Intensive care units (ICUs) are the high incidence sites of hospital-acquired infections, where impinging jet ventilation (IJV) shows great potential. Thermal stratification of IJV and its effect on contaminants distribution were systematically investigated in this study. By changing the setting of heat source or the air change rates, the main driving force of supply airflow can be transformed between thermal buoyancy and inertial force, which can be quantitatively described by the dimensionless buoyant jet length scale ( $\overline{l_{\text{m}}}$). For the investigated air change rates, namely 2 ACH to 12 ACH, $\overline{l_{\text{m}}}$ varies between 0.20 and 2.80. The thermal buoyancy plays a dominant role in the movement of the horizontally exhaled airflow by the infector under low air change rate, where the temperature gradient is up to 2.45 ℃/m. The flow center remains close to the breathing zone of the susceptible ahead, resulting into the highest exposure risk (6.6‰ for 10-μm particles). With higher heat flux of four PC monitors (from 0 W to 125.85 W for each monitor), the temperature gradient in ICU rises from 0.22 ℃/m to 1.02 ℃/m; however, the average normalized concentration of gaseous contaminants in the occupied zone is reduced from 0.81 to 0.37, because their thermal plumes are also able to carry containments around them to the ceiling-level readily. As the air change rate was increased to 8 ACH ( $\overline{l_{\text{m}}}$= 1.56), high momentum weakened the thermal stratification by reducing the temperature gradient to 0.37 ℃/m and exhaled flow readily rose above the breathing zone; the intake fraction of susceptible patient located in front of the infector for 10-μm particles reduces to 0.8‰. This study proved the potential application of IJV in ICUs and provides theoretical guidance for its appropriate design.

Origin of differently sized respiratory droplets is fundamental for clarifying their viral loads and the sequential transmission mechanism of SARS-CoV-2 in indoor environments. Transient talking activities characterized by low (0.2 L/s), medium (0.9 L/s), and high (1.6 L/s) airflow rates of monosyllabic and successive syllabic vocalizations were investigated by computational fluid dynamics (CFD) simulations based on a real human airway model. SST k–ω model was chosen to predict the airflow field, and the discrete phase model (DPM) was used to calculate the trajectories of droplets within the respiratory tract. The results showed that flow field in the respiratory tract during speech is characterized by a significant laryngeal jet, and bronchi, larynx, and pharynx–larynx junction were main deposition sites for droplets released from the lower respiratory tract or around the vocal cords, and among which, over 90% of droplets over 5 μm released from vocal cords deposited at the larynx and pharynx–larynx junction. Generally, droplets’ deposition fraction increased with their size, and the maximum size of droplets that were able to escape into external environment decreased with the airflow rate. This threshold size for droplets released from the vocal folds was 10–20 μm, while that for droplets released from the bronchi was 5–20 μm under various airflow rates. Besides, successive syllables pronounced at low airflow rates promoted the escape of small droplets, but do not significantly affect the droplet threshold diameter. This study indicates that droplets larger than 20 μm may entirely originate from the oral cavity, where viral loads are lower; it provides a reference for evaluating the relative importance of large-droplet spray and airborne transmission route of COVID-19 and other respiratory infections.