Abstract
Purpose :
Models for TBU include many effects: surface tension of the tear/air interface; tear viscosity; evaporation of water to the environment; variation of surface tension from lipids and proteins at the aqueous/lipid interface (the Marangoni effect); osmotic flow from the epithelium; fluorescent intensity (to name just some). We hypothesize two main modes of TBU: Case I due primarily to evaporation and Case II due primarily to the lipid-driven Marangoni effect. The physical effects must combine differently in each case. We aim to clarify when the different effects are important in TBU.
Methods :
Tear films of subjects were simultaneously recorded using either: (1) fluorescein (FL) and retroillumination (RI) methods or (2) FL and lipid microscopy (LM) for lipid layer thickness (Braun et al, 2015, PRER, 45, 132). Using these images for close comparison, math models were solved for the tear film thickness, insoluble surfactant concentration (representing the polar part of the lipid layer), as well as osmolarity and fluorescein concentrations inside the tear film. Fluorescein concentration was converted to fluorescent intensity as described by Nichols et al (2012, IOVS, 53, 5426).
Results :
In Case I, experimental FL+LM images show that the lipid layer is practically stationary, and TBU develops over a period of seconds or more through what are effectively holes in the lipid layer. Theory confirms that this TBU mode has localized evaporation competing with surface tension driven flow and diffusion of solutes; the details of the balance depend on TBU spot size. Osmolarity is steadily elevated in this case. In Case II, FL+RI or FL+LM experiments show sub-second spreading of the lipid layer and tear film and subsequent TBU. The spots tend to widen much more quickly and to a larger extent than for Case I. Theory confirms that Marangoni flow is dominant in the thinning here but it may cooperate with evaporation sometimes. Osmolarity does not immediately rise from the Marangoni flow as in Case I. Case II TBU can be driven by patches of lipid that are either mobile (bubbles bursting) or less mobile (globs). FL intensity results from the math models is consistent with, and helps interpret, the experiments.
Conclusions :
The experimental data from subjects and mathematical models point to two modes of TBU, with Case I dominated by evaporation and Case II dominated by Marangoni effects. Subcases within the latter are possible.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.