Thread by @antiviral_mktng

I donned a 3M Aura 9210+ respirator 20 times to measure loss of protection.

Repeated donnings appear to gradually erode protection, but the impact measured is small, and not detectable using candle smoke, only saline aerosol.

🧵

I had performed another experiment (yet to be published) which required multiple donnings, and noticed that fit-tests performed using candle smoke failed to detect a gradual loss in fit that was apparent when using saline.

So I went on a side quest to explore this phenomenon.

This was also a good time to explore the impact of redonnings on respirator fit. We know that extended wear causes a loss of fit. But how much of this loss can be attibuted to the *time* spent wearing the resperator, versus the process of putting it on?

jamanetwork.com/journals/jaman…

This experiment is pretty straightforward.

I took a new 3M Aura 9210+ and donned it 20 times, measuring the fit factor for a single normal breathing exercise each time. I used a PortaCount 8020A in N99 mode. I performed this experiment with saline, and again with candle smoke.

The respirator was purged for 60 seconds before each fit-test, along with passing a positive pressure seal check. I wiped sweat from my face before each donning to mitigate that variable. I carefully ensured that the respirator was donned properly each time using a mirror.

I have recently switched to using a single overhead Tygon tube suspended from a microphone stand to sample particles from within the respirator. This setup minimizes the weight of the tube, which can pull on the respirator and impact the measured fit factor.

Before each experiment I use an air purifier to flush the room as low as the steady state particle concentration will go, then turn off the purifier and generate aerosols.

For saline, I use a 0.2% w/v concentration of pickling salt in distilled water, dispersed using a Levoit Classic160 ultrasonic humidifier.

For smoke, I light and then extinguish ECLHTLNY paraffin wax tea candles, purchased from Amazon.

For both particle types, I generate aerosols until the concentration is high, then cease generation and use the purifier to reduce particles until the concentration reaches 3000 particles/cm³, as measured by the PortaCount, then turn off the purifier and begin the experiment.

I have settled on 3000 particles/cm³ for most experiments based on this US Army study finding a condensation particle counter (PortaCount technology) slightly undercounts particles when compared with an aerosol electrometer, at higher concentrations.

share.ansi.org/PUBHDSSC/All%2…

Unfortunately due to time constraints, I was not able to perform both halves of the experiment on the same day. So I performed the saline experiment on day 1, and the smoke experiment on day 2.

Here are the results!

You can see the saline fit factor begins around 1000, then seadily decreases toward 500 as I reach 20 donnings. For smoke, the fit factor is more constant the entire time at approximately 300.

I forgot to write down the fit factor for the 12th smoke donning, so I plotted the null value in all graphs shown here. 😅

I should emphasize how small this loss in protection is by the 20th donning. For the saline test, the total inward leakage (percentage of particles penetrating the respirator) began at 0.11%, and ended at 0.21%, for a very smol 0.1% absolute loss in protection. After 20 donnings!

In an attempt to show more clearly the loss in fit factor, here I am comparing the harmonic mean fit factor (average leakage) from the first 5 and last 5 donnings of the respirator, for each aerosol type.

So how do I interpret these results? I'll begin with the practical application, and then discuss the seemingly weird difference between saline and smoke.

For starters, the loss in protection from redonnings in Aura 9210+ is shockingly low, at least for my face. Although there does appear to be a slow loss, it's small enough that I wouldn't worry about it on a day-to-day basis.

It's difficult to say how much this result generalizes to other respirator types and other faces. I'm essentially using my head as a mannequin. Different heads have different circumferences, which could stretch the straps more. Different respirators have different strap material.

At the same time, I was honestly expecting a larger loss in fit factor after *so many* donnings. With the caveat that more replications on more faces and respirator models are needed, this experiment has shifted my opinion toward: donning count is less of an issue than I thought.

I would NOT use this data to determine your own personal limit of safety. If you want to do that, the most accessible way is to replicate this experiment using a Bitrex fit-test every N donnings. Pass one test, then do 10-20 donnings and see if you can still pass.

I have been trying to figure out a way for people to track the usable lifespan of their respirators. Right now I am experimenting with what I call the "dot system": I keep a Sharpie next to my respirator, and put a dot on the strap every time I wear it.

I keep one respirator for each use case. Right now I have one for picking up packages and groceries from my front door (5 minutes per wear), and another for walks around the neighborhood (30 minutes per wear).

The accumulated wear time is then easy to estimate. If you don't want to put dots on your respirator, you can just write them on some paper.

Bitrex fit-test every N accumulated wear time until it fails. Stay well within that limit for safe use. Re-test occasionally to verify.

I also think it's reasonable to use this experiment's result to calibrate intuition toward: "one more donning is probably fine". If you're hyper-aware about economizing respirator wear and tear, I don't know that donning count is a place worth focusing.

I have seen some people claim that respirators should be donned carefully to avoid wearing them out. More data is required for other face types and models, but so far there doesn't appear to be a compelling benefit in doing this.

One thing that probably *does* generalize better from these results: newer respirators are likely to be more protective. Going into a hospital? Fresh respirator. Going for a walk outside? Used respirator is fine.

Never wear a respirator that fails a positive pressure seal check unless you have no other choice. If it gets that far, you're way past its usable lifespan on your face.

sci-hub.se/https://www.ta…

It would be really helpful to have more replications so we can understand the magnitude of the redonning impact on performance. For my part, I'd like to test an Aura 9210+ (blue strap), because I expect the weak straps will lose protection more quickly. Hello PortaCounters! 👋

Okay so that's all great, but why would this loss of protection be detectable with saline, but not with smoke?

This probably happens because the PortaCount counts all particles that enter the respirator in the 0.02-1 μm size range. It does not care about size, it just counts.

Meanwhile, of all particle sizes, electrostatic respirator filter material allows the most penetration around the 0.05 micron size (the MPPS, or most penetrating particle size). It is much better at blocking smaller and larger particles.

ncbi.nlm.nih.gov/pmc/articles/P…

The size distribution of particles in air can vary based on the type of aerosol used. So different aerosol generation techniques can produce different amounts of MPPS particles.

Here is a histogram of NaCl particle distribution, in an experiment which utilized the PortaCount.

onlinelibrary.wiley.com/doi/full/10.10…

While I have not been able to find literature describing expected distributions of various aerosol types, and I have not measured the distribution of aerosols I am using, conceptually they can differ.

Which means you can end up with more smoke particles getting past the filter.

If more particles are getting through the filter, as a fraction of total particles inside the respirator (the rest going around the face seal), then face seal penetration is weighed less heavily when counting *all* of the particles that get inside.

With more particles coming through the filter, a small loss in literal fit (from repeated donnings) has less impact on the resulting fit factor.

If the fit is very poor, total particles coming around the face seal represent more of the particles inside the respirator, and in that scenario all aerosol types should produce similar scores.

So aerosol type scores should only differ for *small* losses in fit.

In addition, the electrostatic filter is vastly superior at capturing positively and negatively charged particles, including MPPS particles.

So if your two aerosol types have different charges, one of them can be superior at penetrating the filter.

In this study, TSI quantified the loss in filter penetration when particles are charged versus charge-neutralized. You can see that particles with a neutral charge are much better at penetrating the filter.

tsi.com/getmedia/f4177…

Ultimately whether the particles fail to penetrate the filter due to size or charge is irrelevant. The point is simply that some particles penetrate better than others, and thus some aerosols are more sensitive to changes in respirator fit! These produce higher PortaCount scores.

As a person with a PortaCount in a plague, I am primarily interested in testing respirator mods so that the community can feel safer in modding, so that modding and decorating will grow, and hopefully respirator wearing will become more popular!

Certain mods or wear patterns (such as redonnings) are only likely to impact the *fit* of the respirator, not the filter performance. To test those mods, I want the most sensitive aerosol for measuring fit.

So a second conclusion I draw from this experiment, which aligns with several informal experiments I have done, is: saline is more sensitive than candle smoke for measuring respirator fit.

I'm intending to do a more thorough comparison of different aerosol types in the future.

It's important to note that this DOES NOT mean smoke is unsuitable for testing a respirator on your face! It simply means that saline allows a more precise measurement. If you can pass any PortaCount or Bitrex fit-test on your face, that's a good standard for protection.

Your home science experiment for this week: buy some Bitrex and a cheap nebulizer, get some wear time on your daily driver, and do a fit-test after some accumulated wear to see if it still passes. Tweet your results!

https://x.com/antiviral_mktng/status/1743473281208910126

And of course, let me know if there are any mods you would like me to test! ❤️😷

cc: @DenHaagChris

You requested an experiment similar to this one a while ago! ✨

https://x.com/DenHaagChris/status/1760205719813378400

@DenHaagChris Thank you @ahc_dot_txt for asking an important clarifying question!

I neglected to mention: all of the donnings in this experiment occurred in rapid succession. I sat down and donned/doffed the respirator 20 times in a row, testing in between, without stopping.

@DenHaagChris @ahc_dot_txt Each cycle of don-check-purge-test-doff takes about 4 minutes on average. By the end of the experiment, the total accumulated wear time for the saline Aura was about 82 minutes, and for the smoke Aura it was 78 minutes.

@DenHaagChris @ahc_dot_txt I also neglected to mention: after each doffing, the respirator was folded flat into its original shape, and the nose wire was straightened. This prevents the nose foam from coming unglued during storage.