Aerosols from speaking can linger in the air for up to nine hours

Abstract

Airborne transmission of respiratory diseases has been under intense spotlight in the context of coronavirus disease 2019 (COVID-19) where continued resurgence is linked to the relaxation of social interaction measures. To understand the role of speech aerosols in the spread of COVID-19 globally, the lifetime and size distribution of the aerosols are studied through a combination of light scattering observation and aerosol sampling. It was found that aerosols from speaking suspended in stagnant air for up to 9 h with a half-life of 87.2 min. The half-life of the aerosols declined with the increase in air change per hour from 28 to 40 min (1 h^-1), 10-14 min (4 h^-1), to 4-6 min (9 h^-1). The speech aerosols in the size range of about 0.3-2 μm (after dehydration) witnessed the longest lifetime compared to larger aerosols (2-10 μm). These results suggest that speech aerosols have the potential to transmit respiratory viruses across long duration (hours), and long-distance (over social distance) through the airborne route. These findings are important for researchers and engineers to simulate the airborne dispersion of viruses in indoor environments and to design new ventilation systems in the future.

Keywords: Airborne transmission; COVID-19; Lifetime; Respiratory diseases; Speech aerosol.

Graphical abstract

Fig. 1

Schematic of the experimental setup. (A) 3D schematics of a 0.5 m³ cubic stainless-steel chamber and air supply system. (B) The top view of the chamber system. A vertical laser sheet (height: 16.8 cm, thickness: 0.5 cm) is introduced through transparent windows across the chamber, normal to the speaking direction. A camera is located at the center of the right panel, with a hole for the speaker on the left panel.

Fig. 2

Decay and lifetime of aerosols from speaking in stagnant air. (D) The decay of aerosols from speaking in terms of particle count per frame by light scattering observation (100-s moving average for the original data) across “background” (14 min), “speaking” (~10 min), and “decay” (for 9 h in stagnant air) phases with five repeats. In the speaking phase, the phrase “stay healthy” was repeatedly spoken in a loud voice (maximum 83 dB at the distance of 40 cm; average 77 dB) with 2–4 s of pause in between the phrases. Accumulated images of 500 successive frames (20 s) for one of the tests at (A) 5 min, (B) 30 min, (C) 120 min, (E) 360 min, and (F) 565 min. A movie is available online at https://doi.org/10.5281/zenodo.4703075.

Fig. 3

Decay and half-life of aerosols from speaking and coughing under different ventilations. Panel (A) The decay of aerosols for both the speaking and coughing in terms of total particle number concentration (0.3–10 μm by optical particle sizer, OPS, 100-s moving average for the original data) in the “decay” phase under three different ventilation conditions (air change per hour, ACH: 1, 4, and 9 h⁻¹) with three repeats. (B) The half-life of the aerosols based on the exponential decay rates of particle concentrations in the chamber. “O” indicates that the half-life is obtained by OPS, “P” measured by ultrafine particle counter (P-Trak, 20–1000 nm), and “L” by the light scattering observation (LSO, D₅₀ = 2.45 μm, D₅₀ means that 50 % of particles are visible, which is also regarded as minimum visible particle size, see Figs. A2-A3 of Appendix A). Plots show the means and standard errors across three replicates. Note that decay of aerosols demonstrated in panel A is by OPS measurement since its sensitivity is better than that of LSO, and corresponding results measured by the LSO and P-Trak are shown in Figs. A4-A5 of the Appendix.

Fig. 4

Time-resolved size distribution of aerosols from speaking. Time-resolved size distribution of aerosols across “background” (14 min), “speaking” (~10 min), and “decay” measured by (A) scanning mobility particle sizer for particles from 10 to 420 nm, and (B) optical particle sizer (OPS) for particles from 0.3 to 10 μm in the chamber (air change per hour: 4 h⁻¹). Size distributions under different ventilations are shown in Fig. A6 of Appendix. (C) Average size-resolved number of aerosols in 12 size bins of optical particle sizer from the 10-min speaking (average across the 9 tests). (D) Estimated size-resolved number of SARS-CoV-2 virus-laden aerosols from 10-min speaking based on average particle size distribution by OPS (Fig. A7A of Appendix) and average virus load of oral fluid of COVID-19 patients (7 × 10⁶ copies per milliliter), assuming the size of aerosols measured shrinks to 20 % of its original size due to dehydration [15].