From the early summer of 2020 to late 2021, I have worked with colleagues at the University of Bristol to try and better understanding how face coverings/masks filter out potentially infectious droplets. Masks are air filters we wear on our face, idea is shown in schematic below by Ioatzin Rios de Anda:
Masks contain tiny fibres, and the idea is that the air flows (blue curves and arrows) around these fibres (allowing you to breathe in and out) while any nasty particles (magenta) in the air collide with the fibres, and are filtered out. Or rather proper masks, i.e., FFP2/N95/KF94 etc, and surgical masks are made of tiny fibres. Cotton and other woven masks are made of tiny fibres twisted into yarns that are then woven to form the fabric, as you can see in this image Jake Wilkins (with Blender) made from imaging data from Ioatzin Rios de Anda:
The thinner strands above are cotton fibres but you can also see one yarn clearly: this the almost vertical bundle of fibres in the centre third of the image. As the fibres are bundled into yarns, almost all the air goes through holes between the yarns and so is not filtered. You can see two such holes above, one in the top right and one in the bottom left.
As a result, cotton and other woven masks are pretty terrible at filtering out particles. Our simulation results (in Physics of Fluids and on arXiv) for the fabric above predict they filter out between 2.5 and 10% of particles (1.5 micrometres in diameter). So I suggest you wear a proper FFP2/N95/KF94 mask, they are lot better at filtering, in particular the material they are made of will typically filter much more than 95% of the droplets in your breath, not maybe 10%.
You can see my blog for latest thoughts on COVID-19 transmission and masks.
Work on modelling masks filtering virus
There are fair number of details involved in how masks filter virus, for example, we breathe out droplets over a huge range of sizes, and it is bit hard to know which sizes are the ones to worry about. I also contributed to a paper in Aerosol Science and Technology (arxiv preprint) that discusses these details, most of the work was done by Josh Robinson.
The paper also discusses the two Fs of a good mask: Filtration and Fit. Making a mask out of a material that filters well (as FFP2 etc are, and as some surgical masks are) is only half the battle. The mask also needs to fit well (as FFP2s etc do but surgical masks don’t). Otherwise, a lot of the air goes past the mask and is not filtered.