Q. Is the BE mask NIOSH approved? Have these masks undergone ASTM or OSHA fit testing?
A. Not yet. BE masks have undergone clinical fit tests in local hospitals and comparative pre-NIOSH filtration tests with N95 masks. We have submitted our PPE design to NIOSH for review. Our results are early and validation is a long process involving many different standards. Some of the standards include approval by the FDA, CDC, NIH, OSHA, ASTM, and NIOSH. We have shared our mask on the NIH 3D Print Exchange and are awaiting review from the VHA. For more details, please skip to our results below.
|The BE mask is offered to the general public for general, non-medical purposes. In accord with FDA Emergency Use Authorization, it has been submitted for review to NIOSH. The safety and efficacy of any mask depends on fitting, sanitization practices, and filter materials. Any mask should be tested and fit to meet local needs and standards. Please see our user liability disclaimer for details.|
Team Testing Process March/2021
This document describes the equipment, method, and process used to perform filtration
efficiency testing and differential pressure testing of respirators (masks) and filter material alone.
Filtration efficiency testing uses a particle counter to measure particle counts of a number of
sizes upstream of the filter and downstream with a fixed flow generated by a vacuum pump.
Higher efficiency means more particles are filtered by the mask (filter). Differential pressure
testing measures pressure drop across the filter and indicates breathability (higher differential
pressure means more difficult to breathe through for extended times).
Mask Tests & Comparisons
The BE mask is a design derivative of the Montana Mask, which has undergone its own series of formalized tests in relation to the standard N95 mask.
|Design||NIOSH OSHA 1910.134Fit Test|
Via QNFT TSI Portacount
|Clinical Variations Comparison Fit Test Report||Respiratory Test vs. N95||Sanitization Tests of bacterial growth (PLA)||Anecdotal Endorsement|
The BE Mask includes iterative improvements to materials, assembly process, seal, sizing in relation to geometric scaling, and filter cartridge design. Details can be found within the design documentation. Test details in table below.
Testing Results Summary
|Saccharine Fit Testing||Hausdorff Analysis||Filtration Testing||Sanitization Test||Material Print Test: TPU||ASTM |
Testing Details Table
|Anecdotal Fit Testing||Industry expertsMedical expertsPublic||Demonstration of seal, sizing, and fit comfort by industry and medical experts. |
|Saccharine Fit Test||Buffalo General Employee Health Services Protocol|
The test was conducted by UB Clinical Informatics, Surgery and ICU fellows.
|An aerosolized solution of saccharine (sodium saccharin) is used during a set of tests to evaluate the filtering of the respiratory mask. Detecting a sweet flavor would indicate inadequate filter or face-mask leak. The tests follow the OSHA Protocol and are standard for all healthcare employees caring for patients.|
One minute each:Thirty seconds of facing forward with normal and deep breathing. Bending over for 30 seconds. Moving head side to side, up-down, counting from 0 to 20 or reading a short paragraph provided.
|Sanitization Degradation Test||UB Biomedical Engineering & Orthopaedics||Test of materials related to degradation and durability using sanitization solutions recommended by CDC and WHO).|
|Filtration Efficiency Test||Buffalo Manufacturing Works||Filter media tested for filtration efficiency relative to particle size and differential pressure. Six particle sizes recorded: 10μm to .3μm. Results averaged across 3 consecutive tests. Results compared with standard N95 material and ASTM standards.|
|Porosity Bio testing using viruses||UB Immunopathology||Testing mask surface material for ability of viral hosting and transmission.|
Report In Process
|Hydrophobic Water Resistance||UB Biomedical Engineering & Orthopaedics||Liquid saturation testing degradation with water|
Report In Process
|Hausdorff Distances Analysis||UB Computational Anatomy||Comparison of geometry and contour design of all masks in NIH Exchange using STL files|
|TPU Material Print Test||3D Universe||Testing Porosity of TPU Material|
Test Model 4/15
|F2299/F2299M-03(2017)||ASTM||Standard Test Method for Determining the Initial Efficiency of Materials Used in Medical Face Masks to Penetration by Particulates Using Latex Spheres|
Filter Material Tested by Supplier
|F2101-19||ASTM||Standard Test Method for Evaluating the Bacterial Filtration Efficiency (BFE) of Medical Face Mask Materials, Using a Biological Aerosol of Staphylococcus aureus|
TPU Report in Process
PLA Report from Billings
Details at www.makethemasks.com
|F2100-19||ASTM||Standard Specification for Performance of Materials Used in Medical Face Masks|
Report In Process
|F1862/F1862M-17||ASTM UB Biomedical Engineering & Orthopaedics||Standard Test Method for Resistance of Medical Face Masks to Penetration by Synthetic Blood (Horizontal Projection of Fixed Volume at a Known Velocity)|
|F1494-14||ASTM UB Engineering||Standard Terminology Relating to Protective Clothing|
Biocompatibility Assessment 4/15
Hausdorff Distances Analysis
Summary and Methods
STL files of each mask model were downloaded from the NIH 3DPX repository (https://3dprint.nih.gov/collections/covid-19-response). The models were imported into Geomagic Wrap 2019 software. Three alignments were performed to bring the masks into the same 3D space and for distance calculations:
- A coarse alignment was performed using the n-point registration function, and four landmarks were assigned to each mask (mask edge at the mid-sagittal location of the bridge of the nose, mask edge at the midsagittal location of the chin, and the maximal points of the mask curvature on either side of the cheek region). The four landmarks were used to roughly align the masks.
- A medium alignment using the best-fit function, minimizing linear distances between the reference mask (BECMv1 size large) and all other masks, using 1,000 random sampling vertices across the reference mask.
- A fine alignment phase, where 5,000 random sampling vertices were used to bring each pair of masks into fine alignment.
Once alignment is done, all meshes were exported as STL files and imported into the R programming environment. The Rvcg and rgl packages were used in a custom script to calculate pairwise mean Hausdorff distance criterion (https://people.eecs.berkeley.edu/~malik/cs294/Huttenlocher93.pdf) between all possible pairs of masks, using the Metro algorithm (P. Cignoni, C. Rocchini and R. Scopigno. Metro: measuring error on simplified surfaces. ComputerGraphics Forum, Blackwell Publishers, vol. 17(2), June 1998, pp 167-174). The vertices used for comparison were defined as all the vertices present on the first mesh being compared (each pairwise comparison was performed twice, once with each mesh being the reference shape for vertex sampling). The resulting distance matrix was used in a cluster analysis (using complete linkage and Euclidean distance) to obtain the dendrogram included below, within the Minitab software. The diagonal of the distance matrix was kept in to verify that the Hausdorff distance calculation was performing correctly.
Table 1: Mask Design Comparison Matrix
Distance matrix showing pairwise (bi-directional) Hausdorff distance values among the 32 NIH 3DPX mask designs, plus the five BECMv1 mask designs. Distance values rounded to nearest whole number for readability. Hotter colors indicate higher Hausdorff distance (=larger difference) between a given pair of mask designs.
Editable/analyzable numerical table here: https://docs.google.com/spreadsheets/d/1sgL43Jm-UmI3xB6KKw2np32huxluS1F78vHzzcJmC7E/edit?usp=sharing
Table 2: Cluster Analysis
Clustering analysis dendrogram showing distances among NIH 3DPX mask designs (see above table for model #) based on bi-directional Hausdorff distances. This analysis suggests there are three major “families” of mask designs. BECMv1 (1 to 5) is in a different “family” as the mask approved by NIH for use in a clinical setting (#11, 12).
The BECM mask can be produced in two versions – one using PLA and one using TPU. The PLA mask includes a gasket seal that touches the face. The TPU mask touches the face directly. Both masks incorporate a common filter material and common strapping. Each material is described in terms of specs and biocompatibility below.
Polylactic acid thermoplastic. Based on review of MSDS from Ultimaker for PLA, mild skin irritation is possible: “May cause eye / skin irritation. Product dust may be irritating to eyes, skin, and respiratory system. Caused mild to moderate conjunctival irritation in eye irritation studies using rabbits. Caused very mild redness in dermal irritation studies using rabbits (slightly irritating).”
Thermoplastic polyurethane TPU is a hygroscopic material. Based on a review of MSDS from Ultimaker for TPU, there is no available test data related to skin irritation. Based on our use by the development team, wearing the masks for short durations of 1-2 hours has not resulted in adverse effects.
The filter material is made from polypropylene plastic and does not contain hazardous materials like fiberglass. The filter rating is MERV-15, filter efficiency is listed as 95-99%, and removes particles down to 0.3 microns.
The supplier for our material: https://www.mcmaster.com/22905k63
Strapping material is typically an elastic material similar to the material used for surgical masks and N-95 masks.
Gasket material is made from silicone foam. The manufacturer – Reiss Manufacturing, Inc. – notes that “the product does not contain any toxic or hazardous substances as defined in OSHA Standard for General Industry (29 CFR 1910)…therefore does not require the Safety Data Sheet…” So, there is no available data regarding skin irritation. This was part of a note passed along on request from our industrial supplier, McMaster-Carr. Based on our use by the development team, wearing the masks for short durations of 1-2 hours has not resulted in adverse effects.
TPU Print Test
Created April 15, 2020
This test establishes print quality and porosity for TPU materials. The following shape needs to pass inspection before printer is approved to create mask.
The STL file can be found here
The “fins” that stick up at the top are to create an extra challenge to test stringing, as the print head needs to travel back and forth between those parts when printing.
Results will depend on the specific printer and TPU material being used. Even with more-or-less optimized settings, there is likely to be some stringing, which can be easily cleaned up with a sharp blade and a heat gun. Then you should have something like this:
Then we do the water test… Fill the cup with water and let it sit for 30 minutes.
Then move the part aside carefully and check for any water on the table surface. There should not be any water visible.
Sanitization Degradation Test
Test performed by Mark Ehrensberger on 4/14/2020
Evaluate the possible degradation of the TPU used for the BECMv1 mask as a function of exposure to basic sanitation methods.
Materials and Reagents
- 3% Hydrogen Peroxide
- 70% ethanol solution
- 70% ethyl alcohol gel
- Dilute bleach (10mL Clorox added to 90mL tap water)
- Ultimaker TPU 95A (thermoplastic polyurethane) filament
- Ultimaker 3 extended printer
- Test coupons (1 inch x 1 inch x 1/8inch depth) of TPU were printed.
One test coupon of TPU was individually immersed in 20mL of 3% hydrogen peroxide, 70% ethanol, dilute bleach, or bathed in 70% ethyl alcohol gel for 5mins. A control TPU sample was not exposed to any sanitation method. Coupons were removed, dried with a paper towel, and visually examined for signs of material degradation. Following coupon extraction, the solutions were also examined for signs of particulate debris. Finally, the flexural mechanical properties of the test coupons were qualitatively assessed by flexing coupons manually. Subsequently, the coupons were re-immersed in the original test solutions/beakers for a total of 3 hours and the same assessments were performed.
General test setup
5 Minute Immersion
TPU coupons floated on surface of all solutions and were periodically pushed below the surface during the tests. As shown in the figures below, during the 5 minutes of immersion there were no noticeable reactions occurring on the surface of any TPU coupons. Note that bubbles were already present in the bleach solution prior to adding the TPU coupons.
Test coupons during 5 min immersion
Following the coupon extraction, visual inspection of test solutions (see figure below) indicated that no residual debris was left in any of the solutions. Furthermore, there were no visual indications of deformation or degradation of the test coupons.
Photos of test solutions and coupons following 5 min immersion
Manual flexing of the coupons following the 5 minute immersion revealed no qualitative differences in the flexural properties.
3 Hour Immersion
The test coupons continued to float on the surface of the solutions. The coupon for the 70% ethyl alcohol gel test was placed flat on a bed of gel on the bottom of the beaker. No reactions were noted during the immersion test period. Following coupon extraction, the 3% hydrogen peroxide, 70% ethanol, and 70% ethyl alcohol gel did not have any residual TPU debris in the beaker. The dilute bleach solution had 1 small TPU particle floating in solution.
Photos of test solutions following 3 hour immersion
Visual inspection of the test coupons showed that the coupons exposed the 70% ethanol solution and 70% ethyl alcohol gel appeared to have a slight bowing (concavity towards top surface ). This perhaps indicates the TPU swelled on the surface that was exposed to the alcohol solution/gel. These coupons were also reported to have a slight reduction in the qualitative flexural rigidity as compared to the other test samples (including control samples). It was also noted the sample in the 70% ethanol made an audible squishing sound when flexing.
Photo of TPU coupon showing slight bowing following 3 hour immersion in 70% ethanol
It is envisioned that the TPU BECMv1 mask might be immersed or sprayed/wiped with some of these sanitation solutions to decontaminate the mask between uses. The 5 minute immersion is likely a better approximation of the actual exposure times that the TPU BECMv1 will be subjected to during a single decontamination cycle. The 3 hour immersion may represent an accelerated test of an accumulated exposure after several decontamination cycles. Therefore, it would appear that the 3% hydrogen peroxide or the dilute bleach solution might be preferred due to the lack of apparent swelling and no noticeable changes in flexural properties. More extensive testing is required to evaluate any quantitative changes in mechanical properties, swelling, and coupon degradation. In addition, the antimicrobial and antiviral efficacy of these decontamination solutions for these TPU surfaces should be confirmed.
Synthetic Blood Test
Material: Bag Air Filter MERV-15 from Grainger (22905K63)
Date: April 15, 2020
Description of Test
The goal was to replicate the F1862 Resistance of Medical Face Masks to Penetration by Synthetic Blood (Horizontal Projection of Fixed Volume at a Known Velocity).
Samples under test are MERV15 filter material, two layers (AKA two-ply in this document). This material is air handling filtration material with a MERV-15 rating. MERV-15 rating indicates particle filtration size (95% of particles larger than 0.3 micrometers blocked).
Due to unavailability of synthetic blood, 2% cow milk was used with ~.5ml red food dye added to 75ml of milk.
- Temperature was room temperature (68 F)
- Humidity was 30% (less than recommended, but we did not have control over humidity)
Test setup shown in figures below. Two layers of filter material secured vertically.
- Outside = side facing the liquid stream
- Inside = side away from liquid stream
Liquid dispensed from a 60ml squeeze bottle from a distance of 12 inches. Horizontal stream sent from bottle through a 3/16” hole toward the mask. Hole is in a cup 1” from the sample surface. Weight of the liquid dispenser documented before and after each test to show amount of liquid used. Not all liquid hit the target. Assuming 1gram weight of liquid is equivalent to 1 ml volume.
Figures 1-3 shows the test setup used.
Three samples were used. Samples were placed in the holder and the bottle was held 12 inches from the sample. The bottle was squeezed to dispense a strong, fine stream toward the sample, through the hole in the cup. The stream lasted at least five seconds, to allow for sufficient liquid to reach the sample. Not all the liquid reached the sample due to aiming error.
Top View of test setup.
Front (head on) view of test setup.
Close-up side view of test setup.
All samples passed the test. No moisture nor discoloration was detected on the side opposite the stream (aka, inside of the sample). Very little of the liquid hitting the samples stayed on the sample and soaked in; most was repelled. Table 1 summarizes the test configuration for three samples, and the results. Figure 4 shows front (exposed) side of the masks.
Test details and results.
Pass = no liquid, no moisture, no discoloration of two-ply material on side opposite liquid stream.
Three filter material samples shown post-test. Shown is the front side for all (the side hit with the stream). Reverse side showed no change, even minutes after the test.
An additional test was performed with sample 2. This sample was subsequently held 3 inches from the liquid dispenser and a strong stream was directed at it for 4 seconds. This resulted in no dampness nor discoloration on the opposite side immediately following this test. The sample was examined again two minutes post-test and there was still no discoloration nor dampness (held horizontal the entire time). After two minutes the top layer of the two layer material was removed and discoloration and dampness existed on the inside of the outer layer. Yet, this did not penetrate the second layer (second ply). As a result, the material is deemed to pass the test of exposure to a small volume of liquid similar to blood.
Finally, sample 3 was held horizontally and 3 g of liquid were poured onto the surface and remained there for 10 seconds. No discoloration nor dampness were found on the opposite side immediately after this test, nor minutes after.