the equilibrium vapor fraction is shifted by curvature according to

$\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$

with growth then controlled by quasi-static diffusion and supersaturation (Wilkinson, 23 Mar 2025).

Within LSW theory, the growth law is written in diffusion-limited form as

$R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$

with the corresponding bubble or particle count decreasing as

$N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$

This yields the familiar asymptotic law $R\sim t^{1/3}$ in diffusion-limited coarsening (Cima et al., 2017, Cima et al., 2018). Molecular-dynamics simulations of bubble nuclei directly confirmed the self-similarity of the bubble-size distribution predicted by LSW theory and showed a crossover from interface-limited growth with $R\sim t^{1/2}$ to diffusion-limited growth with $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$0 as temperature increases (Watanabe et al., 2014).

A useful interpretive summary is that Ostwald ripening is not defined by any one material class, but by a transport architecture: curvature sets local chemical potential, diffusion transmits the imbalance, and mass redistributes from high-curvature to low-curvature domains.

2. Lifshitz–Slyozov–Wagner theory and its generalizations

The classical LSW framework treats a dilute population of spherical objects coupled through a mean field. In one formulation, the size distribution $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$1 obeys the continuity equation

$$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$2

and the diffusion-controlled kinetic law can be written as

$$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$3

with bubbles larger than the mean radius growing and smaller ones shrinking (Inoue et al., 2023). In the late stage, the cube of the mean radius and the mean volume are predicted to evolve linearly in time: $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$4 Experiments on aqueous microbubble solutions found both $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$5 and $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$6 approximately linear in time during the first $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$7 minutes, in agreement with the LSW scaling form, though with larger coefficients than the dilute theory predicts (Inoue et al., 2023).

The LSW picture also provides explicit self-similar distributions. In the aqueous microbubble study, the scaled distribution

$$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$8

was nearly time-independent, although broader than the ideal LSW distribution because of finite-volume-fraction and confinement effects (Inoue et al., 2023). Large-scale molecular dynamics further showed that cumulative distribution functions collapse when plotted against the scaled variable $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$9, providing direct numerical evidence of LSW self-similarity in bubble coarsening (Watanabe et al., 2014).

Several papers extend this structure rather than abandoning it. In buoyancy-driven microbubbles, the dependent variable is redefined as the density distribution of the buoyancy-induced flux,

$r<R$0

which maps the steady population balance onto an LSW problem with height $r<R$1 replacing time and an effective exponent $r<R$2 (Uematsu, 28 Mar 2025). In that setting the mean radius obeys

$r<R$3

so the fifth power of the mean radius grows linearly with height rather than the third power growing linearly with time (Uematsu, 28 Mar 2025).

A different generalization considers externally supplied material. When total aggregate volume grows linearly in time,

$r<R$4

classical Ostwald ripening is recovered only in the zero-growth limit $r<R$5, whereas for $r<R$6 the system enters a qualitatively different constant-number-density regime in which ripening ceases and the size distribution focuses rather than approaching a universal LSW profile (Vollmer et al., 2014). This suggests that LSW theory is best regarded as a limiting structure within a larger class of curvature-driven transport problems.

3. Experimental realizations in bubbles, droplets, and solid islands

Direct experiments across multiple systems reproduce the core signature of Ostwald ripening: the number of dispersed objects decreases while the mean size increases.

In glycerol–water mixtures designed to mimic some rheological aspects of blood, experiments on air bubbles measured the time evolution of mean radius $r<R$7, number $r<R$8, and the radius distributions $r<R$9 and $P_r>P_R$0. The observations were monotonic: the mean bubble radius increased while the number of bubbles decreased, and the distribution shifted toward larger radii as smaller bubbles disappeared (Cima et al., 2017, Cima et al., 2018). The 2018 study further reported that the initial normalized radius distribution follows a Tsallis ($P_r>P_R$1-Weibull) form with fit parameters approximately $P_r>P_R$2, $P_r>P_R$3, and $P_r>P_R$4 at $P_r>P_R$5 (Cima et al., 2018).

In aqueous microbubble solutions confined in a thin glass capillary, image analysis showed that the average growth and shrinkage speed of individual bubbles follows diffusion-limited Ostwald ripening quantitatively. Using literature values $P_r>P_R$6, $P_r>P_R$7, and $P_r>P_R$8, the kinetic coefficient

$P_r>P_R$9

was computed with no fitting parameters and found to agree quantitatively with the measured kinetics (Inoue et al., 2023). The same study noted anomalies for very small bubbles below about $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$0m, including slowdown, temporary pinning, and one apparent long-lived state at $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$1m, suggesting additional interfacial physics beyond the ideal diffusion-controlled law (Inoue et al., 2023).

Solid-state and supported-system realizations show the same morphology at a different scale. In ultrathin FeO islands on Ru(0001), annealing caused the number of islands to decrease, the average island size to increase, total coverage to remain constant within about $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$2, total edge length to decrease by about $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$3 after the first anneal, and defects to anneal out, with the smallest islands disappearing before larger ones expanded laterally (Michalak et al., 2021). The interpretation is explicitly Ostwald ripening rather than coalescence: small islands decompose, material diffuses across the surface, and larger islands grow while adopting more equilibrium-like shapes (Michalak et al., 2021).

The following comparison summarizes several experimentally observed signatures.

System	Observed signature	Reference
Air bubbles in glycerol–water	$$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$4 decreases, $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$5 increases, distributions shift to larger radii	(Cima et al., 2017, Cima et al., 2018)
Aqueous microbubble solutions	$$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$6 follows diffusion-limited law; $$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$7 linear in time	(Inoue et al., 2023)
FeO islands on Ru(0001)	small islands disappear, larger islands grow, edge length and defect density decrease	(Michalak et al., 2021)

A plausible implication is that the experimentally robust observables of Ostwald ripening are ensemble-level rather than object-specific: count loss, mean-size increase, and scaled-distribution evolution recur even when the microscopic carriers are bubbles, droplets, or oxide islands.

4. Deviations from dilute mean-field behavior

Although LSW theory provides the canonical asymptotic reference, multiple studies emphasize that real systems frequently violate its ideal assumptions.

In the glycerol–water bubble experiments, the classical LSW forms for mean radius and bubble number did not fit the data well. The authors instead used empirical relations

$$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$8

and

$$C(r) = C^\infty \exp\!\left(\frac{2\sigma V_m}{rRT}\right),$$9

with reported parameters $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$0, $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$1, and $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$2, which described the observed evolution better than classical LSW theory (Cima et al., 2018).

In aqueous microbubble solutions, the scaled distributions were broader than the ideal LSW form and closer to finite-volume-fraction theories. The authors attributed this to finite effective volume fraction and the quasi-2D wall-attached geometry, introducing

$\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$3

Observed coarsening rates exceeded dilute-theory predictions by factors of a few, with normalized coarsening speed $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$4 about 2.8 for $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$5 mm and 3.4 for $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$6 mm (Inoue et al., 2023).

Spatial interaction effects also violate simple mean-field intuition. Finite-element simulations of 3, 5, and 50 bubbles showed that distance between bubbles strongly influences the kinetics, and that even bubbles larger than the nominal critical radius can still decrease if they feed other bubbles in a multi-bubble system (Cima et al., 2017). This directly qualifies the simplest interpretation of the critical-radius criterion. A related statement appears in porous media, where confinement makes curvature a non-monotonic function of bubble volume, so equilibration does not reduce to the growth of a single largest bubble (Bueno et al., 19 Dec 2025, Laku et al., 8 Apr 2026).

In buoyant bubble columns, the departure is more structural: the universal distribution is approached as a function of height rather than time, and the transformed problem corresponds to LSW with $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$7, not the classical $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$8 or $\Phi(a)=\Phi_{\rm e}+\frac{\Lambda}{a},$9 cases (Uematsu, 28 Mar 2025). In continuously fed aggregate systems, classical $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$0 ripening is described as dynamically unstable, while sustained external growth drives the system into a focusing regime where the number density becomes constant and the standard deviation of the aggregate radius decays monotonically (Vollmer et al., 2014).

A recurrent misconception is that all curvature-driven coarsening must asymptotically display the textbook $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$1 law and universal LSW distribution. The collected work does not support that generalization. Instead, it shows that the LSW structure is exact only under a restricted set of dilute, mean-field, and conservation assumptions; once finite volume fraction, geometry, buoyancy, external feeding, or network connectivity intervene, the asymptotics can change substantially.

5. Porous media, subsurface systems, and confined ganglia

In porous media, Ostwald ripening is mediated by dissolved-species transport through the wetting phase between disconnected ganglia or trapped gas clusters. The process drives local capillary pressure toward uniformity and changes the long-time interpretation of trapping (Adebimpe et al., 2024). In this context, classical percolation-with-trapping becomes, after sufficient diffusive equilibration, percolation without trapping (Adebimpe et al., 2024).

A key practical consequence is that conventional short-duration or fully immiscible measurements can overestimate capillary trapping. A pore-network study reported trapped saturation reductions of about $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$2–$R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$3 when Ostwald ripening is accounted for, with a Bentheimer example decreasing from $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$4 to $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$5, a reduction of $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$6 (Adebimpe et al., 2024). The same work proposed the first-approximation correction

$R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$7

when no better data are available (Adebimpe et al., 2024).

Recent porous-media research adds two further layers: realistic geometry and multicomponent composition. In underground gas storage, ultra-high-resolution microfluidic experiments identified a two-stage process: a fast local equilibration over the first $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$8 hours during which the standard deviation of curvature drops to nearly zero while residual gas saturation $R_{\rm LSW}(t)=\left[R^3(0)+Kt\right]^{1/3},$9 and average curvature remain nearly constant, followed by a slower global ripening stage over $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$0–$N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$1 days in which gas diffuses toward large boundary bubbles acting as low-chemical-potential sinks (Salehpour et al., 24 Aug 2025). The corresponding continuum model links dissolved-gas transport to a pore-morphology-derived $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$2-$N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$3 relation and predicts saturation evolution without fitting parameters (Salehpour et al., 24 Aug 2025).

An image-based pore-network model extends this further by resolving multi-pore ganglia and discrete capillary events such as invasion, snap-off, retraction, fragmentation, coalescence, and dislocation within a unified framework coupling two-phase flow, solute transport, and ripening (Laku et al., 8 Apr 2026). This work emphasizes that in confined geometries the curvature–saturation relation is non-monotonic and discontinuous, so ripening kinetics cannot be captured solely by smooth continuum closure at the ganglion level (Laku et al., 8 Apr 2026).

Multicomponent kinetic theory introduces a number-density function in the 3D state space

$N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$4

where $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$5 is pore size, $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$6 bubble saturation, and $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$7 a compositional mole fraction. Evolution then follows a population balance equation

$N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$8

closed by mean-field approximations that include pore-size correlations and mass conservation (Bueno et al., 19 Dec 2025). This theory is reported to agree well with pore-network simulations across homogeneous, heterogeneous, correlated, and uncorrelated networks without adjustable parameters (Bueno et al., 19 Dec 2025).

These developments suggest that, in confined media, Ostwald ripening is best viewed as a redistribution process in a geometry-defined state space rather than merely a scalar bubble-radius problem.

6. Non-equilibrium, reversal, arrest, and acceleration

Several studies identify regimes in which the usual coarsening direction is altered or its kinetics are qualitatively reorganized.

Reverse Ostwald ripening arises when adsorption of a second species lowers the effective surface energy enough that smaller particles become more stable than larger ones. In a binary $N_{\rm LSW}(t)=N(0)\,\frac{R^3(0)}{R^3(0)+Kt}.$9–$R\sim t^{1/3}$0 solution, the effective surface energy of an $R\sim t^{1/3}$1-particle is written as

$R\sim t^{1/3}$2

with

$R\sim t^{1/3}$3

When this renormalized surface energy becomes negative, mass transfer proceeds from larger particles to smaller ones and the system approaches a monodisperse array (Burlakov et al., 2014). The sufficient condition given for reverse ripening is

$R\sim t^{1/3}$4

This is not merely slower coarsening, but the opposite thermodynamic trend (Burlakov et al., 2014).

At the opposite end, Ostwald ripening can be arrested in driven systems. Molecular-dynamics simulations of droplets under stochastic state switching $R\sim t^{1/3}$5 or random momentum kicks reported that Ostwald ripening is absent only away from equilibrium, with many small droplets persisting in non-equilibrium steady states (Keister et al., 31 Jul 2025). The paper interprets this as a possible mechanism for stabilizing liquid droplets in living cells (Keister et al., 31 Jul 2025).

Chemical activity can also accelerate, rather than suppress, coarsening. In reaction-diffusion models where sticky and nonsticky forms interconvert while total protein mass remains conserved, the asymptotic exponent remains the classical one,

$R\sim t^{1/3}$6

but the prefactor can be enhanced. In the large-droplet limit, an effective diffusivity

$R\sim t^{1/3}$7

increases the coarsening rate, and the acceleration factor can become arbitrarily large when reactions occur only outside droplets and are negligible inside them (Sorkin et al., 4 Jun 2025). This is an important qualification: activity changes the prefactor, not the long-time exponent, under the paper’s mass-conserving assumptions (Sorkin et al., 4 Jun 2025).

Another control mechanism involves additives of a sparingly soluble component. In two-component droplets, the less soluble additive becomes enriched in shrinking particles, and its Raoult effect can counterbalance the Laplace driving force. In the high-$R\sim t^{1/3}$8 regime, the ripening rate follows the cubic law with an additive-controlled prefactor,

$R\sim t^{1/3}$9

and the size distribution remains essentially LSW-like (Kabalnov, 26 Apr 2026). In the low-$R\sim t^{1/2}$0 regime, however, the distribution becomes bimodal: a fines fraction enriched in the additive coexists with a large-particle fraction that ripens almost classically (Kabalnov, 26 Apr 2026).

These cases collectively show that Ostwald ripening is not a single irreversible fate. Depending on surface adsorption, non-equilibrium driving, reaction localization, or multicomponent solubility, the process can be reversed, arrested, or strongly accelerated.

7. Applications and broader significance

The applications represented here span cloud microphysics, diving medicine, active matter, catalysis, batteries, and subsurface energy systems.

In cloud and precipitation physics, Ostwald ripening has been invoked in the context of rain initiation and in modified models with continuous droplet injection and removal at a maximum radius. The latter exhibits a transition from steady state to a limit cycle, with onset controlled by roots of a Laplace transform of a response kernel and an oscillation period scaling approximately as

$R\sim t^{1/2}$1

The relevance to atmospheric precipitation is discussed explicitly, particularly as a non-collisional growth mechanism that could generate periodic precipitation events (Wilkinson, 23 Mar 2025). For the rain-initiation paper itself, only the metadata supplied here are available, so detailed derivations or formulas cannot be verified from the provided text (Wilkinson, 2011).

In diving medicine, Ostwald ripening is treated as a possible contributor to decompression illness because, even at constant ambient pressure, it shifts a bubble population toward fewer but larger bubbles. Experiments in blood-like fluids and simulations motivated incorporation of this effect into the Reduced Gradient Bubble Model (RGBM), with one study noting that for a closed circuit rebreather dive to $R\sim t^{1/2}$2 fsw using $R\sim t^{1/2}$3 Heliox, including broadening can lengthen total decompression time by about $R\sim t^{1/2}$4 under certain broadening-time assumptions (Cima et al., 2018).

In active matter, passive beads in bacterial baths undergo “Ostwald-like” coarsening. The characteristic cluster size follows

$R\sim t^{1/2}$5

with $R\sim t^{1/2}$6 and $R\sim t^{1/2}$7, while the onset time scales as

$R\sim t^{1/2}$8

This is presented as a nonequilibrium analog of Ostwald ripening driven by bacteria-induced attraction and enhanced diffusion (Bouvard et al., 2023).

In electrochemistry, lithium nucleus evolution under constant-current deposition is described as competition between electroplating and electrochemical Ostwald ripening. The growth law

$R\sim t^{1/2}$9

predicts SEI-limited and electrolyte-limited asymptotic regimes, with mean radius scaling as $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$00 in the former and recovering classical 3D ripening in the latter (Zhang et al., 7 May 2025). The same framework relates morphology to Coulombic inefficiency through

$$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$01

linking surface-energy-driven redistribution directly to battery performance (Zhang et al., 7 May 2025).

In subsurface gas storage, Ostwald ripening changes both security and recoverability. Recent work argues that equilibration times for $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$02 can be $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$03 to $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$04 days with median $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$05 days, while $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$06 equilibration times can be $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$07 to $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$08 days with median $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$09 days for representative aquifer conditions (Salehpour et al., 24 Aug 2025). This suggests that ripening can act much faster than convective dissolution in $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$10 sequestration and on timescales comparable to seasonal $$P_{r} = P_{\rm amb} + \frac{2}{r}, \qquad P_{R} = P_{\rm amb} + \frac{2}{R},$$11 storage operations (Salehpour et al., 24 Aug 2025).

Overall, the body of work represented here supports a broad definition: Ostwald ripening is a curvature-driven redistribution mechanism whose canonical mean-field expression remains foundational, but whose practical manifestations are strongly conditioned by geometry, transport topology, compositional coupling, and non-equilibrium forcing.