Droplet Formation inside the Human Larynx: In Vivo Observations and Ex Vivo Models*

Phonation during human speech involves saliva-coated vocal folds opening and closing about 200 times per second.  This process is known to produce large quantities of micron-scale expiratory droplets that are believed to be responsible for transmitting infectious pathogens into the air, leading to millions of deaths, but little is known about the formation of these salivary droplets.  Here we present two sets of video evidence elucidating the nature of droplet formation during phonation.  (1) In collaboration with an otolaryngologist, we inserted a laryngoscope (i.e., camera fed in through the nose) to directly visualize vocal folds during phonation.  We observe that salivary filaments or ‘strings’ often stretch between the vocal folds and pinch off during each cycle, and occasionally exhibit the ‘beads-on-a-string’ instability that occurs with viscoelastic fluids. (2) We used a monster-truck subwoofer to construct artificial vocal folds that oscillate up to 200 Hz with multi-millimeter amplitudes.  High speed video of saliva in the apparatus exhibits behavior like that observed in vivo, affording an opportunity to systematically test droplet formation under controlled conditions. We interpret both sets of observed behavior in terms of a balance between capillary, elastic, and inertial forces, and we discuss the implications for understanding unresolved questions about human-to-human variability in rates of expiratory aerosol emission.