UV VALIDATION: HOW TO INTERPRET A VALIDATION REPORT AND CURRENT VALIDATION PRACTICE TRENDS ABSTRACT Peter D Adamo, Ph.D., P.E. and Chengyue Shen, Ph.D., P.E. HDR The USEPA Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR, January 2006) recognized germicidal Ultraviolet (UV) light as an effective disinfectant to improve control of Cryptosporidium and other microbial pathogens. However, unlike chemical disinfectants (e.g., chlorine and ozone) whose dose can be quantitatively determined by a CT value, UV doses delivered by UV disinfection systems have to be determined and credited through full-scale validation testing. In current practice, the Ultraviolet Disinfection Guidance Manual (UVDGM, USEPA, November 2006) for the LT2ESWTR primarily directs UV validation in drinking water applications. Due to the complexity of the UVDGM validation protocol on field operation and subsequent data analysis, it is critical for design engineers and regulators to understand the components of a typical validation report for a commercial UV system. These include: What test analysis determines the performance of the validated UV system? What does a credited RED based on a biological challenge surrogate mean? What is the definition of the Validation Factor and what is its impact on system sizing? How is the information outlined in the validation report used for design and operation of a UV disinfection facility? How does one compare dose delivery performance of different UV systems based on their validation reports, especially when they are validated with different surrogates? In addition, UV-related technologies and the industry s understanding of the UV disinfection process have both advanced significantly in the UV community since the release of the UVDGM protocol. New issues have been raised and discussed to improve the validation process and the subsequent data analysis and reporting to ensure public health protection. For example: According to recent research results, validations following the UVDGM (2006) protocol may overestimate the disinfection credit for UV systems using medium-pressure lamps with polychromatic UV output on Cryptosporidium disinfection. Corrective actions are being studied by a group of UV experts and final recommendations are expected to be offered. Recent practice has suggested alternative microbial surrogates, including Aspergillus brasiliensis conidia and Bacillus pumilus spores for high dose UV validation for adenovirus disinfection. This presentation will address these topics within a case-study: a UV disinfection system that has been validated, its validation report written to reflect new validation trends, and use of the results for the design, commissioning, and operation. This presentation will benefit utilities and regulators considering or evaluating UV disinfection. KEYWORDS UV Systems, UV Validation, LT2ESWTR, Biodosimetry, Reduction Equivalent Dose (RED) INTRODUCTION Disinfection using Ultraviolet light (UV) is becoming more prevalent in the drinking water industry due to the ability of UV light to inactivate pathogenic microorganisms without the formation of disinfection by-
products. Some utilities have included UV disinfection as additional barrier in their treatment train particularly when their raw water supplies are vulnerable to microbial contamination. The USEPA Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR, January 2006) recognized germicidal UV light as an effective disinfectant to improve control of Cryptosporidium and other microbial pathogens. UV disinfection is one of the tool box strategies under the LT2EWSTR for addressing water supplies vulnerable to Cryptosporidium. However, unlike chemical disinfectants (e.g., chlorine and ozone) whose dose can be quantitatively determined by a CT value, UV doses delivered by UV disinfection systems have to be determined and credited through full-scale validation testing. In current practice, the Ultraviolet Disinfection Guidance Manual (UVDGM, USEPA, November 2006) for the LT2ESWTR provides critical design and operational input regarding UV systems including requirements for monitoring, system reliability, redundancy requirements, and lamp cleaning, replacement and breakage. Key regulatory requirements include UV dose, UV reactor validation, monitoring, reporting and off-specification compliance as it pertains to drinking water applications. This paper focuses on the basics of the UV validation process, key components of a UV validation report and current practice trends. UV VALIDATION PROCESS Under the LT2ESWTR, the EPA developed UV dose requirements to receive credit for inactivation of Cryptosporidium, Giardia, and viruses (Table 1). The UV dose depends on the UV intensity, the flow rate of the water treated and the UV transmittance. Table 1. UV Dose Requirements (mj/cm 2 ) 1 Target Log Inactivation Pathogens 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Cryptosporidium 1.6 2.5 3.9 5.8 8.5 12 15 22 Giardia 1.5 2.1 3.0 5.2 7.7 11 15 22 Virus 39 58 79 100 121 143 163 186 1 40CFR 141.720(d)(1) There can be sources of uncertainty with full-scale UV disinfection systems related to the hydraulic conditions, the equipment selected and overall monitoring approach. Because of these uncertainties, the LT2ESWTR requires UV reactors undergo validation testing. The objective of the validation testing is to determine the operating conditions that the reactor will deliver the required UV dose to achieve the log inactivation credit identified in Table 1. As required under the LT2ESWTR, validation testing requires full scale testing of reactors that conform to the specific application as well as demonstrate inactivation of a test microorganism. Specific validation requirements under the LT2ESWTR are summarized in Table 2. Table 2. LT2ESWTR Validation Requirements Requirement Conditions Citation Validated operating conditions Flow rate 40 CFR 141.720 (d)(2) UV intensity via UV sensor UV lamp status Validation testing includes Full-scale testing of conformed reactor Test organism inactivation whose dose response characteristics have been quantified with a low-pressure mercury vapor lamp 40 CFR 141.720 (d)(2)(ii)
Validation testing accounts for UV absorbance of test water Lamp fouling and aging On-line sensor uncertainties Lamp and critical component failure Inlet/outlet/channel configuration 40 CFR 141.720 (d)(2)(i) EPA s recommended validation protocol is based on the biodosimetry process summarized in Figure 1. Figure 1. Biodosimetry Process The Biodosimetry Process is a two step approach involving both full-scale reactor testing and bench scale collimated beam testing. Collimated beam testing determines the UV-dose response for the target microorganism. UV dose is calculated from the intensity of the incident UV light, UV absorbance and time of exposure. A dose-response curve and equation is generated from this testing. During full-scale reactor testing, the log inactivation of the target microorganism is measured under specific flow rates, UVT and UV intensity. Full-scale testing log inactivation values are input into the bench scale derived dose-response equation to estimate the Reduction Equivalent Dose or RED. RED values are specific to the bench scale target microorganism and the full-scale validation test conditions. The RED value can be adjusted for uncertainties and biases to produce the validated dose of the reactor for the specific operating conditions tested. The validated dose is defined as the RED value divided by the Validation Factor (VF) and is compared to the required dose for compliance purposes.
UV VALIDATION REPORT KEY COMPONENTS The validation report provides detailed information from the validation testing. Key components of the validation report include: Objective of Validation o Generic validation vs. site-specific validation Area of Application o DW, WW, water reuse Description of the UV System Technical Testing Bioassay o Challenge surrogates, dose-response, field tests QA/QC Data Analysis o Validation factor, and credited dose/log inactivation A key element of the validation report is development of the calculated RED algorithm. Critical equations for the RED algorithm and Credited/Validated RED, and Log Inactivation (LI) and VF are summarized below. RED Algorithm where: e f g h 1 RED calc = 10 ( S) ( Q) ( N) ( ) A RED calc = Calculated surrogate dose, mj/cm 2 S obs = observed intensity, mw/cm 2 Q = flow rate, gpm N = number of lamp modules or rows, unitless A 254 = water abs at 254 nm, Log(UVT 254 /100), cm -1 e, f, g, h, I = equation coefficients 254 i Credited/Validated RED and LI where: RED val > Dose Required RED val = RED Calc /VF or LI val = LI Calc /VF LI calc and RED Calc is the dosimeter calculated log inactivation or reduction equivalent dose (based on LI or dose algorithm), as a function of operating variables VF is the Validation Factor, which accounts for certain potential biases and experimental uncertainty Validation Factor (Expanded version) VF = B RED x B POLY x CF as x {1 + (U val /100)}
where: B RED is the RED Bias, a correction factor that accounts for the difference between the UV sensitivity of the target pathogen and the test surrogate and for unknown dose-distributions time and intensity variations B POLY is the polychromatic bias, which accounts for potential impacts of non-germicidal output spectra of MP lamps Modern systems have germicidal sensors (<300 nm, peak response between 240 and 280 nm) B POLY = 1.0 CF as is a correction factor if MS2 or B. subtilis spores are not used, and the action spectra of the surrogate is significantly different than the target pathogen. Ignore if CF as is <1.06. U val is the experimental uncertainty factor Validation Factor (Reduced Version) VF = B RED x {1 + (U val /100)} B RED is a function of the UVT and the comparative sensitivity of the test surrogate vs. the targeted microbe U val is the combined uncertainty associated with the Dose-Response method, the field RED measurements and the sensor variability U val is defined as follows: where: U val = (U IN 2 + U S 2 + U DR 2 ) 1/2 U IN is the uncertainty of interpolation U S is the uncertainty of the duty sensors used for validation (QA Component Ignore if criteria are satisfied) U DR is the uncertainty of the collimated beam dose-response relationship used to determine the RED (QA Component Ignore if criteria are satisfied) In the equation above, U in can be determined statistically from the calculated and measured RED using the formulas for the RED calc and U in as defined below. where: U in = 100 * (t*sd)/red RED = Average RED value measured for each test condition SD = Standard deviation of the RED values measured for each test condition t = t-statistic for a 95-percent confidence level defined as a function of the number of replicate samples A typical data plot of Observed and Calculated RED is shown in Figure 2.
70.0 60.0 Calculated RED (mj/cm 2 ) 50.0 40.0 30.0 20.0 10.0 0.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 Figure 2. Calculated vs. Observed RED Measurements Selection of a target or challenge organism is also an important component of validation testing. It is important to consider the following when selecting a challenge organism: Use a microbe that closely matches the target organism s sensitivity to UV o Reproducible kinetics over the dose range of interest o Culturable in large quantities at high titers o Easily enumerated o Non-pathogenic, environmentally safe Typical challenge organisms for Cryptosporidium include coliphages MS2, Q-Beta, T7, and T1UV. Use of a less-sensitive organism (such as MS2) relative to targeted pathogen may yield an oversized system, but may also mask inefficiencies. The RED Bias must be determined for specific challenge organisms. The RED Bias is a correction factor that accounts for the difference between the UV dose measured with a surrogate microorganism and the UV dose that would be delivered to a target pathogen due to differences in the microorganisms inactivation kinetics. RED Bias and VF for specific microorganisms are summarized in Table 3. Table 3. RED Bias vs. Test Organism Test Organism RED Bias 1 Validation Factor 2 MS2 1.77 2.21 Q-Beta 1.42 1.78 T1UV 1.18 1.48 T7 1.03 1.29 1. 3-log Crypto, UVT 90%, U val = 25%, B poly =1 2. VF = 1.00 under ideal conditions Observed RED (mj/cm 2 ) In summary, with good QA/QC and modern UV reactors (LP or MP), the VF will be comprised of the B RED and the U IN in accordance with the following relationship:
VF = B RED [1 + (U IN /100)] U IN is minimized by close control of field and laboratory tests while B RED is dependent on surrogate selection and UVT (higher at lower UVT). To minimize VF, it is important to use surrogates that closely mimic target pathogen sensitivity or measure dose-distribution directly. CURRENT VALIDATION PRACTICE TRENDS As the UV industry evolves, new data and trends are observed that impact the validation process. Some key trends observed post UVDGM 2006 relates to Medium Pressure (MP) UV lamps under low wavelength conditions. Specifically, non-representative MP lamp output spectrum was used in the UVDGM 2006 when developing key VF components. An example of this phenomenon is shown in Figure 3. Figure 3. MP Lamp Output Spectrum Another trend observed under low wavelength conditions for MP lamps is a significant difference in action spectra between popular biological surrogates (MS2, T1UV, etc.) and target pathogens in drinking water (Cryptosporidium) at the low wavelength range (200 to 240 nm). This is demonstrated in Figure 4. Some surrogates (e.g., MS2 phage) are inactivated more easily than target pathogens (e.g., Cryptosporidium) by the light in the low wavelength UV spectra. Given this observation, using MS2 (or T1UV) as the validation surrogate may overestimate the disinfection credit for Cryptosporidium.
Figure 4. Microbial Action Spectra Based on recent trends and concerns, the American Water Works Association (AWWA) and International Ultraviolet Association (IUVA) working group released recommendations related to UV disinfection and low wavelength, medium pressure disinfection. These recommendations are summarized below. Rely on Conservatism in UV Design. Apply a Blanket Action Spectra Correction Factor (ASCF) (For example, California, 1.30) Apply a Site-Specific ASCF based on CFD simulations Modify Validation Protocols (Use doped quartz in validations (block <240 nm UV)) ADENOVIRUS VALIDATIONS Validating UV systems for virus removal is particularly applicable for groundwater systems where no additional treatment may be required. Based on the high UV resistance of adenoviruses, the EPA s LT2ESWTR) and Ground Water Rule (GWR) require a UV dose of 186 mj/cm 2 for 4-log inactivation of viruses. Using live adenovirus as the challenge organism poses many difficulties including problematic virus stock preparation, the existence of few qualified laboratories to conduct the analysis and high validation costs. Possible alternative challenge organisms include Aspergillus brasiliensis spores and Bacillus pumilus spores. Key considerations in using these challenge organisms are summarized in Table 4. Table 4. Alternative Adenovirus Challenge Organisms for Virus Removal Validation Testing Aspergillus brasiliensis Spores High resistance to UV and the stock can be prepared in relatively high titer to meet fullscale validation needs Has been applied in full-scale UV validations and performance is demonstrated Action spectrum information is unknown; application only in low-pressure monochromatic UV validation for virus credit Bacillus pumilus Spores Also high resistance to UV UV sensitivity can be manipulated, but fixed once prepared Easier to handle when comparing to live adenovirus or aspergillus brasiliensis spores Has been demonstrated in full-scale tests, and ready to be used in actual validations Action spectrum shows a significant sensitivity at the lower wavelengths, similar to that observed by the various Adenovirus strains. This makes it a potential candidate for both LP and MP applications
CONCLUSIONS The use of UV disinfection has increased dramatically over the last 20 years with applications for surface water and groundwater treatment systems and reuse systems increasing markedly during this period. UV disinfection is one of the primary toolbox treatment strategies for Bin 2 and above surface waters under the LT2ESWTR. Because UV dose cannot be quantitatively determined by a CT value, doses delivered by UV disinfection systems have to be determined and credited through full-scale validation testing. UV validation requires both full-scale testing as well as bench scale testing to derive the validated dose needed to achieve the level of inactivation required or desired for target pathogenic organisms such as Cryptosporidium, Giardia and viruses. Important considerations in the validation process include determining the RED, VF and appropriate challenge organisms for validation testing. Key components of the UV Validation test include a statement of the validation objectives, application area (i.e., drinking water), UV system description, technical and bioassay test procedure and results, QA/QC analysis and data analysis including computation of the VF and credited dose/log inactivation. REFERENCES Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced Surface Water Treatment Rule, EPA 815-R-06-007, November 2006.