The masks were cut into pieces (~ 70 mm × 70 mm) and tested in a circular acrylic air duct with an inner diameter of 50 mm. Indoor ambient aerosols or KCl aerosols were loaded as the pollutant source. The counts of 0.3-10 μm particles were measured by an optical particle counter, and were record every 1 min for 2 times upstream the mask sample and then 2 times downstream. The single-pass filtration efficiency η(dp) of particles with a certain size of dp (μm) was calculated by: equation image where Cup and Cdown are the average particle number count (pcs) at upstream and downstream of the filter, respectively.
The pressure drop across the respirator filter was measured by a differential gauge. The air temperature, relative humidity, and filtration velocity were measured by an airflow/temperature meter at the air duct exhaust.


1. Sample size: 5 cm diameter disk (active area = 19.6 cm2)
2. Ambient aerosol: simply use indoor air in the lab, T = 25 ± °C (not controlled), relative humidity RH = 30 ± 20% (not controlled), 0.3 μm partic
le number concentration < 104 pcs/L, typical aerosol size range: 0.3-1 μm. 3. KCl aerosol: using 10 wt% KCl solution, pumping at 1 L/min. Temperature T = 25 ±°C (not controlled), relative humidity RH = 30 ± 20% (not controlled), 0.3 μm particle number concentration < 3×104 pcs/L, typical aerosol size range: 0.3-5 μm.
4. Filter pre-conditioning: none
5. Flow rate: face velocity > 0.1 m/s (preferably > 0.3 m/s, optional).
6. Number of tests: 3 total (optional).
7. Filtration efficiency reported as a function of time (optional), filtration pressure reported as a function of time (optional), filtration efficiency reported as a function of particle size (optional)

How accurate? Comparison with NIOSH approved Lab

Though this is not a sufficient test to establish conformity with NIOSH, EN149 or GB2626 standards, we have shown that results using this set-up conforms well to NIOSH pre-approval results (see in the Supplementary materials).

We have collected data on a set of N95 masks (n = 10) exposed to various sterilization procedures using the equipment described here and a test to NIOSH standards at a commercial laboratory (ICS Laboratories, USA) equipped to perform pre-certification testing. Mask instantaneous filtration efficiency values measured using the two tests had a correlation coefficient of 0.89 and all masks that passed our testing also passed ICS tests. Nonetheless, the data described here results should be interpreted as a relative, not absolute, measures of filtration efficiency.

ICS Instantaneous Test Our setup for 0.3μm particles
Model Cycles Filter Efficiency(%) Model Cycles Filter Efficiency(%)
3M 1860 1 99.33 3M 1860 1 99.2
2 99.47 2 99.2
5 99.36 5 98.76
KC/Halyard 46767(duckbill) 1 99.89 KC/Halyard 46767(duckbill) 1 99.83
2 99.78 2 99.91
5 99.89 5 99.89
Gerson 21 1 98.69 Gerson 21 1 96.06
2 97.43 2 96.46
3M 8210 1 99.76 3M 8210 1 99.86
2 99.82 2 99.52

All of the information above can be found in this paper.