How well do aerosols transmit SARS-CoV-2?

The novel coronavirus spreads through respiratory droplets expelled by a person with COVID-19. Aerosols—the fraction of those droplets about five µm and smaller—tend to remain suspended in the air longer than their larger relatives and may thus contribute substantially to the community spread of the disease.

Now Scott Smith and Daniel Bonn of the University of Amsterdam and their colleagues have created a model system to evaluate the suspension time and survival of aerosols from coughing and speaking in confined spaces. Although a highly infected, symptomatic individual may still spread the disease via aerosols, the good news is that they are an inefficient transmission source for those with mild or no symptoms.

To assess the transmission of aerosols, the researchers sprayed a mixture of glycerol and ethanol into an experimental chamber. The ethanol evaporated a second later and left only the glycerol aerosols, with sizes comparable to those from a person speaking or coughing. The droplets passed across a laser sheet, which illuminated each one. To simulate the number of droplets in the air over time, the researchers used a model that estimated the rate of the aerosols’ evaporation and sedimentation.

Smith, Bonn, and their colleagues found that their results (blue line in the figure) agree with the experimental observations (blue circles) and are similar to results from a previous study’s data set and model (green circles and green line, respectively). Given the six or so minutes it takes for the number of aerosols to halve and the estimated number of virus particles it takes to infect an individual, the researchers conclude that a healthy person is unlikely to become infected by a single cough from someone with COVID-19.

The confined, poorly ventilated experimental chamber could be seen as a worst-case scenario: The authors suggest that the risk of transmission during short exposure times would be smaller in larger, better ventilated indoor spaces. (S. H. Smith et al., Phys. Fluids 32, 107108, 2020 .)