Article Text

Download PDFPDF
Original research
Fit and filter integrity of a respirator mask after multiple cycles of autoclave
  1. Nidhi Desai1,
  2. Hasan Naveed2,3,
  3. Michael Koenig3,4,
  4. Jörg Huber5,
  5. Nikil Kapur6,
  6. Mahmood F Bhutta3,7,
  7. Saul Rajak2,3
  1. 1 Brighton and Sussex University Hospitals NHS Trust, Brighton, UK
  2. 2 Ophthalmology, Sussex University Hospitals NHS Foundation Trust, Brighton, UK
  3. 3 Brighton and Sussex Medical School, Brighton, UK
  4. 4 Cellular Pathology, Brighton and Sussex University Hospitals NHS Trust, Brighton, UK
  5. 5 School of Health Sciences, University of Brighton, Brighton, UK
  6. 6 School of Mechanical Engineering, University of Leeds, Leeds, UK
  7. 7 Ear, Nose and Throat, University Hospitals Sussex NHS Foundation Trust, Brighton, Brighton and Hove, UK
  1. Correspondence to Saul Rajak, Ophthalmology, Brighton and Sussex University Hospitals NHS Trust, Brighton, BN2 5BF, UK; saul.rajak{at}


Background The COVID-19 pandemic has led to profound shortages in personal protective equipment worldwide. The availability of filtering facepiece class 3 (FFP3) respirators could be greatly increased if they could be reused after sterilisation.

Aims To determine the effects of repeated autoclave cycles on the fit and filter function of the 3M 1863 FFP3 disposable respirator.

Methods Participants underwent fit tests with 3M 1863 FFP3 respirators. Respirators were subjected to autoclave cycles and a repeat fit test was conducted after each cycle until failure. The filter function of both unused and autoclaved respirators was determined by quantitatively assessing the differential pressures and filter penetration of aerosolised sodium chloride particles. Mask structural inspection was also carried out by light microscopy.

Results A total of 38 participants were recruited. Repeat fit testing with a new respirator was passed by 30 of 38 (79%) participants in comparison with 31 of 38 (82%) of participants after the respirator had undergone one autoclave cycle. There was fit test failure with further rounds of autoclave. There was no evidence of structural changes after one autoclave cycle, but the nose foam began to separate from the mask following further cycles. Filter efficiency of all 15 autoclaved respirators that underwent filter testing was 97.40% or more. Differential pressure (breathability) of respirators was unaffected by autoclaving.

Conclusions 3M 1863 FFP3 respirator retains good fit and filter function after a single autoclave cycle. Addressing nose foam separation and further testing to EN149 standards would be required before respirators could be considered for reuse.

  • health services research
  • healthcare quality
  • access
  • and evaluation
  • infections
  • virus diseases

Data availability statement

Data are available upon reasonable request. The datasets generated during the current study are available from the corresponding author on reasonable request.

Statistics from

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.

Summary box

What are the new findings?

  • 3M 1863 filtering facepiece class 3 masks retain good fit and filter function following one autoclave cycle.

  • One autoclave cycle does not substantially alter the light microscopic structure of this type of face mask.

How might it impact on healthcare in the future?

  • Autoclaves may provide to be a useful method of reusing respirators in healthcare settings.

  • Further investigations regarding separation of nose foam from respirator material are needed prior to clinical use.


Filtering facepiece (FFP) respirators are an essential component of respiratory protective equipment. They protect wearers against airborne infection, including the novel SARS-CoV-2 virus which is predominantly spread via droplet and aerosol inhalation.1 Healthcare workers are advised to wear respirators in high-risk encounters, such as close contact with patients and when performing aerosol-generating procedures.2 The COVID-19 pandemic has led to profound personal protective equipment shortages in almost all healthcare sectors, which may be exacerbated as respirator use becomes widespread if the disease persists and there are further waves. Therefore, extended use, reuse and sterilisation of respirators are being explored.3

Respirators are regulated against strict standards such as the USA National Institute for Occupational Safety and Health standard 42CFR84 for N95 (specifying 95% filter efficiency) or the European standard EN149:2001 for FFP class 3 (FFP3) respirators (specifying 99% filter efficiency).4 Respirator performance is assessed by specific tests measuring particulate filtration efficiency and breathing resistance. There are three categories of FFP respirators: FFP1, FFP2 and FFP3 are tight-fitting respirators with minimum filter efficiencies of 80%, 94% and 99%, respectively.5 There is wide variation in availability and usage of these respirators, but typically FFP3 respirators are used in higher risk settings such as operating theatres or wards designated for patients with respiratory infections such as COVID-19.

Optimal function of respirators is dependent on fit, whereby an airtight seal is formed around the mouth and nose of the wearer, and filter efficiency of the respirator which reflects the barrier capability against particles. Effective respirator function also depends on its breathability (resistance to air flow), which must be balanced with the conflicting requirement for non-penetration of particles. Successful reprocessing of respirators requires elimination of viable infective particles, without compromising respirator seal, filter efficiency or breathability. Sterilisation methods using chemicals or radiation are effective at destroying pathogens but can cause damage to filter integrity.6–9 There is still a relative paucity of research on the effect of autoclaves on respirator masks.

The 3M 1863 FFP3 respirator (3M, Bracknell, England) is commonly used in the UK National Health System. It is a valve-less mask that consists of three panels of non-woven material with a nose clip in the top panel allowing the wearer to mould the respirator for a better fit. This study evaluates the effect of multiple rounds of autoclave sterilisation at 121°C on seal integrity (assessed via qualitative fit testing) and microscopic filter integrity (assessed via filter aerosol penetration) of the 3M 1863 FFP3 respirator.

Materials & methods


Participants were a convenience sample of consenting healthcare workers recruited from ophthalmology and ear, nose and throat departments of Brighton and Sussex University Hospital. Participants were excluded if they failed an initial fit test on an unused 3M model 1863 respirator, or if they had symptoms consistent with COVID-19. Participants were only fit-tested on an unused respirator, or one that they personally had worn previously. All participants were trained how to correctly wear an FFP3 mask prior to each session and a user seal check was performed during every fit test.


The 3M model 1863 respirators went through one or more cycles of autoclave treatment using a B type vacuum autoclave (Eschmann Little Sister SES3000B, UK), which has a 17-litre chamber suitable for processing porous materials. Masks were placed directly onto metal trays and then placed on shelves inside the autoclaves. A standardised, preprogrammed cycle to the requirements of EN13060 (standard for small steam sterilisers) was used with a total duration of approximately 60 min, comprising of an initial air removal stage, a heat to sterilisation stage, a sterilisation hold stage of 121°C–124°C at an absolute pressure of 205–225 kPa for 15 min, with a subsequent discharge and vacuum drying phase.10

Assessment of seal integrity

The integrity of the respirator seal was assessed through both visual inspection and qualitative fit testing.11

Visual inspection comprised inspection of: (1) the outer and inner respirator surfaces for damage or deformity, (2) the attachment and elasticity of the elastomeric headbands, (3) the shape of the nose clip area. Participants were also asked to comment on the comfort and fit of respirators.

Fit testing was conducted by a single certified fit testing practitioner in accordance with the Occupational Safety and Health Administration protocol. In brief, a plastic hood (3M FT30 kit) was placed over the participant’s head. Participants initially underwent a sensitivity test which involves a nebulised bitter (Bitrex, 3M FT31) or sweet (saccharin, 3M FT11) solution sprayed through a hole in the hood. If participants could not detect the taste, they were excluded from the study. Those remaining donned the respirator and the plastic hood, and performed seven different repeated head exercises for 1 min each. A Bitrex (3M FT32) or saccharin (3M FT12) solution of higher concentration than that used for sensitivity testing was sprayed into the hood at 30-second intervals. If the participant tasted the solution at any point, the fit test was classed as failed.

Anecdotally, it is recognised that fit testing is not entirely reliable, and reliability is likely specific to each model of respirator. Therefore, all participants underwent consecutive fit testing on two unused respirators to establish a benchmark for test–retest variation. Results of fit testing after autoclaving were compared to this benchmark using McNemar’s test (p<0.05 significance threshold).

Light microscopy

A respirator that had been autoclaved at 121°C and a new respirator were examined using a conventional light microscope (Olympus BR-40) by a histopathologist, who was blind to which respirator was which. The filter was separated into its three component layers, with each layer inserted between two glass slides and examined at low and high magnification under polarised and non-polarised light for structural degradation or distortion.

Assessment of filter performance

Filter efficiency and breathability was assessed in a commercial filter testing laboratory (Filter Integrity, Sedgefield, UK). Efficiency was tested on a sample of respirators in the study (three each after one to four cycles of autoclave and three new respirators) using an aerosol of solid sodium chloride (NaCl) particles generated from solution.

The respirator was secured to a holder using hot melt and tape along its perimeter and placed inside a controlled test chamber comprising of a sealable metal cabinet with inlet and outlet connections (figure 1). A dispersion bar was fitted on the inlet port to prevent aerosol jetting onto the filter during testing. An aerosol of solid NaCl was generated from 1%–2% NaCl solution and passed through the test chamber. Temperature and humidity levels inside the chamber were maintained in accordance with standard requirements (<40% relative humidity, 22°C±3°C). The full area of the mask was exposed to the aerosol flow and the domed adaptor spread the mask in a similar fashion to when it is on the face.

Figure 1

Schematic of sodium chloride aerosol test apparatus.

Upstream aerosol distribution was measured using a Palas U-SMPS (an ‘electrical mobility method’) according to BS EN 13274-7-2019 standard specifications. Filter efficiency was measured over a 0.2–10 µm range using a Palas Promo 3000 aerosol spectrometer fitted with a single sensor located downstream of the test filter.

A benchmark reading was taken prior to respirator attachment to measure upstream aerosol samples. After sealing the respirator to the holder, five downstream samples were taken at flow rates of 30 and 55 L/min. A further test run was carried out using U-SMPS in classifier mode to assess specific number efficiency at 0.1 µm at 55 L/min. Both the challenge and the penetrating fractions were measured.


We recruited 38 participants. Their average age was 41 years (range 19–65) and 22 participants were female.

The test–retest reliability of new 3M 1863 respirators was 30 of 38 (79%). After one cycle of autoclave, 31 of 38 (82%) participants passed fit testing (p=0.743). Of the 31 respirators that had passed one cycle, 12 of 26 (46%), 4 of 10 (40%) and 0 of 3 participants passed fit testing after two, three and four autoclave cycles, respectively (figure 2) with five, two and one participants being lost to follow-up after each autoclave cycle.

Figure 2

The number of respirators passing/failing fit testing after one to four cycles of autoclave.

On visual inspection before and after autoclaving, a variable degree of separation between the grey nose foam and the respirator fabric layers was observed (figure 3). There was no evidence of other macroscopic changes or deformities after autoclaving. Seven of 38 (18%) of participants reported that material felt less tight around the nasal bridge in post-autoclave respirators, with no other subjective changes noted.

Figure 3

A 3M 1863 respirator before (left) and after (right) one autoclave cycle, with arrow marking the separation of nose foam.

Inspection with light microscopy revealed that the interlacing material patterns in all layers of the respirator materials were visually identical, and there was no difference in the refractile patterns of the middle layer under polarised light (figure 4).

Figure 4

Light microscopy at 40× magnification of the individual layers of a new respirator (A) and a respirator after one autoclave cycle (B): (i) outer layer, (ii) middle layer, (iii) inner layer.

Filter integrity testing

The filter efficiencies for particles of 0.3 µm and 0.1 µm in size at flow rates of 55 L/min and 30 L/min are shown in table 1. There was a trend towards slightly lower efficiency with repeated autoclave cycles, but the lowest recorded efficiency was 97.40%.

Table 1

Filter testing results of 15 respirators (three each of one to four autoclave cycles and three new respirators)

The operating differential pressure at 55 L/min did not exceed 62.80 Pa (table 1), which is below the maximum breathing resistance allowance (1.00 mbar/100 Pa) defined by the EN149 standard.


The COVID-19 pandemic has led to large-scale depletion of single-use respirators. In low-income countries, respirator availability is restricted, and therefore reuse may be the only option. Furthermore, the environmental harms could be significant: it has been estimated that daily respirator use by the UK population could generate over 128 000 tonnes of unrecyclable waste, putting extreme strain on waste disposal services and potentially exposing disposal workers to infection risk.12 Strategies for increasing reuse of respirators are urgently required.

Autoclaves are widely used for equipment reuse in a variety of healthcare settings, and should ensure a six log (99.9999%) reduction in microorganisms if saturated steam is held at a temperature of a minimum 121°C for at least 15 min.10 Previous research suggests that a single cycle of autoclaving should destroy influenza viral particles on a respirator.9

Our study indicates that the filter function of the 3M 1863 respirator is materially unaffected by up to four cycles of autoclave sterilisation. Throughout testing, all the respirators retained breathability of less than 1.00 mbar and 97.40% or greater efficiency for 100 nm (0.1 µm) and 300 nm (0.3 µm) diameter particles. Although 300 nm particles are the size tested when certifying to the EN149 standard, 100 nm particles are thought to be the most penetrative for this type of filter, and representative of the SARS-CoV-2 virus which is 50–200 nm in size.13 FFP3 respirators should have a minimum efficiency of 99%, however, a drop-off to 98% will not place the user at materially increased risk, and FFP2 or N95 respirators with 95%–96% efficiencies are already deemed acceptable for use in high-risk settings.5 We found that the 3M 1863 respirator seal was reliably maintained for only one cycle of autoclave, with an increasing proportion of masks failing on fit testing after further cycles. We suspect this derives from adhesive failure leading to the separation we noted between the nose foam and fabric layers.

Several previous studies have investigated the effect of autoclaving on the integrity of respirators, but have used a variety of autoclave protocols and respirators, which may explain varying outcomes.6–9 Viscusi et al reported a marked worsening of fit of the 3M 8210/MHI and the Moldex 2200 respirator after 30 min of 60°C moist heat treatment.8 Lore et al reported that 3M model 1860 and 3M model 1870 respirators retained at least 95% efficiency for 300 nm sized particles after being treated with moist heat reaching up to 65°C±5°C for 3 hours but did not assess respirator fit after treatment.9 Harskamp et al reported shape deformities in 50% of tested FFP3 respirators following 34-minute autoclave cycles at 121°C, whereas all FFP2 respirators retained functionality with no reduction in breathing resistance.14 Lin et al reported a notable change in physical appearance and a rise in particle penetration in a N95 respirator (model not specified) after a 15-minute 121°C autoclave cycle.6 Meisenhelder et al reported in an abstract (full paper not available) that 3M FFP3 masks maintained adequate fit and seal after 40-minute 121°C autoclave cycles, but significantly declined in filter efficiency (<95%) following loading cycles.15 A further non-peer-reviewed study of the 3M NR D 8822d FFP2 respirator reported marked deformation of the mask after 90°C autoclave cycles such that fit testing was not attempted, and no deformation after a 60°C, but failure of fit testing.16 However, in contrast, de Man et al recently reported that the ‘functionality’ of the 3M 1862 FFP2 mask was unaffected by a 15-minute autoclave cycle, and a non-peer-reviewed manuscript by Kumar et al reported that there was no loss of structural or functional integrity of four different models of N95 mask (3M Aura 1870, 3M VFlex 1804S, 3M 9210 and AO Safety 1054S) after 10 cycles of autoclave sterilisation (121°C for 15 min) but there was failure of the 3M 1860 and 8210 respirators after the first cycle.17 18

For the 3M 1863, we found that the limiting step to reuse was a failure of the seal integrity rather than of the filter. Two differing and potentially complementary strategies may overcome this limitation: use of different parameters for the autoclave, or a change in respirator materials. We found no evidence that a temperature of 121°C materially affected filter function at up to four cycles of autoclave, however, others have reported that heating non-woven polypropylene materials above 100°C can cause irreversible filter damage and lead to loss of electret charge thought to be important for repelling microorganisms.19 Many microorganisms are destroyed at temperatures lower than 100°C: for example, exposure of the SARS-CoV-2 virus to 56°C for 20 min leads to inactivation.20 Therefore, a lower temperature may lead to better preservation of the materials of the mask without compromising sterilisation. We have conducted preliminary testing of the 3M 1863 masks with autoclave temperatures of 85°C and 110°C (data not presented) and found that it does not appear to cause the same degradation of adhesive around the nose foam. It is also possible that altering the position of the mask during autoclave (for example, straightening the mouldable nosepiece) may reduce stress on the join between the nose piece foam and the fabric and prevent failure.

Nevertheless, most autoclaves run on preprogrammed cycles, and adjustment to such cycles may not be possible or feasible at scale. A better solution may lie in changing the materials used in the manufacture of respirators to those resistant to a higher autoclave temperature. To date, we are not aware that any manufacturer has attempted such an endeavour.

The respirator used in this study was designed, manufactured and certified for single use only. While the present study provides promising data on respirator fit and filter function after autoclaving that may assist respirator manufacturers and policymakers, there are limitations of this study and further important considerations and potential differences between new and sterilised respirators: first, all the respirators autoclaved in this study had not been worn by the user other than for fit testing and we are therefore unable to determine if the fit or filter function is compromised by respirator wear. However, as the fit and filter function are not markedly affected by 60 min of steam autoclave reaching temperatures of over 120°C, it is unlikely that a short period of wear would cause significant damage. Second, in this study, the respirators were not washed prior to autoclaving (which would likely damage the woven fabric) in contrast to the standard process for surgical equipment sterilisation in which debris is removed by washing prior to autoclaving. However, respirators are very unlikely to be in contact with macroscopic patient debris and if they are soiled during usage, reuse would not be considered. Third, this study examined filter function using a particle count method, rather than aerosol photometry which is specified in EN149:2001 requirements.21 Mask filtration was only tested with aerosol loading in the laboratory. Further testing would be required to determine the potential effect of loading by wearer use to ascertain the effect on filter function in a ‘real-world’ setting. However, our aim was to ascertain respirator filter function for a range of flow rates and particle sizes (including ultrafine particles which approximate viruses and other pathogens)22 rather than to achieve certification. Fourth, we performed filter testing with non-neutralised aerosolised salt, which may reduce the ability to test the residual electrostatic charge function of the fibres. N95 respirator testing in the USA requires the use of neutralised salt, but EN149 standards do not. Fifth, qualitative fit testing was used, which provides a subjective measure of seal integrity rather than quantitative fit testing which uses objective measures. This qualitative approach could introduce bias as participants were not blinded to the autoclave status of their respirator. Sixth, we tested the 3M generation 1 1863 respirator, but the generation 2 1863+ respirator is likely to be increasingly widely used, and although it would be expected to perform similarly after autoclave cycles, it would need to be formally tested. Seventh, participants were not blinded to if/how many times their mask had been autoclaved in order to ensure that each participant was only fit tested with their ‘own’ mask. This could introduce bias to the subjective fit test outcomes. However, the noxious agent taste is extremely strong and we therefore think it is unlikely that participants would report the taste not to be present when they could in fact taste it, although it is possible that if they were consciously or subconsciously biased towards masks not being reused that they could erroneously report fit test failure.

We have demonstrated that despite a small amount of degradation, the 3M model 1863 FFP3 respirator retains good fit and filter function following one autoclave cycle of 121°C, and that there are no substantial changes to the light microscopic structure after such treatment. Further research is required to address the separation of the nose foam from the mask fabric, which likely underlies the failure of the seal and the effect of loading of the mask through user wearing in a real-world setting. As autoclaves are widely available in healthcare settings, overcoming such barriers may enable an accessible and affordable method for reusing respirators.

Data availability statement

Data are available upon reasonable request. The datasets generated during the current study are available from the corresponding author on reasonable request.

Ethics statements

Ethics approval

The Health Research Authority ethics decision tool determined that formal ethical review was not required for this study.


The authors would like to acknowledge: Eschmann Technologies, Lancing, UK for the provision of autoclaves and sterilisation expertise; and Filter Integrity, Durham, UK for testing the respirator filters and filter expertise.



  • Contributors MFB and SR designed and conceived the study. NK assisted in planning the study. ND and MK collected the data. ND, MK, JH, MFB and SR performed data analysis. ND, HN, MK, MFB and SR wrote the manuscript. All authors reviewed and approved the final manuscript.

  • Funding This research study has been supported by a grant awarded by the Applied Research Collaboration Kent, Surrey and Sussex (ARC KSS).

  • 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.

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