Data-informed Fabric Masks and Disposable Masks

After a year and a half of sewing masks and struggling to decide which fabrics are likely to provide enough filtration... I have data. After reading dozens of papers... I know what to make of it. After seeing kids spill out of schools... I've learned the difference between theory and practice.

This is my place to share my insights with you.

Each blue box is a link

Masking in our community:

Infographic: The spectrum of masks

Masks perform really quite differently. This summarizes the differences.

Infographic: Mask fit

Mask fit matters a lot. Air that goes around a mask is not filtered by the mask. This inforgraphic shows data from a study of inward filtration and offers guidance on how to assess mask fit.

Infographic: Masks for source control may not protect

One intermittently-masked individual infected half a classroom with Covid. All the others were masked. This illustrates how masks intended for source control may not offer personal protection.

Infographic: How does source control work?

How can my mask protect you more than me? An infographic boiling down evidence that we exhale wet particles large enough to be well-filtered by many fabrics (good source control), but once they dry and shrink and waft to another person they are too small for most fabrics to filter (poor personal protection).

PoweCom size reduction

I find the PoweCom KN95s to be the most comfortable mask for me... but a smidge big. This is my trick to reduce the size.

Infographic: KN95 counterfeits vs fabric

Counterfeit KN95s with sub-standard filtration are a real concern. This infographic shows measurements from a study in which even most counterfeit KN95s outperformed two-layer cotton filtration.

Infographic: Recorders in MVWSD

The present plan for playing recorders includes indoor and outdoor play options, depending on weather. The indoor playing plan provides no source control, and poorly fitting fabric masks which (I expect) provide very little personal protection, a scenario overall equivalent to an unmasked activity.


On sewing fabric masks:

  • Woven cotton is ok but not great. Cotton masks help reduce transmission and so are effective for public health. But if you want to protect yourself, incorporate a better fabric.
  • Non-woven polypropylene and non-woven polyester and flannel are better than cotton. For the same level of breathability, they provide better filtration. Two layers of a medium-weight interfacing will hugely improve your mask's filtration with little impact on breathability, but the stiffer construction may require adjusting your sewing pattern.
  • Washing is fine. Ten machine wash/dry cycles with a laundry bag don't affect filtration.
  • Particle size is a complex topic without which we can't properly interpret fabric filtration data nor public health recommendations.

MakerMask post: Recommendation for a fabric combination better than cotton

In this MakerMask post, I discuss the data and science behind the recommendation for two layers of ~65 gsm (gram/m2) spunbond nonwoven polypropylene (NWPP) plus one layer of ~130 gsm woven cotton. Filtration was tested against 0.5μm-12μm particles at Colorado State University.

Post: Summary of CTT data for sewists

An explanation of measurements of filtration of submicron NaCl and of Bacterial Filtration Efficiency, and what fabrics I had tested, stated as plainly and simply as I can.

Post: Summary of CTT data for scientists

Deeper details and mathematical analysis.

Sewing my mask pattern

Sewists have asked how to sew the mask featured in the MakerMask post.

On particle sizes

Particles >10μm are filtered well by most fabrics, resulting in clear reduction of community transmission with mask use regardless of type of mask. However, 1μm to 10μm particles also can be infectious, and so excellent protection for the mask wearer necessitates adequate filtration in this range as well. Particle sizes used for filtration testing vary across standards and studies.

Microscope photos

Microscope photos of various fabrics used for masks.

Layering Single layer data is not enough, since the first layer in a stack-up seems to filter with higher efficacy than subsequent layers.

Last updated 11 October 2021
© Anna Mitros
Back to Ania's Home Page