Amid the scramble for protective equipment, factories in China have popped up to seemingly overnight to meet the demand. Yet tests of masks from 21 manufacturers shows half of respirators being marketed as N95 fail to meet that standard – with some no better than a cotton cloth.
Factories Spring Up to Meet Mask Demand
Amid the rush, the high prices masks are commanding have led to what The Wall Street Journal called a “wild-west” for mask manufacturing.
Dozens of manufacturers and suppliers have contacted Smart Air selling everything from surgical masks to CE-certified FFP2 and N95 masks. Some suppliers even claimed to produce hundreds of thousands of masks per day.
Testing How Well China’s New Masks Perform
To find out just how well these masks are performing, we started contacting suppliers and testing their masks from the middle of April. In total, we tested 31 masks from 21 suppliers. The suppliers claimed their masks were N95 or KN95, meaning they filter 95% of particles (similar to the FFP2 standard in Europe). What we found shocked us.
We tested masks’ filtration efficiency using a method that resembled the “particulate filter efficiency” test in the official US test standard. The standard tests normally use a TSI 8130 machine that costs $95,000.
Our engineering team simulated this setup using three Met One particle counters (which cost about $5,000 each). This test provides a reasonable approximation of the official test, although it cannot be used to formally certify masks for government standards.
We adjusted the airflow to 85L/min to match the NIOSH test standard and measured the percentage of particles 0.3 microns and above the masks were capturing (full test methodology and how it compares to NIOSH standards). We tested two samples of each mask and repeated the filtration efficiency tests 3 times for each sample.
Results Find Most Masks Not Up to N95 Standard
Out of the 31 masks tested, 13–nearly half–failed to meet the N95 standard.
In other words, buying N95 masks in this new supply chain has odds of flipping a coin that buyers will actually get an N95 mask.
To meet the N95 (US) or KN95 (China) requirement, masks need to capture 95% of particles. Yet only 18 masks managed to capture more than 95% of tiny particles.
What’s more, many of these failing masks were not just a few percentage points below the standard. On average, the masks that failed the tests captured just 51% of particles. The worst of the worst captured just 24% of particles, which is on par with the filtration ability of a bed sheet.
Masks Show Poor Quality in Straps, Stitching, Noseband
We ran simple tests of the structural quality of the masks. We started by pulling the straps of each mask. We only tested the straps on the masks that met the N95 standard.
Straps on some models broke almost immediately. Straps failed on 11 out of 18 masks. These masks are unlikely to last long, especially under conditions where healthcare workers are having to re-use masks.
Next, we tested the masks for fit. This revealed a common issue with the nose pieces and glue. Sub-standard glue meant that metal nose clips on some models came off quickly. In one model, the mask split open in the middle while the engineer was putting it on.
These simple structural tests indicate that, even for masks that meet the N95 filtration standard, they may still prove useless. Only 7 masks – less than one-quarter – passed both the N95 filtration tests and these “wear and tear” tests.
New N95 Masks Made in Mobile Phone and IT Factories
To try to understand why so many masks were failing our tests, we did a background check on each of the 21 companies using a company directory for China.
Out of the 21 factories, we found that only 9 of the factories that sent us mask samples worked in the “health and medical” sectors. The other 12 operate in sectors such as mobile phone accessories, sports products, and even software & IT services.
Oddly enough, there was no correlation between factory sector and mask performance. This could be because the companies actually selling masks are just middle men and not actually producing the masks.
Non-medical companies also sent us more masks on average. Non-medical companies sent us an average of two mask designs versus one for medical companies. This suggests the non-medical companies are merely re-sellers.
New Moves May Help Clean Up Supply Chain
Recent moves may be helping the situation. In an attempt to limit the number of uncertified masks on the market, on April 10th, China tightened the checks on masks leaving the country. It remains to be seen whether these checks will prove useful.
A note on why we’ve anonymized the factory and mask names: we tested masks from only a small handful of suppliers (21 in total). That’s a fraction of all the companies making masks right now. With this data, we didn’t set out to name and blame specific factories, but to provide data to help people keep themselves safe through this pandemic. For more information on the suppliers, and tips on how to source reliable masks, check out our tips for sourcing N95 masks article.
As part of Smart Air’s open-data policy, the test data from these experiments is being released open-source. It can be downloaded from our supplemental data article.
Paddy is the CEO of Smart Air, running operations from Beijing. He has a Masters in aeronautical engineering from Bristol University, UK having specialised in aerodynamics. An advocate for open data, free information and transparent business, he spends his spare time promoting honest business and social enterprise.