Far-UVC eciently inactivates an airborne pathogen in a room-sized chamber

Many infectious diseases, including COVID-19, are transmitted by airborne pathogens. There is a need for effective environmental control measures which, ideally, are not reliant on human behaviour. One potential solution is Far-UVC which can eciently inactivate pathogens, such as coronaviruses and inuenza, in air. When appropriately ltered, and because of its limited penetration, there is evidence that Far-UVC does not induce acute reactions in the skin or eyes, nor delayed effects such as skin cancer. While there is laboratory evidence for far-UVC ecacy, there is limited evidence in full-sized rooms. In the rst study of its type, we show that Far-UVC deployed in a room-sized chamber effectively inactivates aerosolised Staphylococcus aureus. At a room ventilation rate of 3 air changes per hour (ACH), with 5 ltered sources the steady-state pathogen load was reduced by 92.1% providing an additional 35 equivalent air changes (eACH). This reduction was achieved using Far-UVC intensities consistent with current regulatory limits. Far-UVC is likely to be more effective against common airborne viruses, including SARS-CoV-2, and should thus be an effective and “hands-off” technology to reduce airborne disease transmission. The ndings provide room-scale data to support the design and development of safe and effective Far-UVC systems. performed step towards real-world studies are experiments in large, chambers. These chambers, controlled are to replicate a real-room Such spaces have demonstrate effectiveness GUV study of microorganisms They can provide insight into application environmental air indication closer real-world performance. Here we investigate for the rst time the ecacy of Far-UVC for an airborne pathogen under steady-state conditions in a full-scale room-sized bioaerosol chamber.


Introduction
Severe acute respiratory coronavirus 2 (SARS-CoV-2), the virus responsible for the COVID-19 pandemic, can be transmitted by a single individual to one or more people through viral transport in ne airborne particles [1][2][3][4] . The risk of airborne SARS-CoV-2 transmission, from such events, increases in indoor environments where large groups of people congregate, especially when the environment is poorly ventilated 5,6 .
As has been well documented, the high levels of SARS-CoV-2 transmission have overwhelmed national healthcare systems, resulted in millions of deaths and caused long term health problems. The impact on the global economy has been, and will continue to be, devastating, which in turn has resulted in further welfare and public health issues.
It is therefore clear that reducing or preventing SARS-CoV-2 transmission is a critical and unprecedented global challenge.
Transmission control measures have included national lockdowns, restrictions on social and business gatherings, improved indoor ventilation, public health campaigns, protective face coverings and vaccination. These control measures have different success rates and each comes with its own challenges. Vaccination has been one of the most effective measures in reducing death and serious illness, although the evidence is unclear on the e cacy of vaccination in reducing disease transmission 7,8 . Face coverings can be an effective control measure for reducing the risk of airborne transmission but rely on individual behavioural choices, with high levels of compliance required to achieve population level impacts on transmission 9,10 . As the COVID-19 pandemic progresses in time, there is lower acceptance and adoption of control measures that impact on daily life, and therefore an increased need for effective measures that do not rely upon human behavioural choices 11,12 . This is also important beyond COVID-19; airborne transmission has been recognised as an important mechanism for a wide range of other viral infections including in uenza, measles, other human coronaviruses (SARS-CoV, Middle East Respiratory Syndrome MERS-CoV) and Respiratory Syncytial Virus (RSV) as well as for bacterial infections including Tuberculosis and some pathogens responsible for hospital acquired infections [13][14][15] .
Germicidal ultraviolet (GUV) is a control measure which meets the above requirements, with a scienti c track record of success. In 1942, Wells et al. demonstrated less transmission of measles and mumps between children within upper-room GUV irradiated classrooms compared to control groups in rooms without GUV 16 . Similarly, Escombe et al. demonstrated a greater than 70% reduction in transmission of tuberculosis from patients to guinea pigs when upper-room GUV was utilised, with 35% tuberculosis infection in control group and 9.5% infection in the group with GUV 15 . However a major challenge for conventional 254 nm GUV is accidental exposure of humans, which can result in potentially painful sunburn-type reactions in the skin and cornea 17 . This limits traditional GUV to carefully designed upper-room systems, enclosed units or to irradiation of unoccupied rooms. Even when adopted in this manner, accidental exposures can still occur and affect technology adoption 18,19 .
A potential solution is 'Far-UVC', germicidal ultraviolet-C radiation typically in the wavelength range from 200-230 nm. Currently, the most common source of Far-UVC is Krypton Chloride (KrCl) excimer lamps with a primary emission wavelength of 222 nm, and low residual emission throughout the ultraviolet region of the electromagnetic spectrum 20 . The germicidal properties of KrCl excimer lamps have been shown in laboratory experiments to inactivate gram-positive and gram-negative bacteria, drug-resistant bacteria, in uenza viruses and human coronaviruses including the SARS-CoV-2 virus [21][22][23][24][25][26][27] . Importantly, ltered KrCl Far-UVC excimer lamps are much less likely than conventional (254 nm) GUV sources to induce acute adverse reactions on skin and eyes, and studies to date in animal and human models have not demonstrated any long-term adverse health effects 20,28−33 .
Whilst the laboratory results are encouraging, inactivation of a pathogen in a controlled bench-scale laboratory environment does not necessarily translate into reduced disease transmission when the technology is deployed with 'real-world' limitations 34 . Historical precedent with upper-room GUV provides some con dence in the potential for Far-UVC to reduce disease transmission, however there remains an unmet need for real-world evaluations 16,35 . Such studies are complex and must be performed over prolonged periods of time (typically at least 12 months). A translational step towards real-world studies are experiments in large, room-sized, aerosol chambers. These room-sized chambers, with controlled air-ow, temperature and humidity are designed to replicate a real-room environment. Such spaces have been used to demonstrate the effectiveness of upper-room GUV systems and to study the survival and dispersion of microorganisms 36-41 . They can also provide signi cant insight into the application of technologies in rooms where an infectious person may be present over a prolonged period of time, a situation that is common in schools, workplaces, hospitals and hospitality venues. With the continual controlled release of airborne pathogen, achieving a steady-state environment, the air within the chamber can be regularly sampled both with and without the environmental air disinfection technologies, providing an indication closer to real-world performance. Here we investigate for the rst time the e cacy of Far-UVC for inactivating an airborne pathogen under steady-state conditions in a full-scale room-sized bioaerosol chamber.

Results
Five Far-UVC lamps were secured to the ceiling of a room-sized bioaerosol chamber at the University of Leeds, with the lamps arranged in a quincunx pattern ( Fig. 1) with their emission directed towards the oor. Studies were undertaken either with all ve lamps on or with only the central lamp on. The mechanically ventilated 32 m 3 chamber was operated at a ventilation rate of three air-changes-perhour (ACH) and a continuous release of aerosolised Staphylococcus aureus was introduced to the room. After a 60 minute stabilisation period, 10 air samples were taken over a 50 minute period. Then either one (the central Far-UVC lamp) or ve Far-UVC sources were switched on and the sampling continued for a further 50 minutes.
These measurements were repeated using 3 different lamp exposure rates ( Table 1). The exposure rates chosen were motivated by existing optical radiation exposure limits ("Medium" scenario) and proposed threshold limit values ("High" scenario) 17,42 . An additional scenario at much lower lamp intensity was also included ("Low" scenario). Statistical analysis is detailed in Table S1.
As described in the "Methods" sections, the concentration of viable S. aureus pathogens in air at the collection location ( Fig. 1), was serially assayed for 4 minutes every 5 minutes, both before and after the lamps were switched on ("lamp on"). The results, quanti ed as colony forming units per cubic metre (cfu m − 3 ), are shown in Fig. 2 and Table 1, both for the 45 minutes prior to "lamp on", and serially for 50 minutes after "lamp on". The values after "lamp on" are expressed as percentages of the average values prior to lamp on. Again it is emphasized that the pathogen was continuously released into the room throughout the experiment. Table 1 Average percentage pathogen reduction, irradiance and calculated 8-hour exposure dose for three different exposure conditions at two heights from the ground. The bold, italicised 8-hour exposure values are above the 222-nm exposure limit of 23 mJcm -2 . Statistical signi cance is represented by: ns = p > 0.05, *=p ≤ 0.05, **=p ≤ 0.01, ***=p ≤ 0.001, and ****=p ≤ 0.0001.. As expected, the highest reduction in the steady-state airborne viable S. aureus load was with the "High" exposure scenario. Using all ve lamps this reduced the steady-state viable pathogen load by 98.9% compared to ventilation alone (three air-changes-per-hour). The peak 8-hour exposure dose in this "High" scenario is outside the current exposure limits but within, and indeed motivated by, the proposed increase in the American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value (TLV) for the skin (415 mJcm − 2 at 222 nm over 8 hours) 42 . A single lamp in the "High" exposure scenario did not exceed the average 8-hour exposure dose and reduced pathogen load by 93.9%, which produces an estimated equivalent air change rate of 46 eACH. Although the single lamp did not irradiate the full room, good air mixing in the chamber is likely to have resulted in this very substantial effect.

Peak Values
The "Medium" exposure scenario, with a peak 8-hour exposure dose motivated by the current exposure limit at 222 nm of 23 mJcm − 2 , produced a 92.1% reduction in the steady state viable pathogen load using all ve lamps. This corresponds to 35 eACH, equivalent to over 11 times the baseline ventilation with new steady state reached in under 15 minutes. It is relevant to note that while the 8-hour peak exposure dose is slightly higher than the current exposure limits, the average 8-hour exposure dose was more than 5 times lower ( Table 1).
The "Low" exposure scenario, with very low intensity Far-UVC exposure rates (a factor of 10 lower than the "Medium" exposure rate scenario), produced a 9% (one lamp) and 32% ( ve lamps) reduction in viable pathogen load.

Discussion
We have demonstrated for the rst time in a realistically sized room, with typical ventilation and a continuous source of airborne pathogens, the potential for Far-UVC to rapidly produce signi cant reductions in airborne pathogens. With the lamp intensities at a level where the exposure limits would not be exceeded, a ~ 92% reduction in viable pathogens was demonstrated, taking less than 15 minutes to reach the new ambient level. At much higher lamp intensities, more than 15 times higher, a ~ 99% reduction was demonstrated, taking less than 5 minutes. A comparison of the two scenarios described is shown in Fig. 3.
Although our study was not performed with SARS-CoV-2 for safety reasons, aerosolised S. aureus pathogen was used as a surrogate for more relevant (in the current context) airborne viruses such as human coronaviruses and in uenza viruses. The rationale for this is shown in Fig. 4, where inactivation rates by Far-UVC of airborne human coronaviruses (OC43 and 229E), airborne in uenza virus (H1N1), and airborne S. aureus are compared 26, 27 . All these inactivation rates were measured using the same laboratory setup. No corresponding results have been reported for Far-UVC inactivation of airborne SARS-CoV-2, but corresponding results for Far-UVC inactivation of SARS-CoV-2 on surfaces suggest similar sensitivity to human coronaviruses OC43 and 229E 43 . Our results demonstrate that airborne S. aureus is less sensitive to inactivation by Far-UVC than airborne in uenza and human coronaviruses, from which we conclude that S. aureus is a conservative surrogate. It is hypothesised that percentage reductions achievable for airborne coronavirus and airborne in uenza virus would likely be larger, and have shorter inactivation times.
For installers of Far-UVC it may be challenging to interpret and apply the optical radiation exposure limits 17,42 . Many will opt for the conservative approach of assuming an 8-hour exposure at the peak irradiance. However exposure limits are intended to be used with a Time Weighted Average (TWA) irradiance (E TWA ), which considers the actual exposure an individual has received within a space 44 .
This allows for a higher peak irradiance if the E TWA remains within limits. In this study, the peak lamp intensities could have been ve times higher than the "Medium" scenario, thereby improving inactivation, and the average 8-hour dose would still be within exposure limits.
This highlights the importance of correct installation of Far-UVC, to ensure the designated space is appropriately and safely irradiated.
For example, whilst a single lamp in the "High" scenario produced an overall ~ 94% pathogen reduction, there were areas of the chamber which were not fully irradiated. For real rooms, which may be larger and have potentially less effective air mixing than the chamber used in these experiments, the actual pathogen reduction may be signi cantly lower. As a result of previous modelling studies, we introduced a diffusing material to all of the Far-UVC sources within the chamber to broaden their irradiation pattern and increase Far-UVC coverage 45 .
Our results also provide some initial data that enable comparison to other technologies particularly portable air cleaners. These typically have a clean air delivery rate (CADR) between 200 m 3 h − 1 and 500 m 3 h − 1 depending on the size of units. For the experimental chamber this would result in between 6.2 and 15.5 eACH assuming that the portable air cleaner could mix the air su ciently in the room to achieve the theoretical maximum performance. Therefore the "Medium" Far-UVC scenario with 5 lamps performs substantially better than even a higher ow HEPA based air cleaner. Although the design and installation of a Far-UVC system has a higher degree of complexity than a "plug and play" portable air cleaner, the approach has the potential to offer far greater eACH and is also silent.
All methodologies designed to reduce airborne transmission of diseases such as COVID-19 would ideally be used within a layered approach involving, as appropriate, vaccination, social distancing, masks and ventilation. Further work is required to explore the in uence of parameters such as temperature, humidity, ventilation rates and proximity to infectious source but the results reported here should provide con dence that Far-UVC, when deployed appropriately, and conforming to current (or future) safety regulatory limits, is likely to be an effective, human behaviour independent, control measure to inactivate key airborne pathogens such as human coronavirus and in uenza -and thus reduce the airborne, and potentially surface, transmission of these diseases.

Bioaerosol chamber
Experiments were conducted in a controlled bioaerosol chamber with dimensions 4.26 m in length, 3.35 m width and a height of 2.26 m. The chamber is mechanically ventilated and operated under negative pressure with a full fresh air system that is HEPA ltered on the supply and extract to provide both experimental control and safety in operation. Ventilation air was supplied through a high level wall mounted inlet grille located in one corner of the room. The wall mounted air outlet is located diagonally opposite at low level. The chamber was operated at an air ow rate of 0.027 m 3 s − 1 equivalent to three air-changes-per-hour (ACH). The release location of the aerosolized Staphylococcus aureus was at a height of 168 cm from the ground, 50 cm from the air inlet and 64 cm from the adjacent wall (Fig. 1). The sample collection point was at a height of 50 cm, positioned 20 cm from the air outlet and 64 cm from the adjacent wall. Prior studies have indicated that this location is representative of the average concentration within the chamber. Care was taken to ensure the bacteria release point and sample point were not located directly under a Far-UVC source (Fig. 1). The chamber was operated at a temperature of 28 o C ± 1 o C and relative humidity 50 % ± 2 %. As a biocontainment facility, experiments were conducted with the chamber sealed and nebulisation, aerosol sampling and operation of the Far-UVC devices were carried out remotely.

Choice of Aerosolized Pathogen
In practice, the bioaerosol chamber could not be used with aerosolized level-3 pathogens such as SARS-CoV-2. In order to choose a usable aerosolized pathogen which would be a reasonable but conservative model for airborne human coronavirus, we undertook some preliminary studies using the Columbia University laboratory-based aerosolized pathogen UV irradiation system, as described by Welch et al. 27   The airborne Staphylococcus aureus was allowed to establish a steady state within the chamber over a period of 60 minutes. This steady state is similar to having an infected individual in the corner of the room breathing aerosolised pathogen into the room. Then, ten air samples of four-minute duration were taken every ve minutes at the collection point (Fig. 2), with the other minute being used to prepare the next sample. The Far-UVC lamps were then switched on and the sampling was repeated in the same manner. An average of the rst ten air samples was used to determine the concentration of Staphylococcus aureus (cfu m − 3 ) present in the chamber prior to switching on of the Far-UVC lamps. The concentration (cfu m − 3 ) of each subsequent air sample was then plotted as a percentage of the average initial steady state concentration.

Analysis
Concentrations of Staphylococcus aureus were normalised by comparing to the mean concentration of all samples prior to switching on the Far-UVC devices to enable comparison within and between experiments. Steady state concentrations with the lamps switched on were determined from the six measurements taken between 20 and 50 minutes when the decay period after switch on had ended.
The equivalent air change rate due to the Far-UVC was calculated from the steady state concentrations before (C) and after (C uv ) the lamps were switched on based on Here N is the ventilation rate of the room (ACH) and N uv is the equivalent air change rate (eACH) due to the Far-UVC.

Data Availability
All data generated or analysed during this study are included in this published article (and its Supplementary Information les).  Percentage of viable airborne S. aureus remaining plotted on a linear y-axis for two of the exposure scenarios motivated by current exposure limits (5 lamps "Medium") and proposed increased ACGIH Threshold Limit Values (5 lamps "High"). Note that the pathogen was continuously released into the room throughout the experiment with a mechanical ventilation rate of 3 air changes per hour.

Figure 4
Inactivation of aerosolized human coronaviruses HCoV OC43 and HCoV 229E and H1N1 in uenza virus at relevant low far-UVC doses, compared with aerosolized S. aureus. Measurement taken at the Columbia University laboratory-based aerosolized pathogen irradiation system. HCoV OC43, HCoV 229E and H1N1 in uenza published previously.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.