Inactivation of airborne Enterococcus faecalis and infectious bursal disease virus using a pilot-scale ultraviolet photocatalytic oxidation scrubber

Abstract

High microbial concentrations and emissions associated with livestock houses raise health and environmental concerns. A pilot-scale ultraviolet photocatalytic oxidation (UV-PCO) scrubber was tested for its efficacy to inactivate aerosolized Enterococcus faecalis and infectious bursal disease virus (IBDV). Microbial reduction was determined by the difference in microbial concentrations measured in the upstream and downstream isolators that were connected to the two ends of the UV-PCO scrubber. Two UV irradiance levels were tested by using one or two UV lamps. The theoretical average UV irradiances were 6,595 µ W cm⁻² with one UV lamp and 12,799 µ W cm⁻² with two UV lamps. At the tested ventilation rate (70 m³ hr⁻¹), the contact time was 1 sec. Reduction rate and other two indexes (k-value and Z-value) that normalized UV radiation were calculated to describe the extent of microbial inactivation. The UV-PCO scrubber eliminated >99.7% of airborne E. faecalis from the incoming airstream under one UV lamp irradiance, and the reduction was further increased by 0.2–0.3% when the second UV lamp was added. The reduction rate for airborne IBDV was 72.4% with one UV lamp. The calculated k-values were 0.501–0.594 cm² mJ⁻¹ for airborne E. faecalis and 0.217 cm² mJ⁻¹ for IBDV. The Z-value of airborne E. faecalis to UV irradiance was 9.3 (±1.6) × 10⁻⁴ cm² µ W⁻¹ sec⁻¹. The results indicate that a UV-PCO scrubber can serve as an effective and efficient technology for inactivating airborne bacteria and virus. Scaling up of the pilot-scale scrubber for field use will require considerations such as design air treatment capacity, UV irradiance level, contact time, dust concentration, susceptibility of target microorganisms, and expected reduction rate.

Implications:

This work demonstrated that a UV-PCO scrubber can be used to inactivate animal-associated airborne microorganisms, thus reducing microbial emissions from livestock houses and minimizing the biological impact to ambient environment. The microbial reduction efficiency by the UV-PCO scrubber varied depended on the level of UV irradiation and the target microbial species. The tested viral species (infectious bursal disease virus) was more resistant to the UV-PCO scrubber as compared to its counterpart bacterial species (E. faecalis).

Introduction

Emission of airborne microorganisms from concentrated animal feeding operations (CAFOs) has increasingly received concern with respect to its health and epidemiological impacts (Seedorf et al., 1998). The airborne microorganisms emitted and their components (e.g., endotoxins and lipoteichoic acid) may not only trigger inflammatory respiratory dysfunction in humans and negative immune responsiveness in animals but some pathogenic species are also suspected to be aerially transmittable between farms, which is especially concerning in countries with high-density livestock farming (e.g., the U.K. and The Netherlands). Some infectious viruses have been detected at the exhaust outlet of animal houses (Zhao et al., 2013) and other highly infectious diseases have been proven to be spread through airborne transmission (Gloster et al., 2003). Effective technologies are needed to reduce microbial emissions and prevent pathogen transmission, thereby safeguarding animal and human health in the vicinity.

A variety of air scrubbers have been designed to reduce emissions of air pollutants from CAFOs. While gases (e.g., ammonia) and particulate matter are constantly reduced, their efficacies in reducing microbial emissions vary largely depending on the scrubber type. Aarnink et al. (2011) reported that a sulfuric acid scrubber eliminated 70% of total airborne bacteria; however, higher bacterial counts were found in the outlet air compared to the inlet air of a bioscrubber, which is consistent with the study by Seedorf and Hartung (1999). Considering the high bacterial levels associated with CAFOs, 5–7 log₁₀ colony-forming units (CFU) m⁻³ in general (Zhao et al., 2011a; Wang-Li et al., 2013), even a 70% reduction might not be completely satisfactory and “not significantly prevent emissions of microorganisms” (Aarnink et al., 2011, p. 1921).

Recently, a novel pilot-scale ultraviolet (UV) photocatalytic oxidation (PCO) scrubber has been developed for evaluation of its antimicrobial efficacy. Photocatalytic oxidation is a technology that produces hydroxyl radicals (·OH) and superoxide (O₂ ⁻¹) from reactions between water/oxygen and a semiconductor (normally titanium oxide) under UV radiation. Both hydroxyl radicals and superoxide have antimicrobial capacity and have been proven to be capable of inactivating airborne Staphylococcus aureus (Vohra et al., 2006), Escherichia coli (Vohra et al., 2006; Pal et al., 2008), Aspergillus niger (Vohra et al., 2006), MS2 phage (Vohra et al., 2006), etc. Ultraviolet, especially that at 254 nm wavelength, inactivates microorganisms by inhibiting DNA replication or inducing the formation of thymine dimers (Prescott et al., 2005). It has been widely applied to disinfection in water and food industries for decades (Hijnen et al., 2006; Begum et al., 2009). For air disinfection, efforts have been directed to investigating the antimicrobial effect of UV lamps in the confined human environment or flow chambers (Miller and Macher, 2000; Peccia et al., 2001; Xu et al., 2003), showing that a UV lamp effectively inactivated airborne microorganisms including Mycobacterium parafortuitum (Xu et al., 2003), Mycobacterium bovis (Xu et al., 2003), Mycobacterium vaccae (Griffiths et al., 2005), Micrococcus luteus (Miller and Macher, 2000), E. coli (Miller and Macher, 2000), Bacillus subtilis (Miller and Macher, 2000; Xu et al., 2003), MS phage (Griffiths et al., 2005), etc. These environments do not require high UV irradiation for microbial inactivation because the air exchange rate is low, thus giving a long contact time between the microorganisms and UV irradiation. A high volume of air is associated with summertime ventilation of CAFOs, which would require high UV irradiation for rapid-onset microbial inactivation. However, little is known concerning the performance of UV scrubbers, as incorporated with PCO technology, in exhaust air treatment for CAFOs.

The objective was to evaluate the efficacy of a pilot-scale UV-PCO scrubber, with high UV irradiance level, on reduction of aerosolized Enterococcus faecalis and infectious bursal disease virus (IBDV) at a short contact time (1 sec). E. faecalis, a gram-positive and commensal bacterium, is a major cause of human enterococci infections and is suggested to transmit from animal reservoirs (Poulsen et al., 2012). Infectious bursal disease virus is a nonenveloped double-stranded RNA virus that causes chicken immunosuppression and high mortality in birds (Hirai et al., 1979; Sharma et al., 2000). This viral species was recently recovered in the exhaust air of an experimental broiler room (Zhao et al., 2013) and is speculated to be airborne transmissible between farms (Edgar and Cho, 1976; Lasher and Shane, 1994; Sanchez et al., 2005).

Materials and Methods

Experimental scrubber

The stainless steel UV-PCO consisted of a rectangular contact cell with two funnels mounted on both sides (Figure 1). The contact cell measured 0.20 × 0.32 × 0.32 m (L × W × H). Three UV lamps (TUV, PPL 35W HO, Phillips, Amsterdam, The Netherlands) were installed at different heights in the middle of one side wall of the contact cell. Each lamp had UV wattage of 11 W. The gap was 0.06 m between lamps and 0.10 m between the lamp and ceiling/floor. Three UV irradiance levels could be created in the contact cell; that is, zero irradiance (all lamps turned off), low irradiance (middle lamp turned on), and high irradiance (two side lamps turned on). The values of the UV irradiance levels were calculated using the method described in the Calculations subsection. Two pleated PCO filters were installed in parallel at a distance of 20 cm on both sides of the UV lamps. The funnels were used to connect the scrubber with the testing system.

Figure 1. Longitudinal section and cross-sectional views of the UV-PCO scrubber (dimensions in centimeters).

Testing system

The UV-PCO scrubber was tested in an controlled environment room at Animal Health Service (GD, Deventer, The Netherlands). An upstream and a downstream isolator were connected to the inlet and outlet of the UV-PCO scrubber, respectively (Figure 2). The two isolators were identical (1.94 × 0.75 × 0.95 m; Beyer and Eggelaar, Utrecht, The Netherlands). High-efficiency particulate air filters were installed at the inlet of the upstream isolator and the outlet of the downstream isolator in order to filter incoming air and to ensure bio-safety of the outgoing air. Beneath each isolator was a water lock through which air samples were taken out after sampling. Negative-pressure ventilation was provided to the testing system by programing the control panel on top of the isolators. The air stream was directed from the upstream isolator to the scrubber and then to the downstream isolator. In this experiment, a ventilation rate of 70 m³ hr⁻¹ (maximum ventilation capacity of the system) was used in all tests. At this ventilation rate, the aerosolized microorganisms passed through the contact cell of the UV-PCO scrubber at a speed of 0.19 m sec⁻¹, resulting in a 1 sec exposure/contact time between the microbes and UV and PCO components.

Figure 2. Schematic drawing of the testing system.

Microbial suspensions for aerosolization

E. faecalis originating from spontaneous cases of amyloid arthropathy in chickens was cultured as described by Zhao et al. (2011b) and was suspended in buffered peptone water (BPW) at a concentration of approximately 9 log CFU mL⁻¹. A suspension with a lower concentration of E. faecalis was also prepared by serially diluting (10⁻²) the above-mentioned suspension, yielding approximately 7 log CFU mL⁻¹. Virus from 24 vials, each containing 7.4 log 50% egg infectious dose (EID₅₀) of IBDV (Merial B.V., Velserbroek, The Netherlands), was suspended in 100 mL BPW, yielding a viral concentration of 6.7 log EID₅₀ mL⁻¹.

Aerosolization

A Walther Pilot spray gun with a nozzle of 0.5 mm diameter (Walther Spritz- und Lackiersysteme GmbH, Wuppertal, Germany) was used for aerosolization of the microbial suspensions. It was connected to an air compressor (Mecha Concorde type 7SAX, 1001, SACIM, Verona, Italy), supplying compressed air at 2 bar for aerosolization. The aerosol spectrum of the spray head was previously characterized by laser diffraction (Mastersizer-S long bed; Malvern Instruments, Malvern, UK). The volume median diameter of the aerosols generated by the spray gun was 10 μm.

To test the efficacy of UV-PCO scrubber on reducing airborne E. faecalis and IBDV, a total volume of 22 mL of either bacterial or viral suspension was aerosolized near the air inlet of the upstream isolator and the concentrations of airborne microorganisms were determined at the outlet of the upstream isolator and at the inlet of the downstream isolator (Figure 2). Bacterial suspensions with high (9 log CFU mL⁻¹) and low (7 log CFU mL⁻¹) concentrations of bacteria were aerosolized with either one or two UV lamps in operation. Each treatment combination was repeated four times. Because physical loss of airborne microorganisms could occur during aerial transportation in the testing system, three control treatments were also performed with the scrubber connected but the UV lamps turned off or with the scrubber disconnected (i.e., the two isolators were directly connected with tubes). Suspensions of IBDV were aerosolized in triplicate with a control (no scrubber) and a treatment (one UV lamp turned on). A summary of the treatments is listed in Table 1.

Table 1. Summary of treatments in this study

Suspension	Microbial concentration in suspension	Treatment	Repetition
E. faecalis	9 log CFU mL^−1	Control (no scrubber)	3
	9 log CFU mL^−1	Control (lamp off)	3
	9 log CFU mL^−1	One lamp on	4
	9 log CFU mL^−1	Two lamps on	4
	7 log CFU mL^−1	One lamp on	4
	7 log CFU mL^−1	Two lamps on	4
IBDV	6.7 log EID50 mL^−1	Control (no scrubber)	3
	6.7 log EID50 mL^−1	One lamp on	3

Each aerosolization event lasted for 5 min. During aerosolization, continuous ventilation was supplied to the entire system (isolators and scrubber). A fan was used in each isolator to facilitate uniform distribution of the airborne microorganisms. Before aerosolization, the testing system was ventilated for at least half an hour to remove the remaining airborne microorganisms. When UV treatment was used, the UV lamps warmed up for 15 min before aerosolization. Temperature and relative humidity (RH) of the air in the controlled environment room (supply air) and in both isolators were continuously measured at a 5-sec intervals using Rotronic sensors (HygroClip2, ACIN Instruments, The Netherlands).

Air sampling and sample processing

Two Airport MD-8 air samplers (Sartorius, Göttingen, Germany) mounted with gelatin filters (pore size = 3 μm) were used for sampling airborne microorganisms in the upstream and downstream isolators. Air sampling in both isolators started simultaneously 3 min after aerosolization onset and lasted for 2 min. The sampling air flow rate was 30 L min⁻¹. After sampling, microbe-laden gelatin filters were carefully wrapped in sterile waterproof boxes and taken out of the isolator through the water locks. Liquid samples were made by dissolving each gelatin filter in 50 mL BPW (37°C) for further analysis of viable microbial counts.

Microbial analysis

Liquid samples of E. faecalis was serially diluted (in 10-fold steps) in physiological salt solution (bioTRADING Benelux B.V., Mijdrecht, The Netherlands). One tenth of a milliliter of each dilution was plated on a petri dish with sheep blood agar. The petri dishes were then incubated at 37°C for 48 hr. The number of colonies on plates with 30–300 colonies were counted and the concentration of the bacteria in the 50-mL liquid samples was calculated, following international standards (International Organization for Standardization, 1985).

Concentrations of IBDV in the liquid samples were determined with an egg embryonic death test. The samples were serially diluted in 10-fold steps. A volume of 0.5 mL of each serial dilution was injected into the allantoic cavity of five 9-day-old specific pathogen-free embryonated eggs. The inoculated eggs were incubated at 37°C for 7 days, and the viral concentration was calculated based on death of the embryos and specific abnormalities of the living embryos using the Spearman-Karber method (Spearman, 1908).

Based on the microbial concentration in liquid samples, the air concentrations of airborne E. faecalis and IBDV were calculated by dividing the total microbial count in liquid samples by the air volume filtrated by gelatin filters, expressed in log CFU m⁻³ or log EID₅₀ m⁻³.

Calculations

The theoretical UV irradiance in the contact cell with the operation of one or two lamps was estimated according to Bolton and Linden (2003): the UV lamp was divided in 20 equal sections (1 cm long each), and the scrubber's contact cell was divided into 2 × 2 cm grids. The total irradiance at a grid intersection was calculated by summing the irradiances from all lamp sections. Assuming laminar air flow and constant travel speed of the airborne microorganisms in the contact cell, average irradiance of a vertical gridline, a horizontal gridline, and overall average irradiance can be calculated as the mean of the total irradiances of the grid intersections involved.

The concept of inactivation in this study refers to the loss of microbial culturability after being treated by the UV-PCO scrubber (i.e., includes both culturable-to-dead and culturable-to–viable but nonculturable transitions). The extent of inactivation of airborne microorganisms by the UV scrubber was indicated by reduction rate (R, %) using eq 1:

1

where C _up is the concentration of airborne microorganisms in the upstream isolator before the scrubber (CFU m⁻³ or EID₅₀ m⁻³); C _down is the concentration of airborne microorganisms in the downstream isolator after the scrubber (CFU m⁻³ or EID₅₀ m⁻³); α is the physical loss of airborne microorganisms during transport in the testing system; that is, the difference in microbial concentrations between upstream and downstream isolators without a scrubber connected (%).

Response of microbial survival to UV radiation was also expressed by the k-value and Z-value. Equation 2 was used to calculate the k-value:

2

where k is the k-value (cm² mJ⁻¹), R is the reduction rate, and Fluence (mJ cm⁻²) is the product of UV irradiance (mW cm⁻²) and the contact time (1 sec in this study).

The Z-value reflects the dose–response of microorganisms to germicidal treatments. The Z-value of airborne E. faecalis is the slope by linear regression of the inactivation rate (sec⁻¹) against UV irradiance (μW cm⁻²) to which the bacteria were exposed (Xu et al., 2003). The inactivation rate (sec⁻¹) is the natural logarithm slope derived by linear regression of the normalized airborne microorganism concentration against contact time. Because only two levels of contact time and UV irradiance were used, the Z-value in this study was simply calculated using eq 3:

3

where Z is the Z-value (cm² μW⁻¹ s⁻¹), C ² _down and C ² _up are the E. faecalis concentrations (CFU m⁻³) at the downstream and upstream isolators when two UV lamps were on, C ¹ _down and C ¹ _up are the E. faecalis concentrations (CFU m⁻³) at the downstream and upstream isolators when one UV lamp was on, and Irradiance is the UV irradiance (μW cm⁻²). C ² _down and C ¹ _down were both corrected for physical loss.

Statistical analysis

Statistical analysis was performed using analysis of variance (SAS 9.2, SAS Institute Inc., Cary, NC) to examine the reduction rate of airborne microorganisms and the k-values as affected by the microbial species and initial bacterial concentrations.

Results

UV irradiance in the contact cell of the scrubber

The UV irradiances of vertical and horizontal gridlines in the contact cell with one or two UV lamps on are shown in Figures 3 and 4, respectively. It can be seen that the distribution of UV irradiation in the contact cell was not uniform. Much higher average irradiance was present for the gridlines crossing the UV lamp(s). The overall average UV irradiation in the contact cell was 6,595 μW cm⁻² for one lamp and 12,799 μW cm⁻² for two lamps.

Figure 3. (a) Average UV radiations of vertical gridlines in the scrubber's contact cell with one middle lamp on. (b) Average UV radiations of horizontal gridlines in the scrubber's contact cell with one middle lamp on. Example gridlines (in red) are shown in the upright subfigures. (Color figure available online.)

Figure 4. (a) Average UV radiations of vertical gridlines in the scrubber's contact cell with two side lamps on. (b) Average UV radiations of horizontal gridlines in the scrubber's contact cell with two side lamps on. Example gridlines (in red) are shown in the upright subfigures. (Color figure available online.)

Temperature and humidity

Table 2 shows the temperature and relative humidity of the supply air and air from upstream and downstream isolators during sampling (the last 2 min of aerosolization). The temperature and relative humidity of the supply air were 20.5 ± 0.2°C and 47 ± 9% and remained constant before and after sampling. Before sampling, temperatures in the upstream and downstream isolators were 20.6 ± 0.2°C and 21.6 ± 0.4°C, respectively. After 2-min sampling, the temperature in the upstream isolator decreased by 0.4°C as a result of evaporative cooling of the liquid spray. The temperature in the downstream isolator did not change much before and after sampling. Relative humidity in both the upstream isolator and downstream isolator increased after aerosolization. During aerosolization, the average temperatures were 20.4 and 21.6°C in the upstream and downstream isolators, and relative humidity was 67 and 55% in the upstream and downstream isolators, respectively.

Table 2. Temperature and relative humidity (RH) of supply air and air of the upstream and downstream isolators before and after sampling (n = 5). ^a Mean ± standard deviation

	Supply air			Air of upstream isolator			Air of downstream isolator
	Before	After	Average	Before	After	Average	Before	After	Average
Temperature (°C)	20.5 ± 0.2	20.5 ± 0.2	20.5 ± 0.1	20.6 ± 0.2	20.2 ± 0.3	20.4 ± 0.4	21.6 ± 0.4	21.7 ± 0.4	21.6 ± 0.4
RH (%)	47 ± 9	47 ± 9	47 ± 8	62 ± 8	71 ± 12	67 ± 11	51 ± 8	59 ± 9	55 ± 10
Note: ^aDue to malfunction of the sensors, temperature and relative were not measured in every spraying event.

Efficacy of UV-PCO scrubber on microbial reduction

Table 3 shows the physical loss of airborne microorganisms in the test systems with no scrubber connected or with the scrubber connected but the UV lamp off. In both cases, the concentrations of airborne microorganisms were reduced by about 0.8 log in the downstream isolator relative to the upstream isolator, reflecting a physical loss of 82.67 to 85.93%. The physical losses for E. faecalis and for IBDV were not significantly different (P = 0.68). The physical loss obtained without the scrubber was used to correct the reduction rate of airborne microorganisms by the UV-PCO scrubber.

Table 3. Physical loss (α) of airborne microorganisms in the testing system. ^a Mean ± standard deviation

Microbial suspension		Airborne concentration in upstream isolator	Airborne concentration in downstream isolator	Reduction (%) ^b
E. faecalis (9 log CFU mL^−1)	No scrubber	9.47 ± 0.01	8.71 ± 0.03	82.67a ± 1.23
	Lamps off	9.52 ± 0.04	8.66 ± 0.04	85.93a ± 2.79
IBDV (6.7 log EID50 mL^−1)	No scrubber	6.11 ± 0.36	5.31 ± 0.36	84.17a ± 0.50
Notes: ^aUnit of airborne concentration is log CFU m^−3 for E. faecalis and log EID50 m^−3 for IBDV.
^bReductions not followed by the same letter were significantly different (P < 0.05).

The bacterial concentration was reduced remarkably for all treatments with UV lamp(s) (Table 4). The scrubber inactivated 3.5 to 5.3 log airborne E. faecalis with one UV lamp operating and inactivated 7.5 log E. faecalis with two lamps operating. The reduction efficiency for the treatments was higher than 99.7%. Two-factorial analysis of variance showed that the bacterial concentration in suspensions (high or low) did not affect the reduction efficiency (P = 0.63), but the number of UV lamps affected the reduction efficiency (P = 0.03). The calculated k-values for E. faecalis were 0.501–0.594 cm² mJ⁻¹ and did not differ between UV treatments (P = 0.27) or between high and low bacterial concentrations in the suspensions (P = 0.84). Airborne IBDV was inactivated by about 1.4 log with one UV lamp on, representing 72.35 ± 19.91% reduction efficiency by the scrubber. The viral reduction was significantly lower than the reduction for E. faecalis under the same UV irradiance level (P < 0.01). The calculated k-value for IBDV was 0.217 ± 0.046 cm² mJ⁻¹, which was significantly lower than that for E. faecalis.

Table 4. Inactivation of culturability of airborne microorganisms by the UV-PCO scrubber. ^a Mean ± standard deviation

Microbial suspension ^b		Airborne concentration in upstream isolator	Airborne concentration in downstream isolator	Reduction (R, %) ^c	k-Value (cm^2 mJ^−1) ^c
E. faecalis (9 log CFU mL^−1)	One lamp	9.64 ± 0.07	6.09 ± 0.29	99.81a ± 0.09	0.538a ± 0.044
	Two lamps	9.72 ± 0.22	2.34 ± 0.71	99.99b ± 0.01	0.581a ± 0.096
E. faecalis (7 log CFU mL^−1)	One lamp	7.61 ± 0.14	2.34 ± 0.71	99.72ab ± 0.36	0.501a ± 0.051
	Two lamps	7.54 ± 0.11	0.00 ± 0.00	100b ± 0.00	0.594a ± 0.008
IBDV (6.7 log EID50 mL^−1)	One lamp	6.25 ± 0.31	4.82 ± 0.00	72.35c ± 19.91	0.217b ± 0.046
Notes: ^aUnit of airborne concentration is log CFU m^−3 for E. faecalis and log EID50 m^−3 for IBDV.
^bThree replications for IBDV and for replications for E. faecalis.
^cMeans in a column not followed by the same letter were significantly different (P < 0.05).

Because the bacterial concentration in the suspension was an insignificant factor, airborne concentrations of E. faecalis were pooled for calculation of inactivation rates. The inactivation rate was 5.6 ± 1.1 sec⁻¹ with one UV lamp on and 11.0 ± 1.1 s⁻¹ with two UV lamps on (Figure 5). The Z-value of airborne E. faecalis against UV irradiance (and the induced PCO germicidal components) was 9.3 (±1.6) × 10⁻⁴ cm² μW⁻¹ sec⁻¹ (Figure 5). The inactivation rate and Z-value for airborne IBDV were not calculated because only one UV irradiance level was tested.

Figure 5. (a) Inactivation of airborne E. faecalis at two UV exposure levels and (b) Z-value. The inactivation rates are the slopes in (a), which have been corrected for physical loss, α.

Discussion

A pilot-scale UV-PCO scrubber was tested for its efficacy in reducing culturability of airborne microorganisms in short contact time, and its UV irradiance was estimated. It was found that the UV irradiance in the scrubber's contact cell was not evenly distributed, with the highest UV irradiance occurring near the UV lamps. The average UV irradiances of gridlines crossing the UV lamp were close to 40,000 μW cm⁻² (Figures 3 and 4), compared to UV irradiances of gridlines close to the scrubber wall being <10,000 μW cm⁻². It should be noted that the UV irradiances in this study were the theoretical maximum value estimated by assuming perfect air mixing and laminar flow.

The physical loss of E. faecalis during aerial transport in the testing system when the scrubber was installed (85.93%) was slightly higher than that with no scrubber (82.67%), though this was not significant. This indicates a 23.17% of physical filtration of the airborne microorganisms by the UV-PCO scrubber.

Venieri et al. (2011) reported that the UV-PCO technique effectively inactivated E. faecalis in wastewater. Significant inactivation by UV-PCO was also noticed for airborne E. faecalis in this study. The inactivation of the aerosolized E. faecalis was not affected by the initial bacterial concentration in the suspension (P = 0.63). The inactivation of the aerosolized E. faecalis significantly increased when a higher UV irradiance level was applied (two UV lamps vs. one UV lamp). Specifically, the scrubber with one UV lamp on (theoretical maximum irradiance = 6,595 μW cm⁻²) inactivated >99.72% of airborne E. faecalis within 1 sec contact time, whereas two UV lamps (theoretical maximum irradiance = 12,799 μW cm⁻²) achieved an even higher reduction of >99.99%. Positive relationships between inactivation and UV irradiation were also reported for many other microbial species (Peccia et al., 2001; Xu et al., 2003). The reduction of bacteria was not linearly related to the amount of UV power used (μW cm⁻²). The running costs would be doubled if two UV lamps were used; however, the reduction in the bacteria only increased marginally (0.2–0.3%). This finding is consistent with the findings of Xu et al. (2005), who reported that an increase in the UV strength above 5 μW cm⁻² did not give a proportional increase in the inactivation rate. However, for a condition or situation where (almost) complete elimination of bacteria is required, installing enough UV power would be crucial.

Several inactivation indexes, including the k-value and Z-value, have been used to characterize the response of microorganisms to UV. Hijnen et al. (2006) reviewed the k-values for a variety of bacterial species that were determined with collimated beam tests. Their k-values were within a range of 0.059–1.341 cm² mJ⁻¹. The calculated k-value for airborne E. faecalis (0.501–0.594 cm² mJ⁻¹) reported in the current study was in the middle of Hijnen's range. The Z-values were reported as 23–55 × 10⁻⁴ cm² μW⁻¹ s⁻¹ for Mycobacterium tuberculosis (Riley et al., 1976), 23–39 × 10⁻⁴ cm² μW⁻¹ s⁻¹ for M. bovis BCG (Riley et al., 1976), 12 × 10⁻⁴ cm² μW⁻¹ s⁻¹ for M. parafortuitum (Xu et al., 2003), and 6.3–6.6 × 10⁻⁴ cm² μW⁻¹ s⁻¹ for B. subtilis (Peccia et al., 2001). We reported, for the first time, a Z-value of airborne E. faecalis at a temperature of 20.2–21.7°C and 51–71% RH of 9.3 (±1.6) × 10⁻⁴ cm² μW⁻¹ s⁻¹, which is within the range of the previously reported Z-values for bacteria but on the low side.

The IBDV is a relatively resistant viral species to environmental stress. The virus can remain viable for up to 60 days in poultry litter (Vindevogel et al., 1976) and may potentially be spread between farms through airborne transmission (Zhao et al., 2013). With one UV lamp on, the airborne IBDV was inactivated by 1.5 orders of magnitude, yielding a 72.35% reduction. This reduction rate might not be high enough to safeguard animal safety and was significantly lower than that for airborne E. faecalis at the same UV irradiance level. Our results are consistent with those reported by Harris et al. (1987), who revealed that viruses are generally more resistant to UV than bacteria. The UV inactivation of viruses is mainly due to uracil dimerization in viral RNA, whereas the inactivation of bacteria is the result of thymine dimerization of DNA. Uracil dimerization is less readily induced than the thymine dimerization, which may partially explain the greater resistance of viruses to UV inactivation than bacteria (Harris et al., 1987). The k-value reported for airborne IBDV in the current study was within the range for other previously reported viral species—that is, 0.018–0.293 cm² mJ⁻¹—including poliovirus, adenovirus, rotavirus, calicivirus, hepatitis A and coxsackie virus (Hijnen et al., 2006).

Considerations for scaling up of the UV-PCO scrubber for field use include design air treatment capacity, UV irradiance level, contact time, dust concentrations, susceptibility of target microorganisms, expected reduction rate, etc. If scaling up is planned for use in a laying henhouse, the following should be taken into account. The ventilation rate is generally 0.5 m³ hr⁻¹ hen⁻¹ in winter time and 10 m³ hr⁻¹ hen⁻¹ in summer. Adopting the bacterial reduction reported in this study, a scrubber with one UV lamp (11 W UV-C) is suitable for 140 hens in wintertime and 7 hens in summertime in order to achieve >99.7% bacterial reduction. For nonpathogenic microorganisms, such a high reduction rate may not be necessary. Therefore, one UV lamp can treat more exhaust air. In situations where 100% reduction of pathogens is imperative—for example, when it is used for cleaning incoming air in a specific pathogen-free unit—more UV power per air volume will be necessary. Enlarging the size of the scrubber should be considered to increase the residence time of the air inside the scrubber. However, the enlargement would make it more difficult to provide uniform distribution of UV radiation in the scrubber. Possible solutions would be to use more but lower power UV lamps or using a properly designed mirror to reflect UV light.

Conclusion

A pilot-scale UV-PCO air scrubber was evaluated for its efficiency in reducing aerosolized E. faecalis and IBDV supplied with a high ventilation rate and thus low contact time (1 sec). Based on the experimental data, it can be concluded that this UV-PCO scrubber is very effective in removing E. faecalis from the air. When one UV lamp was used (theoretical maximum UV irradiance = 6,595 μW cm⁻²), more than 99.7% airborne E. faecalis could be eliminated from the incoming airstream. This reduction was further increased by 0.2–0.3% by adding one more UV lamp (theoretical maximum UV irradiance = 12,699 μW cm⁻²). The scrubber inactivated 72.4% of airborne IBDV, which was significantly less than the efficacy for E. faecalis removal. The calculated k-values were 0.501–0.594 cm² mJ⁻¹ for airborne E. faecalis and 0.217 cm² mJ⁻¹ for IBDV. The Z-value of airborne E. faecalis for UV irradiance was 9.3 (±1.6) × 10⁻⁴ cm² μW⁻¹ sec⁻¹. Scaling up the UV-PCO scrubber for field use will require consideration of factors such as design air treatment capacity, UV irradiance level, contact time, dust concentrations, susceptibility of target microorganism, and expected reduction rate.

Acknowledgment

We thank Machiel Esmann, Tao Nan, and the staff at the microbiological labs at the Animal Health Service (GD) in Deventer for their assistance with the experiment.