Immersive ultraviolet disinfection of E. coli and MS2 phage on woven cotton textiles

Characterization of UV-C fluence

Characterization of dosimeter fluence

Dosimetry revealed that the distribution of light inside the disinfection cabin is not uniform despite the 360° emission of UV-C irradiation. 15 and 25 W lamps were used during the characterization process. Multiple wattages have been tested due to supply chain constraints that may limit the availability of certain lamp models. The results of the dosimetry are summarized in Table 1. The dosimetry provides a colorimetric range specific to a UV-C fluence rather than a quantitative value. The 30 s irradiation cycle exposed all areas of the cabinet to a fluence of at least 43.8 mJ cm⁻². The minimum fluence delivered for the shortest use case exceeds a 3 log reduction value for pathogenic viruses such as SARS-CoV-2 on hard surfaces²¹.

The exposure cycle of 70 s allowed the device to reach a minimum of 50 mJ cm⁻² fluence in all areas of the cabinet. It should also be noted that several areas were very close to 100 mJ cm⁻² dosimeter threshold. It is essential to obtain a fluence delivered greater than that necessary for a hard surface, because a porous surface reduces the fluence delivered due to shading, geometry and textile characteristics²². The calculated fluence for the disinfection cabinet use case exceeds the required fluence necessary for a 3 log reduction of SARS-CoV-2^21,23,24. Dosimetry coupons have an upper limit of detection and are qualitative in nature, which has limited their use as a characterization tool.

Characterization of fluence by spectroradiometer

Spectroradiometer data indicated asymmetry in the disinfection cabinet in all orientations. Table 2 shows the fluence, considering a cycle of 70 s, averaged over the suspension positions and for each orientation of the radiometer. The upward orientation provided the lowest fluence, which was expected since no lamp is mounted on the ceiling of the cabinet. All UV light detected in the UP position reflects off the interior walls of the cabinet. A similar trend is seen for the front orientation, as the majority of UV light detected from this position is reflected off the cabinet doors. The distance from the ceiling to be detected was approximately twice the distance between the detector and the doors. Notably, the fluences for the Right and Front orientations of the detector were about half the fluence for the Back and Left orientations. The cause of this discrepancy is unclear, as the layout of the cabinet is laterally symmetrical.

The results in Table 2 further highlight the limitations of dosimetry coupons. Each orientation of the spectroradiometer is above 100 mJ cm⁻² upper limit of detection of the colorimetric coupon scale. This result indicates that the proper consideration of the drawbacks of dosimetry coupons must be taken into account before use.

Penetration of UV-C light through cotton layers

Light penetration using a collimated beam simulates light from a single source passing through a textile. An average irradiance of 3.97 W m⁻² (n = 30) was measured by the spectroradiometer when no tissue covered the detector. 82.1% and 93.0% of UV-C light was blocked when measuring light penetration through L1 and L2 respectively.

Blocking UV-C cabinet light via adjacent porous objects

Direct measurement of UV-C irradiance by spectroradiometry quantifies light intensity and addresses the sensitivity limits of coupon dosimetry. The average delivered fluence is constant in all suspension positions and exceeds 100 mJ cm⁻², but there are differences in the measured irradiance for specific orientations of the spectroradiometer. Lower fluences were calculated at all spectroradiometer positions where the detector was arranged in the forward and upward orientations, as shown in Figure 5. Spectroradiometer data is aligned with dosimetry data which showed similar patterns for this part of the cabinet. Despite the differences in irradiance, the experimental results indicate that the light from each direction inside the cabinet far exceeds 100 mJ cm⁻² in conditions where shading and textile micro-geometry are not factors. A detailed table containing this data can be found in the “Additional information”.

The impact of adjacent hanging clothes was examined after determining the light distribution for an empty wardrobe. The spectroradiometer was hung at each of the internal positions inside the cabinet (P1, 2, 3, 4), while T-shirts were hung on each other inside the cabinet. The proportion of light blocked was then calculated and is summarized in Table 3. Positions 1-3 had similar light dynamics while P4 had a lower proportion of light blocked.

The dynamics of light inside the disinfection cabinet changes as objects are placed inside. For example, a full cabinet (objects suspended at each position) would complicate the dynamics of light in the irradiated volume. Characterizing empty and full conditions provides insight into the amount of light blocked in each use case. Overall, the method described in this work for calculating a directional mean fluence provides a tool for understanding immersive UV-C devices.

Microbial disinfection by MS2 and E.coli

MS2 and E.coli were used as challenge organisms to assess the disinfection capabilities of the disinfection cabinet. The relative humidity of the room was measured and ranged from 12 to 24%. Relative humidity may have impacted the recovery threshold for MS2 and E.coli. Dry air can cause viruses and bacteria to dry out more on a garment before it is processed in the wardrobe.

Disinfecting MS2 cotton T-shirt coupons

Logarithmic reaction values ​​for MS2 on each of the inoculated regions are provided in FIG. 6. MS2 control inoculum concentrations ranged from 7.5 to 8.5 PFU cm⁻². The ultraviolet transmittance (UVT) of the inoculum averaged 72%. The recovery efficiency of MS2 was variable, with an average of 15.5% ± 23.5%.

The three front sections (F1, F2, F3) shown in Figure 6 have mean log reduction (LRV) values ​​of 1.58, 1.63 and 1.27 respectively. Additionally, the RA and RS sections achieved 1.34 and 1.36 LRV. LA and LS locations had the lowest LRV with 0.47 for LA and −0.07 for LS. A one-way ANOVA (α = 0.05) with shirt location as factor and LRV value as response confirmed that there is a significant difference between shirt locations. A Tukey post-hoc test revealed that forward sections (F1, F2, F3) and straight sections (RA and RS) are not significantly different from each other (P-value > 0.05). Therefore, there is a significant difference between the LA and LS sections and the rest of the locations tested.

Cotton disinfection E.coli t shirt coupons

E.coli was inoculated in the same way as MS2; however, modifications to the protocol have been implemented to compensate for difficulties in recovery E.coli from t-shirt coupons. The E.coli inoculation stock was adjusted to a concentration of approximately 10.5 log cm⁻² to compensate for losses due to cell death on a dry cotton support. In addition, E.coli had a greater sensitivity to desiccation compared to MS2, which made the inoculum unrecoverable once dried. This led to the elimination of the 20 minute drying step for E.coli experimentation. The UVT of the E.coli the inoculum was practically nil due to the concentration required for recovery. The recovery efficiency of E.coli gave an average of 4.0% ± 1.0.

F1, F2 and F3 reached higher mean LRV values ​​of 3.1, 2.9 and 2.4 respectively, as shown in Fig. 7. Additionally, the RA and RS sections achieved an average LRV of 2.3 and 2.0. LA and LS data showed the lowest LRVs of 0.7 and 0.9 on average. The LA section also resulted in the highest variability, with values ​​ranging from 0.45 to 2.8 LRV.

E.coli LRV for each inoculation location on hanging T-shirts (n=3). Right side (RS) and right arm (RA); Left Sleeve (LS) and Left Armpit (LA); Before 1,2,3 (F1, F2 and F3).

One-way ANOVA (α = 0.05) with shirt location as factor and LRV as response revealed a significant difference between shirt locations. A Tukey Post-Hoc test revealed that the F1, F2, F3 and RA sections were significantly different from LA, LS. The RA section was not significantly different from any of the sections. The lack of difference for this region is believed to be due to data variability and not a physical phenomenon. The orientation of the LA and LS regions inside the cabinet is believed to be the cause of the lack of disinfection in these areas compared to the right (RA and RS) and front (F1, F2 and F3) areas of shirt.

Impacts on immersive UV-C disinfection

The physical layout of the disinfection cabinet is the main source of significant disinfection differences for T-shirt regions. The hanging racks inside the cabinet are centered on the dimensions of the interior wall and not on the arrangement of the UV-C lamps. Misalignment of light sources and hanging positions results in uneven distribution of light in the cabinet. This work highlights the importance of understanding light distribution when designing and implementing immersive UV-C light technologies.

The type of material also has an impact on the effectiveness of disinfection in addition to light distribution. To the knowledge of the authors, there are currently no published manuscripts examining textile disinfection using immersive UV-C technologies. However, there is a large body of knowledge on the use of UV disinfection technologies in water treatment. The differences between disinfection of non-porous and porous surfaces are comparable to the disinfection of drinking water compared to the disinfection of wastewater. UV-C is effective for both water matrices, but there are additional factors that must be considered for proper use. The same concept applies to non-porous surfaces versus porous surfaces.

As UV-based technologies are used in wider applications in areas such as schools, airports or stadiums, the impacts of porous materials on UV effectiveness must be considered. This work provides benchmark data that shows that the fluences needed to achieve disinfection on a porous surface are orders of magnitude higher than the fluences needed to disinfect a hard surface.