Article Text

Download PDFPDF

Disinfection of Pseudomonas aeruginosa from N95 respirators with ozone: a pilot study
  1. Edward P Manning1,2,
  2. Matthew D Stephens3,
  3. Sylvie Dufresne4,
  4. Bruce Silver5,
  5. Patricia Gerbarg6,
  6. Zach Gerbarg7,
  7. Charles S Dela Cruz1 and
  8. Lokesh Sharma1
  1. 1Section of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut, USA
  2. 2Department of Internal Medicine, VA Connecticut Healthcare System, West Haven, VA, USA
  3. 3Air Liquide Advanced Materials, Branchburg, New Jersey, USA
  4. 4IM3 Consulting Group, Quebec City, Quebec, Canada
  5. 5Silver Dental Associates, LLC, Burlington, New Jersey, USA
  6. 6Department of Behavioral Sciences, New York Medical College, Valhalla, New York, USA
  7. 7Gerbarg & Associates Inc, Scottsdale, Arizona, USA
  1. Correspondence to Dr Lokesh Sharma; lokeshkumar.sharma{at}; Dr Edward P Manning; edward.manning{at}


Introduction Personal protective equipment shortages require the reuse of N95 respirators. We sought the necessary conditions for ozone to disinfect N95 respirators for reuse and the effects of multiple cycles of exposure.

Methods Portions of 3M 1870 N95 respirators were exposed to ozone at 400 ppm with 80% humidity for 2 hours to determine effectiveness of ozone on killing Pseudomonas aeruginosa. Entire 3M 1870 N95 respirators were exposed to five cycles of 400 ppm with 80% or higher humidity for 2 hours then evaluated for ozone’s effects on airflow resistance, filtration efficiency, strap strength and quantitative fit.

Results Ozone exposure disinfected 3M 1870 N95 respirators heavily inoculated with P. aeruginosa. Ozone exposure did not negatively affect the airflow resistance, filtration efficiency, strap strength or fit of the 3M 1870 N95 respirator.

Discussion These results suggest that ozone is a feasible strategy to disinfect N95 respirators for reuse during this and future pandemics.

  • occupational lung disease
  • respiratory infection
  • viral infection

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Key messages

  • We aimed to find: (1) conditions necessary for ozone to kill targeted organisms on N95 respirators for disinfection and reuse; and (2) the effect of ozone in similar conditions on the function of N95 respirators after multiple cycles of disinfection.

  • Ozone is a feasible strategy to disinfect N95 respirators.

  • Ozone at 400 ppm with 80% humidity for 2 hours effectively kills Pseudomonas aeruginosa on the 3M 1870 N95 respirators. Five cycles of exposure at these conditions do not degrade the filtration efficiency of the 3M 1870 N95 respirators nor does it negatively affect the quantitative fit of the mask.


The COVID-19 crisis has created a shortage of personal protective equipment (PPE), most critically, National Institute for Occupational Safety and Health (NIOSH)-approved N95 filtering facepiece respirators, hereafter referred to as N95 respirators (or N95’s) for healthcare personnel (HCP). The Centers for Disease Control (CDC) recognised conditions under which the reuse of N95 respirators may be necessary.1 2

We investigated the feasibility of ozone disinfection of N95 respirators to ensure sufficient supplies for HCP during the situations like COVID-19 pandemic, which has created shortage of PPE. Ozone has been shown to inactivate viruses by acting on the protein structure of a virus capsid or on viral nucleic acids including COVID-19.3–6 Therefore, it seemed feasible to explore its potential benefit in inactivating COVID-19. For initial investigations, a bacterial surrogate microbe, Pseudomonas aeruginosa (PsA), was chosen. Ozone can be generated from air, quickly destroyed, naturally degrades to oxygen and is easily accessible through commercial means. We aimed to find: (1) conditions necessary for ozone to kill targeted organisms on N95 respirators for disinfection and reuse; and (2) the effect of ozone in similar conditions on the function of N95 respirators after multiple cycles of disinfection.


N95 respirators from 3M (type 1870) were tested. An ozone chamber provided by Ozone Solutions (Hull, Iowa, USA) was used for testing the efficacy of ozone for killing bacteria on N95 respirators. The ozone chamber consists of an airtight chamber, ozone generator using ambient air capable of concentrating to 500 parts per million (ppm), ozone destruction unit and ozone UV analyser.

Four pieces of the N95 respirator approximately 4 cm2 were dipped in a culture of PsA overnight. We chose PsA as a representative organism to inoculate our masks because it is a vegetative bacteria that the CDC identifies as more difficult to kill than viruses7 and ozone concentrations required to inactivate vegetative bacteria have been shown to be 4–13 times higher than for viruses.3 Half of the pieces were exposed to ozone: 400 ppm 80% humidity for 2 hours. The remaining half were kept in ambient air approximately 35% humidity for 2 hours. The pieces were dipped in 1 mL phosphate buffer solution and vortexed vigorously to dissociate the maximal amount of bacteria from the surface or interior of the respirator. Serial dilutions were plated on agar plates to enumerate surviving bacteria as colony-forming units per millilitre (CFU/mL). This process was performed twice resulting in four pieces in our ozone-exposed experimental group (n=4) and four pieces in our control group (n=4).

The effects of ozone on filtration efficiency were tested on five whole respirators after exposure to 450 ppm 75%–90% humidity for 2 hours for five cycles with 20 min ambient air breaks. To ensure that there was no residual ozone on the respirators themselves, we used a handheld monitor to measure the ozone level on or near the surface of the respirator. We observed no residual ozone level on or near the respirator. Ozone exposures were performed at Ozone Solutions. Airflow resistance and filter efficiency tests were performed by the CDC (Pittsburgh, Pennsylvania, USA) per protocol in NIOSH N95, 42 CFR Part 84 (Respiratory Protective Devices) (TSI) using 0.26 µm aerosolised sodium chloride under a flow rate of 85 L/min.8 9 An additional five whole respirators underwent quantitative fit testing protocol as defined by OSHA 1910.134(f)(7) using a mannequin headform simulating normal and deep breathing. Overall mannequin fit factor (mFFO) scores range from 0 to 200 and are quantitative assessments of fit that can be replicated with other quantitative fit-testing devices. Scoring ranged from 0 to 200, where a score greater than 100 indicates that no change in fit performance is detected. Tensile strength testing of the straps of three respirators was performed to determine per cent change in strap integrity (tensile strength of exposed respirators minus the controls divided by the controls).9–11

Comparisons between ozone-exposed and control groups were made using a two-tailed Student’s t-test with threshold for significance defined as p<0.05. No human subjects were involved in this research study.


Exposure to ozone at 400 ppm with 80% humidity for 2 hours effectively killed bacteria on the 3M 1870 N95 respirator, as shown in table 1.

Table 1

Log kill efficacy in ozone-exposed respirator pieces compared with the controls

There were no significant changes in filtration efficiency (p=0.45) nor filter resistance (p=0.84) between ozone-exposed 3M 1870 respirators and controls after exposure to ozone for five cycles with little to no noticeable wear, as shown in table 2. No visible degradation of the 3M 1870 straps was observed, and there were no significant reductions in tensile strength of top (0.2%, p=0.99) or bottom strap (5.3%, p=0.65).

Table 2

N95 Airflow resistance filtration efficiency and strap integrity testing after exposure to ozone 450 ppm at 75%–90% humidity for 2 hours per cycle

Quantitative fit testing revealed no loss of fit as a result of ozone exposures. Ozone-exposed respirators scored an average 180 (range=118–200) compared with the control average score 169 (range=121–200).


This is the first report demonstrating that an ozone gas application is effective at killing organisms that may contaminate N95 respirators and does not damage or degrade respirator filtration ability.

Ozone achieved high level of disinfection against PsA on N95 respirators, an organism that is more difficult to kill than SARS-CoV-2.7 Additionally, our experiments were performed with very high level of contamination (>10 million CFUs/4 cm2). The necessary conditions were 400 ppm ozone for 2 hours with relative humidity 80%. Similar to previous reports, humidity is essential in bacterial killing by ozone and fails to kill pathogens on respirators below 50% relative humidity.12 Lower concentrations of ozone for shorter periods of time may be effective on viruses based on previous work with influenza virus.13 14 Further testing is necessary to determine if such conditions exist.

Ozone did not degrade the function of the respirator filters after five cycles of ozone at 450 ppm at 75%–90% humidity for 2 hours. The filtration portion of N95 respirators consists primarily of polyethylene and polypropylene with which ozone does not easily interact. There was no change in fit detected by quantitative fit testing, and the straps did not experience significant reduction in tensile strength.

The ozone analyser is critical to this process. The ozone analyser continuously measures the presence of ozone in parts per million before, during and after ozone production. It ensures that a proper concentration is achieved and maintained. After 2 hours of ozone at 400 ppm, the ozone destruct unit was turned on. With the use of the ozone analyser, we ensured that no ozone is left in the chamber. The chamber has internal fans that homogenise the internal atmosphere to ensure no residual ozone lingers within the sealed chamber.

These data suggest that ozone may be an effective and easily accessible method for disinfection of N95 respirators, complementing the methods recently authorised under FDA emergency use authorisations.15–21 Future directions will focus on repeating these experiments using additional gram-negative and gram-positive bacteria, phages which are frequently used surrogates for COVID-19, and an airborne non-pathogenic virus. However, further studies are required to directly assess the effects of ozone on SARS-CoV-2.


The authors would like to thank Erica L Herzog of Yale University for her support; Sannel Patel, Jimmy Moler and Kevin York of Ozone Solutions (Hull, Iowa, USA) for the ozone application equipment used in this work and technical advice; Munroe Chirnomas for his technical advice; Alexandra Stephens for her assistance with the preparation of this manuscript and literature review; and Foad Katirai and Shigeru Saito for their assistance in translating references from Japanese to English. This work is dedicated to healthcare personnel and first responders who are fighting the COVID-19 pandemic.



  • Contributors EPM and MDS wrote the manuscript, planned and organised experiments, interpreted the results; SD planned and organised experiments and interpreted the results. BS originated idea of ozone disinfection of N95 respirators and interpreted results; PG and ZG interpreted the results and edited the manuscript; CSDC provided equipment and supervised experiments. LS planned, organised and performed experiments, interpreted the results and edited the manuscript.

  • Funding EPM was funded by NHLBI Training Grant T32HL007778. LS was funded by Francis B Parker Fellowship and American Lung Association Catalyst Award. Prototype ozone disinfection chamber was provided by Ozone Solutions (Hull, Iowa, USA).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data sharing not applicable as no datasets generated and/or analysed for this study. No applicable.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.