Evaluation of UVLED disinfection for biofouling control during distribution of wastewater effluent

Water reuse is a promising water management strategy to alleviate water scarcity currently affecting every continent (UN-Water 2018). Water sources for potential reuse include stormwater, industry process water, and municipal wastewater (US EPA 2019). Uses for reuse water include environmental restoration, groundwater recharge, potable reuse, and agricultural irrigation. Wastewater reuse for agriculture is particularly promising; for instance, Israel recycles 87% of its wastewater for irrigation of both food and non-food crops (Marin et al. 2017). However, irrigating with wastewater effluent comes with challenges including biofouling of drip lines and other infrastructure meant to transport water.

Biofouling, or biofilm accumulation, is the unwanted deposition of microorganisms and sticky extracellular polymeric substances (EPS) on surfaces (Flemming 2002). Biofilms are established in five stages: (1) attachment of planktonic cells to a surface, (2) attached bacteria divide and EPS is excreted, (3) the EPS cell mixture expands, (4) the biofilm matures, and (5) the mature biofilm releases planktonic bacteria back into the environment (Flemming et al. 2007). Biofilms play a major role in many water reclamation and reuse strategies like membrane bioreactors and filtration systems used in specific reuse applications (Bishop 2007). The higher concentration of nutrients, organic matter, and microorganisms in wastewater effluent can aggravate the biofouling of water distribution systems, like drip irrigation lines (Tarchitzky et al. 2013).

Strategies for biofouling control during the distribution of effluent include filtration, flushing, and chlorination. Filtration is an important factor in preventing biofilm formation as it removes biofouling microorganisms, organic particulates, and some nutrients that facilitate their growth. Most manufacturers of drip irrigation systems recommend filtration for particulates; however, in the case of effluent, filter clogging is a big concern in addition to the clogging of the drip irrigation lines (Ravina et al. 1997). Regardless, filtration is still an important step during irrigation with reuse water to prevent immediate clogging of emitters by large particles (Adin & Sacks 1991).

Flushing is the process of increasing the water velocity in an irrigation system so the flow hydraulic shear force rapidly sheds attached biofilms while also slowing down the clogging frequently caused by the shredded biofilm falling inside the emitter (Li et al. 2015). Flushing studies have identified a method to effectively slow down emitter clogging in a reclaimed water drip irrigation system, and the longer period it was applied, the better controlling was observed (Li et al. 2015, 2019). However, increasing water velocity causes an immediate increase in bacteria numbers as the increase in shear stress resuspends biofilms (Lehtola et al. 2006). In one study examining the effects of flushing during irrigation with reclaimed water, lateral flushing failed to completely solve the emitter clogging problem and pairing flushing with additional measures was recommended (Li et al. 2015).

Chlorination is considered the most effective method of controlling biofouling during irrigation, but it has drawbacks (Song et al. 2017). Aside from being hazardous to humans, one study demonstrated adverse soil health effects when chlorine was applied in high concentrations during short-term use which is the recommended treatment for irrigation lines (Song et al. 2019). When chlorine and chloramine react with organic matter, like that found in wastewater effluent, it forms hazardous disinfection by-products (DBPs) (Doederer et al. 2014). Although the impact of DBPs on soil and crop health is not well studied, their ability to persist in aquatic environments has been reported (Rostad 2002).

UV disinfection is a common and effective treatment technology for drinking water, wastewater, and reclaimed water globally (Song et al. 2016; Nguyen et al. 2019). Recent advancements in UV light emitting diodes (UVLEDs) have allowed for compact, electrically efficient, and customizable point-of-use UV disinfection options. This includes the integration of UVLEDs into devices like toothbrush holders, coffee makers, and water coolers (Linden et al. 2019). UVLEDs have demonstrated an ability to treat wastewater effluent to US EPA guidelines for irrigation of processed food crops and non-food crops (Nguyen et al. 2019). LEDs are particularly well suited for the intermittent flow applications of agricultural irrigation due to their small size, no required warmup time, and ability to turn on/off without negatively impacting the lifetime and performance of the device (Song et al. 2016; Chen et al. 2017). Despite this, the impact of UVLED disinfection on biofouling during irrigation with wastewater effluent has not been studied.

The goal of this study was to examine a commercially available UVLED flow-through device for biofouling control on drip irrigation line material during the distribution of wastewater effluent. The effects of the UVLED device on biofilm growth in a bioreactor fed by effluent were tested through direct quantification assays (total coliforms) and indirect quantification assays (crystal violet (CV), adenosine triphosphate (ATP)). Before implementing the UVLED device for biofouling mitigation studies, a series of challenge tests were performed to establish a relationship between operating flow rate and fluence for varying water qualities in the UV LED flow-through devices. Qβ bacteriophage was chosen as the challenge organism because of its ability to survive high fluence levels, providing countable concentrations for the low-flow rate/high-fluence testing conditions (Malayeri et al. 2016). The log reduction values (LRV) of Qβ bacteriophage were determined in three water matrices described later with varying UV transmittances (UVT) at 254 nm (93, 80, and 70%).

Qβ bacteriophage stock (ATCC 23631-B1) was propagated by GAP EnviroMicrobial Services Ltd (London, Ontario, Canada). An Escherichia coli K-12 (ATCC 23631) host culture was made in tryptic soy broth (TSB) with shaking at 35 °C. Once the E. coli reached early log-phase growth (OD600: 0.1–1.0), 1 mL of 1 × 10¹⁰ plaque forming units (PFU)/mL Qβ stock was added. After overnight incubation with shaking, the stock was centrifuged to remove cellular debris, leaving the phage in the supernatant. A concentration of approximately 1 × 10¹⁰ PFU/mL was achieved. Stocks were stored at 4 °C.

E. coli Famp (ATCC #700891) was used as the Qβ bacteriophage stock host for challenge testing. The Famp was separately streak-plated onto sterile nutrient broth (NB) (BD Difco™) agar plates and incubated for 20 h at 37 °C. One isolated colony was selectively removed with a sterile inoculation loop and suspended in 25 mL sterile NB for overnight incubation for 20 h at 37 °C with shaking (180 rpm). The overnight culture was then centrifuged at 5,000 rpm for 10 min. The supernatant was removed, and the pellet was resuspended in sterile NB. Twenty percent (by volume) of glycerol (100% v/v, Microlytic) was added to produce a final concentration of 9 × 10⁸ colony forming units (CFU)/mL E. coli Famp. Stocks were stored at −80 °C.

Three water matrices with varying UV transmittances (UVT_254nm) (93, 80, and 70%) were used for characterizing the UV devices for fluence delivered as a function of flow rate. Qβ bacteriophage stock was serially diluted into dechlorinated tap water to achieve a test water concentration of approximately 5 × 10⁶ PFU/mL, resulting in a 93–95% UVT at 254 nm. UV transmittance was adjusted for the 80 and 70% UVT_254nm waters using a mix of SuperHume (UAS of America, Lake Panasoffkee, FL, USA) and vanillin according to NSF/ANSI 55 standards. Dechlorination of test water was verified with a Hach DPD Free Chlorine colorimetric test.

US EPA 1602 method and Standard Methods 9224 B.3 were followed with minor adjustments including the addition of a streptomycin/ampicillin antibiotic to prevent microbial contamination (Bridgewater et al. 2017; Sholtes & Linden 2019). The bacterial host was E. coli Famp (ATCC #700891). Enumeration of Qβ was conducted using a combination of spread plate and spot plate methods and each sample was plated in triplicate. For each dilution in spread plating, 1 mL of sample was applied to soft-agar plates. For each dilution in spot plating, 20 μL of sample was serially diluted into 180 μL of phosphate-buffered saline (PBS) and five spots, 10 μL each, were added to the soft-agar plates with a multichannel pipette. After the sample was absorbed into the agar, plates were inverted and incubated for 20 h at 37 °C. Plaques were counted and concentrations were expressed in PFU/mL. For each experiment, the PBS dilution water, dechlorinated tap water, and pre-wash purified water were plated to test for contamination. Microbial reduction at each dose was expressed as log₁₀ (N₀/N_x). N₀ represents the initial concentration in PFU/mL in the control sample without UV exposure. N_x represents the concentration in PFU/mL after exposure to the various experimental flow rates.

Bench-scale collimated beam tests were performed with a PearlBeam UVLED System (AquiSense Technologies, Erlanger, Kentucky) and commercial low pressure (LP) lamps (15 W Sankyo Denki, Japan) to create reduction equivalent fluence (REF) equations for the flow-through UVLED devices. The UVLED collimated beam experiments followed Sholtes & Linden (2019) whereas the LP experiments followed Bolton & Linden (2003). The peak and weighted average wavelength of the UVLED over a 200–300 nm spectrum were 281.5 and 283.5 nm, respectively, measured using a Maya 2000 Pro spectrometer (Ocean Insight, Dunedin, FL). Incident irradiance (mW/cm²) was measured with a radiometer (IL2400, SED240, International Light, Newburyport, Massachusetts) at the calibration factors corresponding to the weighted average wavelength (283.5 nm) for the polychromatic UVLED lamps and 254 nm for the monochromatic LP lamps. The calibrated plane of the detector was placed directly under the UVLED at the same height as the water surface of the irradiated samples. Stirred suspensions of 10 mL (0.55 cm sample depth) were irradiated in 5.4 cm diameter dishes at a 15 cm distance from the UVLED and 41 cm from the LP lamps. The UV irradiance was measured before and after the exposures with a calibrated radiometer to ensure there was no decay in lamp output over the exposure periods and less than a 5% difference between final and initial irradiance as recommended in the US EPA UV Disinfection Guidance Manual (US EPA 2006). The petri factor was >0.99 for all UVLED exposures. UV fluence is a product of average irradiance and time; therefore, pre-determined fluences were achieved by manipulating exposure time once average irradiance was determined. All experiments were performed in duplicate, and each duplicate sample had a replicate measurement. Samples were exposed in random order over the target fluence levels.

The PearlAqua Micro^TM 9C UVLED flow-through units (AquiSense Technologies, Erlanger, Kentucky) were set up in accordance with manufacturer instructions. Greater than five reactor and effluent tube volumes of laboratory-grade deionized (DI) water were pumped through the entire system as a rinse. A sample of this DI rinse water was taken to test for contamination from previous testing in the system line prior to each experiment.

The test water was contained in a glass bottle, connected to integrated tubing, and a peristaltic pump for additional flow control. An untreated sample was taken from the glass bottle after sufficient mixing to quantify the starting concentration. To start experimentation, the peristaltic water pump (MasterFlex L/S 7518-62) was turned on. Once the water was flowing through the UVLED system, a sample was taken before the UVLED device was turned on to test for microbial decay due to non-UV factors. Upon changing the system flow rate, five times the reactor and effluent tube volume were allowed to flush the system before further experimentation. Samples from influent and effluent were taken in duplicate.

Once the relationship between fluence and LRV was established in the collimated beam experiments, the reduction equivalent fluences (REF_254nm and REF_280nm) for various flow rates were back calculated based on the Qβ LRVs in the flow-through units. Similar to Nguyen et al. (2019), a REF at 280 nm (as opposed to 254 nm) was chosen to investigate the wavelength-specific UV-induced stress experienced by microorganisms in the 280 nm UVLED flow-through device. However, a benchmark to 254 nm was included as it is helpful to compare to LP UV applications.

Secondary wastewater effluent was collected at the City of Boulder Water Resource Recovery Facility after biological nutrient removal and prior to UV disinfection. Water was analyzed according to Standard Methods for the Examination of Water and Wastewater, Section 1060B (Bridgewater et al. 2017). The average water quality parameters measured were pH (7.48), temperature (21.3 °C), dissolved organic carbon (DOC) (6.83 mg-C/L), and UVT_254nm (72.7%). The secondary effluent was used unfiltered.

Total coliforms were enumerated through membrane filtration on m-ENDO agar LES (NutriSelect™ Plus). Each fouled coupon was rinsed in 30 mL PBS for 20 s to remove planktonic cells then immersed in a 10 mL PBS solution. A vortex-sonication series was used to disaggregate biofilm from the coupons. The UltraSonic cleaner (Branson 8210, 50/60 Hz) was degassed for 5 min prior to testing. Samples were then sonicated for 30 s, vortexed for 30 s, with each step being repeated three times. Serial dilutions of 10 mL were passed through 0.45 μm membrane filters and the filters were placed on the surface of m-ENDO agar plates, filter side up. Plates were incubated at 35 ± 0.5 °C for 24 ± 2.0 h. Coliforms were identified by red colonies with a golden-green metallic sheen.

ATP bioluminescence is an indirect quantification method using ATP as a proxy marker which infers biofilm viability and biomass. ATP is a nucleoside triphosphate which acts as the primary energy source in all organisms, making it a strong indication of biofilm viability and biomass (Wilson et al. 2017). Coupons were rinsed in 30 mL PBS to remove extracellular ATP then immersed in 2 mL PBS. Coupons were then subjected to 30 s vortexing followed by 5 min of sonication. Samples were then left to incubate at 37 °C for 4 h to diminish any ATP increase due to stress induced by UV irradiation (Rauch et al. 2019). Intracellular ATP was quantified using LuminUltra's standard Quench-Gone Aqueous (QGA) Test (LuminUltra Technologies Ltd, New Brunswick, Canada).

Gram staining is a common and optimized indirect biofilm quantification method (Wilson et al. 2017). CV is a basic trianiline dye that permeates cell membranes in both Gram-positive and -negative cells regardless of inactivation or cell lysis. The dye leaves cells violet in color, and after a decolorization step with an ethanol solution, the biofilm can be quantified via spectroscopy. The CV assay was conducted according to Charlton (2008). A calibration curve was established between CV concentration and optical density, and a linear relationship was found. After treatment, coupons were immersed in a 2 mL 0.3% CV solution for 90 min followed by a PBS rinse and a subsequent DI rinse. Coupons were allowed to air dry, then transferred to a 2 mL 95% ethanol solution. Further dye solubilization was achieved by a 30-s vortex followed by 5 min of sonication. In total, 125 μL of sample were added to a 96-well plate and absorbance was measured at 540 nm. Five replicate measurements were taken per sample. Background absorbance was determined with clean coupons taken through the same stain and rinse steps.

Analysis of variance (ANOVA) full factorial experiments were performed to examine the statistical significance of experimental effects as included in the supplemental information. Model adequacy was verified by checking the three assumptions required for an ANOVA analysis: (1) equal variances, (2) residual normality, and (3) independent data.

The F-specific RNA bacteriophage Qβ was chosen as the biological indicator for flow-through UVLED device challenge testing because it shows first-order rate kinetics at high UV fluences (>40 mJ/cm²) and can capture the high inactivation rates produced under slow flow rates. It is a commonly used for flow-through UV device challenge testing and is the NSF/ANSI 55 challenge organism for UV device challenge testing (Jenny et al. 2014; Oguma et al. 2016a, 2016b; NSF/ANSI 55 2019). An LP 254 nm exposure was performed to ensure quality control of the microbial stocks and comparable results to other QB LP literature values. In addition to the LP 254 nm exposures, 280 nm UVLEDs were chosen to illuminate the response of microorganisms inactivated in the UVLED flow-through device also manufactured with identical 280 nm LED chips.

Using the model equation, at 1 L/min, there is a 0.961 Qβ LRV increase when UVT_254nm increases from 60 to 93%. At smaller flow rates of 0.1 L/min, this difference is larger, with a 2.647 Qβ LRV increase when UVT_254nm increases from 60 to 93%. The model results can be paired with the inactivation rate constants (slopes) found in Figure 2 to convert QB LRV to REF_280nm and REF_254nm. Therefore, the model can be used to contextualize the benefits of technologies that may impact water UV transmittance and the resulting benefit to REF. For example, at a flow rate of 0.5 L/min, a UVLED device paired with a pre-filter that increases UVT_254nm from 65 to 80% would result in an additional REF_254nm of roughly 10 mJ/cm².

The primary objective of this study was to test the UVLED device's ability to inactivate planktonic microorganisms in waters with UVT ranging from 70 to 93%. A target REF_280nm for water reuse at 40 mJ/cm², corresponding to a REF_254nm of 24 mJ/cm², was chosen to investigate biofouling control in the CDC bioreactor. 40 mJ/cm² REF_280nm, or 24 mJ/cm² REF_254nm, were obtained using a 0.16 L/min flow rate in the UVT_254nm 72.7 ± 2.5% water matrix. The following quantification methods describe the results from the flow-through investigations.

Biofilm quantification can be measured through direct and indirect methods. Direct biofilm quantification methods rely on direct observations for quantification of parameters, for example, the number of cells or total biofilm volume (Wilson et al. 2017). Inferring biofilm quantity through viable cell numbers should be supported with other assays as many factors affect biofilms carrying capacity of cells. Carrying capacity is defined as the maximum potential population size of a given landscape is capable of supporting and is a common attribute used to describe population dynamics (Stilling 1996). Biofilm cell carrying capacity is dependent on; but not limited to, carbon and oxygen depletion, shear stress, temperature, and iron availability (Madigan et al. 1949). These factors also regulate the rate at which biofilms shed planktonic cells.

After day 2, the control and UV-treated coupons approach similar concentrations of 10⁴ TC/coupon and 10^3.6 TC/cm², respectively. Similar observations were reported in previous studies (Wenjun & Wenjun 2009; Conrad 2018; Sperle et al. 2020), which showed UV exposures on planktonic cells that only delayed their biofilm formation in a continuous system. One limitation of this investigation is that only sub-lethal UV fluence was used to treat the planktonic cells before biofilm formation. Less viable cells were available for inoculation on the surface in the early stage of biofilm development. But, once the biofilm structure is formed, it may start capturing the incoming planktonic cells and utilizing the nutrients to grow and eventually reach a similar steady-state as the biofilms without UV pretreatment. Long-term biofilm inhibition by UV pretreatment is possible if the UV fluence is high enough to achieve intensive suppression or even elimination of viable planktonic cells. Another method employing UVLEDs for biofilm control is through direct UV exposure orienting the UVLEDs directly above the forming biofilm structure. Previous studies suggest that direct UV exposure can effectively inactivate pre-formed biofilms on surfaces (Gora et al. 2019; Ma et al. 2022) or inhibit biofilm formation (Torkzadeh et al. 2021), achieving a maximum reduction of more than 2 log (99%). In addition, recent studies (Lanzarini-Lopes et al. 2020; Zhao et al. 2023) demonstrated that UVLEDs coupled with side-emitting optical fibers could be an alternative configuration for effective biofilm control via direct UV exposure, providing greater surface area coverage, especially in systems with complex geometries like drip irrigation systems.

An important distinction not investigated in this study is whether the viable planktonic cells are growing within the bioreactor, then adhering to the coupons and forming biofilms, or if the planktonic cells immediately form biofilms then release more planktonic cells as biofilm maturity is reached. To achieve a REF_280nm of 40 mJ/cm², the UVLED device was operated at a specific flow rate corresponding to a 22-min hydraulic residence time within the bioreactors. Under favorable conditions, bacterial doubling time can be less than 20 min (Chalor et al. 2012). This suggests that planktonic cell concentrations could be increasing within the bioreactor independent of biofilm adhesion and maturity.

To understand how the UVLED device may change microbial conditions within the bioreactor with an influent flow rate of 0.16 L/min, the ambient bioreactor effluent coliform concentrations were measured on day 5. A full-factorial ANOVA analysis examining the effects of time in days and reactor (control or UV treated) on effluent coliform revealed that the day 5 effluent coliform values were not statistically different between the non-UV and UV-treated bioreactor (p = 0.65). This supports the previous conclusion that any viable cells surviving UV disinfection could potentially colonize the bioreactor. Interestingly, the cell carrying capacity of the coupons (∼10⁴ CFU/coupon or ∼10^3.6 CFU/cm²) is similar to the planktonic cell concentration of the wastewater effluent within the bioreactors (∼10^3.25 CFU/mL). The data may suggest the planktonic cell concentration environment is influencing the carrying capacity of the biofilms, or the opposite, the biofilm may be influencing the planktonic cell concentration in the bioreactor. If biofilm maturity is reached rapidly, planktonic cell dispersion from the mature biofilm may influence the planktonic cell concentration in the bioreactor effluent.

In addition to direct quantification, biofilms can be quantified indirectly through the use of a proxy marker, such as CV or metabolites. There is a general assumption that proxy markers are directly related to cell concentration within the biofilm (Azeredo et al. 2017; Wilson et al. 2017). However, these indirect biofilm quantification techniques are often dependent on metabolic function and biomolecule production that may not affect cell concentration. This study followed the recommendation provided by Wilson et al. (2017) to pair indirect quantification methods with a direct quantification method.

CV staining was first described by Christensen et al. (1985) and is now one of the most optimized microbiological methods for the identification and visualization of bacteria (Wilson et al. 2017). A CV assay was performed on bioreactor coupons on days 1, 3, and 5 of the experiment. Although Gram-positive and Gram-negative cells can be differentiated via microscopy after CV staining, this was not examined. It should be noted that CV stains cell membranes regardless of cell viability and is therefore a quantification of biomass and not viable cell concentration (Wilson et al. 2017). To accurately capture the amount of EPS, a prerinse step was performed prior to CV staining to attempt to ‘wash’ planktonic cells from the coupon biofilm.

Due to resource limitations, the ATP analysis was only performed on day 5 to provide an understanding of the overall bioactivity after 5 days of biofilm development. ATP was sampled from the bioreactor coupons as well as bioreactor effluent. UV prevents cells from replicating but does not immediately inhibit ATP production in the cell after inactivation. To observe differences between UV and non-UV-treated samples, the coupon and bioreactor effluent samples were incubated for 4 h post-sampling at 37 °C to allow for replication of active cells (Rauch et al. 2019). Allowing cells to undergo replication cycles post-UV provides resolution for the ATP assay between treated and non-UV-treated samples that would not be detectable immediately following UV disinfection.

Rauch et al. (2019) developed an ATP biomass recovery method for E. coli and wastewater communities following UV exposure. Adoption of the method has been successfully reproduced in studies examining the effects of UV on pure cell cultures ATP (Gora et al. 2019). Miller et al. (2020) found that a 2-h incubation after UV exposure was enough to diminish background noise for ATP measurements at a pilot-scale direct potable reuse facility, and higher sensitivity was achieved from ATP than flow cytometry following UV advanced oxidation processes (UV-AOP). Hammes et al. (2010) was able to calculate an average ATP-per-cell value (1.75 × 10⁻¹⁰ nmol/cell) from microbial samples taken from a variety of aquatic environments including drinking water, groundwater, river water, and wastewater effluent. Although the study standardized ATP production per cell, the authors also recognized the high ATP heterogeneity of microbial communities of different samples and variations in ATP extraction and analysis procedures when different natural water microbial communities are analyzed (Wilson et al. 1981; Schneider & Gourse 2004; Eydal & Pedersen 2007; Hammes et al. 2010). Rauch et al. (2019) also recognized that variations in microbial communities (like slow-growing organisms) may complicate the method and indicated the importance of the incoming water matrix on the outcome of the biomass recovery-ATP test. The ATP results presented in this study suggest that the 4-h 37 °C incubation was not sufficient at suppressing heightened ATP production in microbial communities post-UV. Future studies should build upon the recommendations made by Rauch et al. (2019) and examine the relationship between UV irradiation and natural microbial community ATP, as opposed to pure cell cultures, over time frames beyond 4 h.

A different explanation for the ATP results on day 5 is the high level of variability characteristic of biofilms and their quantification. Many coliform-specific biofilm studies have noted this stochastic behavior including a study examining coliform retention and biofouling within irrigation pipes (Shelton et al. 2013) and a study investigating drinking-water systems plagued by coliform regrowth (Camper et al. 1996). Through an 18-month survey of 31 water systems in North America, LeChevallier et al. (1996) concluded that the occurrence of coliforms is dependent upon complex interactions between chemical, physical, operational, and engineering parameters. This reinforces the recommendation made by Wilson et al. (2017), to draw upon multiple biofilm assessment methods that include both direct and indirect techniques to improve understanding and knowledge surrounding biofilms. Table 1 provides an overview of the biofilm quantification methods used in this study including the quantification target, quantification time, and high-level pros and cons of each technique.

Challenge testing was conducted for a UVLED flow-through device using three test waters of varying UVT_254nm (93, 80, 70%). The study concluded that the UVLED device performed similarly to other UV flow-through devices discussed in the literature, with tailing of inactivation performance at higher flow rates and decreased performance for low UVT waters (Oguma et al. 2013; Barstow et al. 2014). When operating with 70% UVT_254nm effluent at low (<0.036 L/min) flow rates, the UVLED device produced fluences >40 mJ/cm² corresponding to the 2012 US EPA Water Reuse Guidelines for the use of reclaimed water for irrigation of food crops (nondetectable fecal coliforms), and processed/non-food crops (<200 fecal coliforms/100 mL) (US EPA 2012). For disinfected tertiary recycled water intended for agriculture, California Title 22 reuse regulations allow disinfection methods that, when combined with filtration, produce 5-log inactivation of PFU of F-specific bacteriophage MS-2, poliovirus, or a similar virus (Title 22 Code of Regulations). According to NWRI analysis of UV fluence at 254 nm and MS2 inactivation, a 5-log removal of MS2 corresponds to a fluence of 121 mJ/cm² (NWRI 2012). Using the regression analysis presented in Figure 3, the UVLED system examined for this study would need to operate at 0.003 L/min to produce a REF_254nm of 121 mJ/cm² for a 70% UVT_254nm water matrix (assuming a worst-case scenario where the pre-filter provided no virus removal). For reference, drip irrigation row flow rates are generally in the range of 0.8 L/min per 100 feet of row length (Burt 2008). This level of fluence necessitates an evaluation or potentially UV LED device such as a PearlAqua Deca (AquiSense Technologies, Erlanger, Kentucky) with higher UV output or potentially pairing UV treatment with pretreatments, like filtration, that improve the UV transmittance of the distributed reuse water.

The UVLED flow-through device was then assessed for its ability to prevent biofouling in a CDC bioreactor fed by unfiltered wastewater effluent operating at a 40 mJ/cm² REF_280nm. This study demonstrated UVLED disinfection retarded biofouling in a bioreactor fed by unfiltered wastewater effluent. However, by day 5, both non-UV and UV-treated bioreactors approached 10⁴ CFU/coupon as well as 10^3.75 coliforms per mL of bioreactor effluent. Examining higher reduction equivalent fluences or more extensive pretreatments, such as filtration, would provide a meaningful contribution to this body of literature surrounding biofouling control by UVLEDs. Understanding the interaction between planktonic cells and sessile cells (i.e., whether the planktonic cell concentration is influencing biofouling; or if a rapidly formed mature biofilm is dispersing planktonic cells) would also help advance the conclusions drawn in this study. Overall, UV pretreatment at REF_280nm of 40 mJ/cm² may slow but will not ultimately eliminate biofouling in drip irrigation lines. The impact on drip emitters also may differ due to the hydrodynamics of the water flow through the emitters, which was not tested herein but is addressed in a companion study as part of this larger body of funded research.

Analysis of biological activity after 5 days of biofilm development using ATP did not align with the quantification of biofilm biomass using enumeration assays. This can be attributed to either (1) the high variability known to exist within biofilm behavior and resulting quantification or (2) known challenges in applying ATP assays to UV disinfection and possible insufficient incubation conditions following UV. Natural microbial communities in wastewater will vary geographically and may exhibit varying physiological UV-stress responses depending on treatment conditions. Future studies examining the relationship between ATP production in mixed cell cultures or microbial communities post-UV should be conducted to advance the understanding of microbial response to UV-induced stress.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.