Why a diffusion-only picture can miss the flow inside the drop
Diffusion-limited evaporation treats the slow step as vapour slowly spreading away from the droplet surface into still air. If that is the bottleneck, the evaporation rate and the resulting internal flow follow from the droplet's shape and the vapour concentration field alone. The trouble is that this picture quietly assumes nothing else is moving the vapour or the liquid. When buoyancy-driven flows develop, the internal motion is no longer set by diffusion, so predictions of the flow pattern can go wrong even when the overall evaporation is still roughly diffusive [1].
The same caution applies to the surrounding gas. If the air around the droplet is being forced to move, convection carries vapour away and reshapes the concentration field near the surface. Under those conditions the evaporation is no longer purely diffusion-limited, and the internal flow that a diffusion-only model predicts can diverge from what actually happens [1]. The breakdown is therefore less about the droplet itself and more about whether the environment is quiet enough for diffusion to be the true bottleneck.
Confinement can switch which step is the bottleneck
A second route to breakdown is geometric. When a droplet sits in a narrow gap between two plates, vapour has very little room to spread, so the removal of vapour by diffusion can stop being the slow step. Modelling of this setup shows a transition between a diffusion-limited regime and a phase-transition-limited regime, where the limiting factor becomes the evaporation at the liquid surface rather than transport through the gas [2]. In the phase-transition-limited case, the assumptions behind a diffusion-only description no longer hold, and predictions of how the droplet evolves need a different treatment.
This matters for internal flow because the flow is driven by where evaporation happens most strongly. If the controlling physics shifts from vapour transport to surface phase change, the evaporation profile across the droplet changes, and so does the circulation it drives. The crossover is modelled for shallow droplets under strong confinement [2], so it should be read as a confined-geometry result rather than a general rule for droplets in open air.
Different evaporation models give different answers near the contact line
Even without buoyancy or confinement, the choice of evaporation model changes predictions in ways that matter for flow. Comparing a one-sided model with the diffusion-limited model for thin droplets shows that the two can differ greatly, and that the differences trace back to how each model treats the region near the contact line, where the droplet meets the substrate [3]. In the one-sided model the contact line stays warmer than the bulk of the droplet, and the apparent contact angle and interface temperature profiles behave qualitatively differently from the diffusion-limited case [3].
Because internal flow is sensitive to temperature and surface-tension gradients along the interface, these near-contact-line differences are exactly where a diffusion-limited prediction is most likely to drift from reality. The comparison also shows that in some parameter regimes the total evaporation rate becomes proportional to the droplet surface area under the one-sided model [3], a different scaling from the diffusion-limited expectation. So part of the breakdown is not a physical failure of diffusion but a modelling choice that becomes inappropriate in thin-droplet regimes.
Where the diffusion picture still earns its keep
None of this makes diffusion-limited evaporation useless. Studies of mixture droplets describe the internal flow patterns during evaporation within a diffusion-limited framework, indicating that the assumption can still organise our understanding of the flow in some settings [4]. An empirical model likewise builds on diffusion-limited evaporation while adding convective transport at the liquid–gas interface, treating convection as a correction rather than a replacement [5]. That framing is useful: it suggests the diffusion picture often captures the leading behaviour, with convection and other effects acting as additions that grow important in specific conditions.
Putting the pieces together, the practical answer is to check the environment and the geometry before trusting a diffusion-only flow prediction. Still air, an open substrate, and a droplet thick enough that the contact-line region is not dominant favour the diffusion-limited picture. Forced convection, buoyancy, strong confinement, or very thin droplets push the system toward regimes where the internal flow needs a model that includes those effects [1][2][3]. The evaporation rate itself may still look diffusive in some of these cases, which is precisely why the internal flow can be mispredicted without anyone noticing from the droplet lifetime alone.
Sources used in this answer
Evaporation of Sessile Droplets: Effects of Substrate Thermal Conductivity and Droplet Shape Under Natural and Forced Convection
Diffusion-limited theory has limits for predicting internal flow in evaporating sessile droplets, with buoyancy-driven flows and forced convection among the factors that cause the assumption to break down.
Sessile drop evaporation in a gap – crossover between diffusion-limited and phase transition-limited regime
Under strong confinement in a narrow gap, shallow sessile droplets show a transition between diffusion-limited and phase-transition-limited evaporation regimes.
Comparison of one-sided and diffusion-limited evaporation models for thin liquid droplets
One-sided and diffusion-limited evaporation models for thin droplets can predict very different behaviour, with differences traced to how each model treats the near-contact-line region, including contact-line temperature and apparent contact angle.
Experimental investigation of evaporation dynamics and evaporation-driven flow in a sessile mixture droplet
Internal flow patterns during evaporation of a sessile mixture droplet were examined within a diffusion-limited framework.
Contribution of convective transport to evaporation of sessile droplets: empirical model
An empirical model treats convective transport at the liquid–gas interface as a contribution alongside diffusion-limited evaporation in sessile droplets.