Open Data —— the Smart Air Fit Tests

As always, I’m making the raw data and more details about the testing methods openly available here.

Extra details:

A second way to parse the data

The results for all N95 and N90 masks in the main article are using the fit machine’s “N95 correction mode.” Explaining this gets into a bit of nerdery, but it’s one of the neatest facts about filters I’ve ever learned.

Let’s start with a mystery. Sharp readers might have noticed something weird about the test results in the main article. For example, take a look at the 9010, which is an N95 mask. That means it should get 95% of particles. But in the test above it got 99%. How is that possible?

Fit Test

That gets to the heart of what fit tests are. They are designed to measure how much air is leaking in the sides of the mask. They are NOT interested in how many particles penetrate the mask itself. So these are truly leakage tests.

The nerd in me cannot stop from explaining: How do they know which particles leaked in and which penetrated the mask? Here’s the trick. People often think, “the smaller the particle, the harder it is to capture.” But when we’re talking about really small particles, that’s actually the opposite of the truth!

See, masks and other filters are actually really really good at capturing the smallest of particles. (The reason is Brownian Motion, which I explain here.) That means that even with an N95, filtration percentages for particles .01 microns is close to 100%.

Fractional Filter Efficiency

Thus, when the machine is testing N95 masks, it only counts .04 micron particles. For particles this small, almost all of them would be captured by the mask. So any leftovers almost certainly leaked in the sides.

But with the N99 masks, the penetration rate is so small, the machine doesn’t bother with restricting the particle range. Instead, it counts its full range (.01 to 1 microns). So the N99 results are the most intuitive: the number of particles inside and outside the mask.

But if we want that intuitive number for all the different masks, we can add back in the percentage of particles that theoretically should be getting through the N95 and N90 masks. For example, the N95 9010 above scored 99% on the fit test. Because it’s an N95, we can assume that 5% of the entire particle range should make it through the mask, so we can subtract 5% from the fit results. That gives a new result of 94%. If we do that, here’s what actual effectiveness looks like:

% Particles Blocked (Corrected)

(I only did this calculation for the 3M masks because most other masks don’t provide specific particle ratings.)