Attachments ( 58 files ):
  • 57 image files
  • 1 octet-stream file

Wastewater Ultraviolet Disinfection 1

Disinfection by ultraviolet irradiation is a relatively cost effective method for disinfecting water and wastewater. It has a similar cost to the combined chlorination and dechlorination processes. Economics for the chlorination process depend on the treatment plant capacity and prevailing regulatory and safety requirements for storage of chlorine.  The UV irradiation process does not produce significant quantities of disinfection by-products, such as does the chlorination process.

Disinfection by ultraviolet irradiation is suitable for inactivating; bacteria (Wilson, 1992) (eg. coliforms, salmonella spp. (Keller, 2003), viruses (Jacangelo, 2003), (Thurston-Enreiquez,  2003) (eg. poliovirus, rotavirus, enterovirus, adenovirus (Durance, 2000), (Thompson, 2003) and bacteriophages) and protozoan parasites, such as Cryptosporidium (Shin, 2000), (Linden, 2000) and Giardia species. Recently two studies have questioned the efficacy of UV radiation for inactivating Giardia in treated wastewater effluent, in full-scale operating plants (Li, 2009), (Cantusio Neto, 2006). It is difficult to inactivate Ascaris eggs with ultraviolet radiation (de Lemos Chernicharo, 2003 ), because the ultraviolet light cannot easily penetrate the outer two layers of skin of the Ascaris egg (Brownell, 2006).

N.B.  This article is divided into 2 parts. Please Click Here to read Part 2

Table of Content

The Action of Ultraviolet Light

Ultraviolet light is able to inactivate microorganisms because it damages the DNA or RNA of the organism. The organism is still able to function, however cannot use its DNA/RNA to reproduce and so the population of the organism dies out. Such injured microorganisms have been shown to display similar metabolic rates, such as respiration rates, compared with healthy organisms (Blatchley III, 2001).

The damage to the DNA primarily consists of the creation of two bonds between adjacent Thymine groups on the single DNA strand (called Thymine dimers). Cytosine-Thymine, and Cytosine dimers are also possible, although less probable. Uracil, Uracil-Cytosine and Cytosine dimers are produced by irradiation of RNA. Pyrimidine-Pyrimidone photoproducts, DNA-DNA crosslinks and DNA-protein crosslinks are also possible, but 1000 times less likely (USEPA, 2006c). The damage to DNA/RNA alters the conformation of the DNA/RNA strand, such that it cannot replicate at that site. When the damage is significant enough the DNA/RNA strand is rendered useless.


Figure 1: Diagram showing typical pyrimidine dimers.


Figure 2: Structure of a Thymine dimer, showing the cyclobutane ring formed. In DNA the two Thymine groups are arranged roughly in parallel planes, one on top of the other (Jagger, 1967).


Figure 3: Ultraviolet light and the electromagnetic spectrum


Chart showing typical UV absorbance by Thymine

Source: Masschelein (2002)

Figure 4: Absorbance spectrum of thymine.

Ultraviolet light in the UV-C band is responsible for this damage. Specifically damage by ultraviolet radiation in the range of 230 to 290 nm is responsible for the damage to DNA/RNA.

There are mechanisms in cells which repair damage to the DNA/RNA. These can be divided into light-catalysed repair, called photoreactivation, and dark space repair mechanisms. Photoreactivation is the cleavage of the Thymine-Thymine (Cytosine or Uracil) double bonds, initiated by enzymes and (blue to UV - 350 to 450 nm) light wavelengths, and in the case of E. coli can account for repair of the order of 1% of original viable organisms. The lower the UV dose administered to the organisms (the less the DNA damage) and the higher the visible light irradiation time the greater is the repair effect of photoreactivation. Dark space repair accounts for a much lower fractional repair of organisms (E.coli or thermotolerant coliforms typically of the order of 0.05% of original viable organisms). Enzymatic mechanisms such as excision and SOS repair account for this repair.


An ultraviolet disinfection bank consists of one or more UV lamps and a conduit or duct in which the water or wastewater to be irradiated flows. A UV light transparent material separates the lamps from the water/wastewater. As ultraviolet light is absorbed by ordinary glass, usually a pure quartz is used as the separating material. The UV lamps are generally enclosed in a cylindrical quartz sleeve and the water to be irradiated flows on the outside of the sleeves. Alternatively, a system where the water is enclosed in a UV transparent fluoropolymer pipe or tube and the lamps arranged parallel to the tube in the air space, are manufactured, in both USA and Australia. Pictures of typical UV disinfection units are presented below:


Figure 5: a) Trojan horizontal lamp open channel unit. 3 modules installed in SS channel. b) Trojan open channel unit during construction. Only half of the lamp modules installed. (Courtesy Aquatec Maxcon. Used with permission and not to be used elsewhere)


Figure 6: a) Severn Trent Microdynamics vertical lamp, in-channel, microwave lamps. 
b) Showing Severn-Trent Microdynamics vertical lamp unit, installed in-channel. (Courtesy Severn Trent Services. Used with permission. Not to be used elsewhere)

Figure 6: a) Severn Trent Microdynamics vertical lamp, in-channel, microwave lamps.
b) Showing Severn-Trent Microdynamics vertical lamp unit, installed in-channel. (Courtesy Severn Trent Services. Used with permission. Not to be used elsewhere)


Figure 7: Trojan UV Swift medium pressure lamp in-pipe UV disinfection units. (Courtesy Trojan UV. Used with permission and not to be used elsewhere.


Figure 8: Trojan UV Fit LPHO Closed pipe UV disinfection system, Torrevieja, Spain. Design is 60,000 m3/d total capacity treated wastewater, 10 mg/L suspended solids, 55% UV transmittance, target 100 thermotolerant coliforms/100 mL. (Courtesy Aquatec Maxcon and Trojan UV. Used with permission. Not to be used elsewhere)

Haylock11.jpgFigure 9: a) Berson In-line medium pressure lamp, closed pipe, UV disinfection system and UV intensity probe at top, b) Cutaway photograph showing internals of the unit, with lamps operating. White rings are part of the lamp wiper system. (Courtesy Berson. Used with permission, and not to be used elsewhere.)

Generally for systems, a number of UV lamps are required to give the required UV dose. The lamps are generally arranged as a number of banks or stages and, in each, a number of lamps arranged in a regular separation or arrangement, such as on a square grid for each bank or stage. The terminology is (NWRI, 2003):

  • Lamp - ndividual lamps or lamp pairs in the same module.
  • Module  - Separate collections of UV lamps, joined together and sharing a common
    electrical feed.
  • Bank - Separate UV modules together, through which the channel flow must pass though each complete bank. Banks can be turned on and off separately in response to water quality and flowrate changes.
  • Reactor - A combination of multiple Banks in series, with a common mode of failure (e.g. electrics, cleaning system).
  • Reactor Train - A combination of reactors in series, which includes inlet, outlet and and level control arrangements.
  • System - Combination of parallel UV reactor trains or side-by-side flow system

Elimination of dead spaces is necessary for efficient design of the irradiated section of the UV reactor, as these areas reduce useful detention time and thus disinfection. Because disinfection aims to typically achieve two to four log (99 to 99.99 %) microbial inactivation, it is important to have minimal backmixing from the water’s inlet to outlet. Having some mixing perpendicular to the axis of flow, ensures that all parcels of fluid receive a broad range of UV intensities through their travel through the reactor. Parcels of fluid that move through the reactor solely a large distance from the UV lamp, and receive minimal disinfection, are thus eliminated.

The UV lamps are electrically powered, generally through a ballast and wiring system. There are controls to switch lamp stages, monitor lamp operation and to adjust lamp UV energy input, to meet disinfection requirements (due to flowrate and UV transmittance changes).

Ultraviolet Dose Requirements

The energy required to inactivate a microorganism is termed the UV dose or UV fluence and is measured in, or, or J.m-2. UV fluence is the modern term, used to represent UV light impacting on a small parcel of fluid from all directions. UV light is scattered by particulate and colloidal material, meaning that it does not merely radiate outwards from the lamp. Light is scattered in all directions, including forward- and back-scattered.

1 = 1000 = 10 J.m-2

This is defined as being the amount of UV energy reaching the organism (the intensity) multiplied by the contact time (the time the microorganism is in the UV irradiation field).

Dose = Average intensity (eg. x contact time (eg. seconds)

Severin (Severin, 1983) investigated the temperature dependence of some bacteria and viruses. Between the temperatures of 5 and 35 ºC. He noted that the inactivation rate increased by less than 20% for f2 bacteriophage, for the above increase in temperature and increased by less than 10% for E. coli and Candida parapsilosis(Is the conclusion correct?). Malley (Malley) (Can someone insert the reference. UVDGM ref is incorrect. Is the conclusion correct?) found that MS2 phage inactivation was temperature independent.

Malley (Malley) (Can someone insert the reference. UVDGM ref is incorrect) found that inactivation is independent of pH, between pH 6 and 9.

Sommer (Sommer, 1999) investigated the reciprocity of time and intensity for E.coli strains, Bacillus subtillis and coliphages. Tests revealed higher inactivation at the higher UV intensities and the same dose, for E. coli. Reciprocity differences for other microorganisms were not statistically significant. It is hypothesized that the difference for E. coli was due to repair mechanisms taking place in the low intensity case. Others (Oliver, 1975) (Rice, 2001) investigating reciprocity over a range of intensity of 1 to 200 mW/cm2 found no difference in inactivation rates

Different microorganism species and individual strains have differing responses to ultraviolet radiation. Typical values are provided in the table below, for essentially free microorganisms. Additional information can be found in Hijnen and USEPA. Organisms attached to particulates in wastewater may require doses of 2, 3 or more times, to achieve the same log reduction as for free organisms.

Table 1: Quantitative UV doses for inactivation of pure cultures of bacteria, viruses and protzoa in water, adapted from Hijnen (Hijnen, 2006) and USEPA (USEPA, 2006f)


Required Fluence (mJ/cm2)

1 log red’n

2 log red’n

4 log red’n

E. colia




E. coli O157d




Salmonella typhia




Clostridium perfringensa




Bacillus subtillisa




Streptococcus faecalisa




Camylobacter jejunid




Legionella pneumophilliad




Shigell dysenteriaed




Vibrio choleraed





Adenovirus type 40




Adenovirus types 2, 15, 40, 41




Calicivirus canine




Calicivirus feline




Calicivirus bovine




Coxsackie virus B5




Hepatitis A




Poliovirus type 1




Rotavirus SA-11




B40-8 phage




MS2 phage




Qb phage




fx174 phage




T7 phage




T1 phage




Cryptosporidium USEPAc




Giardia USEPAc








a  Environmental species.
b  MICmax < 4 log.
c  No correction for environmental species (research needed).
d  Corrected for environmental species.
e No value due to tailing.


Recently two studies have questioned the efficacy of UV radiation for inactivating Giardia in treated wastewater effluent (Li, 2009), (Cantusio Neto, 2006). This may be due to operational issues, aggregation, particle association, or dark space repair, in wastewater effluents.

Shin (Shin, 2009) found that, for inactivation of Adenovirus 2 (Ad2), lower doses from MP lamps gave the same inactivation as higher doses provided by LP lamps. It was hypothesized that the resistance of Ad2 is due to repair mechanisms and that the MP lamp caused a greater range of organism damage, including damage to possible repair enzymes.

Dose-Response Models

The first dose response model was Chick’s Law (USEPA, 1986a):


Equation 1


            k          = the microorganism inactivation constant (log10reduction.mJ-1.cm2)

            Dose    = applied UV dose (

            I           = Average UV reactor intensity (mWcm-2)

            t           = fluid retention time in the irradiated zone of the UV reactor (s)

Free microorganisms in water may exhibit this type of inactivation behavior.

Two effects were observed in plots of log survival versus UV dose:

Tailing – the dose response curve would curve upwards from its downward trend at higher doses. This phenomenon is hypothesized to be due to shielding of microorganisms by adsorption to particles or floc.

Shouldering – At low doses there would be a lag in log survival before the dose response curve would adopt the Chick behavior. This phenomenon is hypothesized to be due to the formation of microorganism aggregates, requiring multi-hit kinetics for inactivation (Masschelien, 2002a), photorepair (Hoyer, 1998) or dark space repair (Morton, 1969) mechanisms.


Figure 10: Simple Chick kinetics and influence of tailing behaviour.


Figure 11: Effect of shouldering.


Figure 12: Combined effect of shouldering and tailing. NB. The magnitudes of shouldering and tailing may be different.

Coliform Inactivation

Coliforms (total, thermotolerant and Escherichia coli) are the common target pathogens for confirming inactivation in wastewater disinfection. In wastewater disinfection the coliform inactivation curve consists of a 3 to 5 log linear (simple Chick form) (perhaps with some shouldering) response and then a significant zone of tailing.

Typical required doses for meeting a median 200 MPN/100 mL coliform standard (1.5 to 3.5 log10 reduction) for a filtered nitrified secondary effluent are 29 to 50 Metcalf & Eddy (Metcalf & Eddy, 2007) provides a comprehensive table of estimated UV doses for coliforms in various effluents.

The first equations to model the ultraviolet inactivation of coliforms in treated wastewater effluent were proposed by (Scheible, 1987) and (USEPA, 1986a).


Equation 2


N         = final coliform concentration

No        = initial coliform concentration

u          = wastewater velocity = x /(Ve/Q)

x          = reactor characteristic length

            = the avg length traveled by water while UV exposed

Ve        = reactor liquid side irradiated volume

Q         = wastewater flowrate

E          = reactor dispersion coefficient

            = accounts for non-plug-flow behavior in the reactor    

a          = UV inactivation constant

Inom      = Nominal intensity in the reactor for new lamps and clean sleeves

Fp         = fractional lamp UV output or UV power at end of lamp life compared

            with initial lamp output

Ft         = fraction of lamp UV output transmitted through the quartz sleeves 

b          = experimental constant

c          = experimental  constant

SS        = wastewater suspended solids concentration

m         = experimental constant

Whilst this is an extremely complex equation, the first grouped term incorporates factors to allow for the dispersion (or backmixing) of the wastewater flowing through the reactor.

The second term provides for the residual coliforms due to particulates at very high UV doses. Scheible fitted data for the Port Richmond facility, USA, as being:


Equation 3

The R2 value for this linear regression was however only 0.46.

The University of California at Davis, put in a significant effort to develop a coliform dose-response model. Their first model was an empirically fitted model, based on experiments at the Univ. Davis wastewater treatment plant. Tests consisted of disinfecting mixtures of mixed liquor and/or clarifier suspended solids with clarifier effluent (Emerick, 1993). The model gives a coliform residual as a function of suspended solids, unfiltered UV transmittance and UV dose. The influence of the initial coliform concentration on the residual coliform concentration was not statistically significant. Thus:


Equation 4

A Monte Carlo method design example and typical model coefficients, for two plants, are presented in Loge (Loge, 1996). The paper notes that the coefficients are site specific for a particular treatment plant’s treated effluent. A similar probabilistic design approach is presented in Tchobanoglous (Tchobanoglous, 1996).

The shortcoming of this model, however, was that it failed to predict residual coliform concentrations, for a range of treatment plants. This was discovered when trialing UV disinfection to meet the future California recycled water (Title 22, Division 4, Chapter 3) regulations, which set limits of median 2.2 total coliforms/100 mL, and maximum 23/100 mL, for water for irrigation of food crops and non restricted recreational impoundments. In particular, unusually high doses, of 300 mJ/cm2 (Kuo, 1997) were required to inactivate effluents from oxygen activated sludge treatment plants, compared with conventional activated sludge plants, where the required dose was 97 mJ/cm2 (Darby, 1993). Model coefficients for two plants are presented in (Emerick, 1999)

The second model for predicting residual coliform concentrations (the “Tailing Model”) was based on the following fundamental concepts:

  • The effluent residual coliform concentration is independent of the initial coliform concentration (i.e. Chick’s Law does not apply in the tailing region), because a significant part of the inactivation occurs in the early part of the UV dose (a very large fraction of coliforms are free from particle attachment);
  • The residual coliform concentration depends on the number of particles, of a particular size range, which have on them at least one coliform. Only one undisinfected coliform is required per particle for that particle to register as a positive or count for coliforms (e.g. plaque on an agar plate). Multiple coliforms on a particle will register as one coliform.

The whole of the work is summarized in three papers (Loge, 1999), (Emerick, 1999), (Emerick, 2000). The findings were that:

·        Light absorption by solids in particles is significant.

·        Particles of size 10 mm and larger can contain at least one coliform.

·        Coliform bacteria are uniformly distributed within particles, independent of particle size.

The fraction of particles of 10 mm and larger containing at least one coliform depends on the process type, but specifically depends on the MCRT of the activated sludge and particle size (there were two groups of particle sizes) (chemically dosed P removal plants an exception because of added chemical precipitate, however if this is taken into account the percentages could align with A/s plants, which vary with MCRT).

The work assumes that the fraction of average UV intensity that reaches the critical coliform (the most embedded coliform) varies between 1 and 0, and that this is distributed uniformly between 0 and 1. Integrating this simplest form of distribution, one arrives at the following equation which determines the numbers of particles remaining with a coliform after disinfection to be:


Equation 5


Np(t)    = Number of particles remaining with at least one coliform after UV

                 irradiation of time t (MPN/100 mL);

Np(0)   = Number of particles containing at least one coliform before UV

                 irradiation (MPN/100 mL);

            = Number of particles greater than 10 μm MPN/L x fraction of particles

                 containing at least one coliform;

k          = UV inactivation rate constant determined from inactivation of dispersed

                coliforms (mJ-1.cm2);

I           = Average reactor UV intensity ( after accounting for reduction

               in lamp output due sleeve fouling and lamp age; and

t           = time of irradiation (s).

The complete equation defining the final effluent coliform concentration is:


Equation 6


Nd(0)   = the initial concentration of disperse or free, coliforms

The form of the relationship is presented in the chart below:


Figure 13: The University Calif -Davis coliform inactivation equation .

The relationship was found to hold for all forms of treatment process (oxygen and air activated sludge, extended aeration, biological N removal and chemical and biological P removal and aerated lagoons). The authors note that this form of relationship only applies to coliforms in treated wastewater.

In particular the number of particles initially containing at least one coliform can be estimated as the total number concentration of particles in the effluent multiplied by the fraction of particles that would contain at least one coliform. The latter fraction appears to be only dependent on the MCRT for activated sludge plants, from oxygen to biological nitrogen and phosphorus removal plants. This relationship is presented in Metcalf and Eddy (Metcalf & Eddy, 2007a), (Metcalf & Eddy, 2003a). Alternatively Emerick (Emerick, 2000) recommends using the value of k from regression of the low dose particle free coliforms and then adjusting Np(0) until the least sum of the squares of the deviations of the tailing region data-points is minimized.

Given, however, the uncertainty in the measurement of coliform concentrations, others have questioned the proof of the basis of the relationship. The assumption of a uniform distribution for the hiding of coliforms in solids particles, independent of size, was questioned. Assuming the fraction of particles containing a coliform is independent of particle size was also questioned (Madge, 2001), (Ginn, 2001). Whilst the original authors present evidence to prove the even particle association for particles, grouped in 10 to 80 μm and >80 μm categories, more evidence, for different effluent types, would be desirable. This model, however, is the best universal model presently available, for modeling disinfection of coliforms in wastewater effluents. It seems to be a universal model; applying to all wastewater treatment types.

Alternative empirical models have been employed for mathematical modeling of coliforms with particle association, in treated wastewater.


Equation 7

(Wright, 2000a)


k1, k2    = empirically fitted constants for the free and particle associated organisms

No,Np   = empirically fitted constants for numbers of free and associated organisms

D         = the UV applied dose or fluence = Iavg.t



Equation 8

(Cantwell, 2007)


Β         = the fraction of UV resistant particles (containing a coliform or

               multiple coliforms)

D         = the UV applied dose or fluence = Iavg.t

Typical values for β are around 10-3.

Tools: The Collimated Beam Apparatus

There are not widely available models used to determine the microorganism response to UV radiation in treated wastewater effluents. The use of UV disinfection for inactivation of treated effluents for reuse, may require determination of UV doses for inactivation of coliforms, other bacteria, protozoan parasites such as Cryptosporidium and Giardia species and viruses, such as adenovirus, poliovirus and MS2 coliphage.

Difficulties such as non-uniform reactor hydraulics, dead space, and polychromatic medium pressure light sources, make design even more difficult.

The collimated beam apparatus is a method where dose-response relationships can be determined on a laboratory scale using the actual wastewater under consideration. This method can thus include the effects of indigenous particles and can employ indigenous or cultured and added microorganisms.

The collimated beam apparatus consists of (USEPA 2006a):

  • a low pressure mercury light source in an enclosed tube, which is ventilated to the outdoor atmosphere and which incorporates a temperature probe to determine the lamp skin temperature;
  • voltage control, ballast and switching system for the LP lamp;
  • a collimating tube, perpendicular to the UV lamp source, fitted with a screw thread, or ribs, to prevent reflections down the collimating tube;
  • shutter for shutting off the UV radiation from the collimated tube;
  • an optional additional collimating tube parallel and of the same dimensions as the other – for the purpose of simultaneously determining the UV intensity from the collimating tube;
  • a height adjustable, magnetic stirrer under the collimated tube, on which sits a sample dish  (of diameter smaller than the collimating tube) containing the effluent, stirred by a stirrer bar timer;
  • calibrated UV radiometer, to determine the UV intensity from the UV source; and
  • UV spectrophotometer for determining effluent UV transmittance or absorbance.

Collimated beam devices are available with medium pressure light sources, to enable lab-scale work on polychromatic aspects of UV disinfection.

A diagram of the apparatus is presented below:


Figure 14; Diagram of a collimated beam apparatus.

The procedure to use the Collimated Beam Apparatus is as follows:

  • Switch on UV light source and allow to warm up to normal operating temperature and maintain constant temperature;
  • determine the UV intensity at the end of the collimated beam, to irradiate the sample;
  • profile the UV intensity across the water surface of the dish, to determine the Petri Dish factor;
  • determine the UV transmittance of the wastewater sample following addition of microorganisms;
  • determine the delivered UV intensity according to Beer’s Law;
  • place a sample of the wastewater in the open sample dish and irradiate the sample for a set time to provide a fixed UV dose;
  • repeat the irradiation for replicate doses and different doses;
  • assay of the pre-irradiation and post irradiation samples for the required microorganism, to determine the microorganism log reduction;
  • plot log reduction or log concentration versus applied UV dose as described in USEPA(USEPA, 2006a).

The application of Beer’s law, to a sample, in an open sample dish, under the collimating tube is given by:


Equation 9

(USEPA, 2006a),(Metcalf & Eddy, 2007c)


            D         = UV Dose at 253.7 nm (

            Io        = incident intensity (at 253.7 nm) at the surface of the sample


            t         = exposure time (s)

            R         = reflectance of the water surface (dimensionless 0.025), or can be

                          estimated by Fresnel’s Law;

            Pf        = Petri Dish factor to allow for non-uniform radiation (dimensionless)

            a        = absorbance coefficient (cm-1, base e) = 2.303 A (cm-1, base 10 or a.u.)

                      = - lne (UV%T/100%).               

            d        = sample depth in the uncovered stirred dish (cm)

             L       = distance from lamp centerline to liquid surface (cm)

The Delivered Dose

The energy required to inactivate a microorganism is termed the UV dose or UV fluence and is measured in, or, or J.m-2.

1 = 1000 = 10 J.m-2

This is defined as being the amount of UV energy reaching the organism (the intensity) multiplied by the contact time (the time the microorganism is in the UV irradiation field).

Dose = Average intensity (e.g. x contact time (e.g. seconds)  

Equation 10

The intensity within a reactor is actually an average intensity across the spacing between the lamps. The further the distance a point is from the lamp in the UV reactor the lower is the UV intensity, due to; (1) as the energy radiates out from the lamp the area which the same amount of energy passes through increases; and (2) absorption of the UV radiation by substances in the water. The latter effect is quantified by a parameter called the UV transmittance, or the UV absorption coefficient.

The intensity is the rate of passage of UV radiation (in watts for example) across a fixed plane, perpendicular to the travel of the radiation, of defined area (in cm2 for example). This is presented diagrammatically in the next figure. If the distance from the source to the point where the intensity is measured is doubled from L to 2L, then the intensity falls off to W x 10^(-A x L)/2cm x 2cm = W.10^(-A.L) /4 cm2, to take into account the spreading of the radiation and water absorbance.


Figure 15: Diagram showing the definition of UV intensity from a point source

Thus the average intensity in a reactor is only a function of the lamp type, electrical power input (usually documented at 100% power input), geometric arrangement of the lamps, lamp spacing (or fluoropolymer tube inside diameter), quartz sleeve outer diameter, and the UV Transmittance of the wastewater. Usually a UV disinfection supplier has a curve of Average Intensity for each model (lamp, lamp spacing and geometry) of UV disinfection unit, as a function of UV Transmittance.

Point Source Summation Methods for Calculation of Average Intensity

Contribution required

Tools: UV Transmittance

The UV transmittance of a water sample is defined as the fraction of incident light at 254 (or 253.7) nm, remaining, after passage through a 1.0 cm pathlength of a sample of the water and is measured in percent.

UV   Transmittance [% (cm-1)] = Iout/ Iin(254 nm, 1 cm pathlength)               

Equation 11

The 100 % transmittance calibration standard is to use distilled water in place of the sample. The zero standard is a totally blocked light path.

This model of light absorption follows Beer’s Law, which states that:

I / Io = exp (-e(l).c.l)                                                              

Equation 12


             e(l).c   =  the extinction coefficient of the absorbing substance times the

                              concentration of absorbing substance in the sample;

             l           =  the pathlength through the sample (usually 1.0 cm).


              I / Io   = UV% Transmittance [% (cm-1)]          

                         = exp (- a [cm-1] x 1[cm])                                                                                                                                                                                         

Equation 13                                                    

                        = 10^(- A [cm-1] x 1 [cm])

Equation 14


            a [cm-1 (base e)] = 2.303 A [cm-1 (base 10)] = - lne (UV%T [% (cm-1)])

Equation 15

Thus A and a are absorbance coefficients, and are alternative measures of the UV transmittance coefficient. They are more convenient to work with when looking directly at numbers of lamps, because the coefficient scales directly with UV system size. If you are using either of these absorbance coefficients, please be sure to qualify the absorbance coefficient, defining whether you are working in base e or base 10. Base 10 matches with the conventional “Absorbance” or “a.u.” scale of spectrophotometers and is commonly referred.

Suspended solids (or turbidity) and colour are the main substances, interfering with or reducing the UV transmittance of the sample, by blocking or scattering the light beam and absorbing the UV energy, respectively. It is also important to identify any other substances in wastewaters which may detrimentally impact on UV Transmittance. These can include:

·        UV absorbers         - textile dyes (can be invisible to the eye);

                                    - washwaters from sunscreen manufacture;

                                    - wastewaters from printed circuit board manufacture

                                      (photoresist, etch, solvents)

                                    - other UV absorbing chemicals, for example used in plastics


                                    - organics which absorb in the 254 nm region

                                    - sodium thiosulfate and sulfite solutions;

                                    - un-reacted BOD in the wastewater

                                    - Refer to (Swift, 2005), (Swift 2007)

·        Colour absorbers    - water supply colour from humic substances (this will increase

                                       through the treatment process);

                                    - humic substances from landfill leachate

                                    - abattoirs (blood - colour, lysed cells and iron);

                                    - colours from confectionary manufacture;

                                    - colour from molasses (carbon particles), etc.;

·        Suspended solids    - suspended and colloidal substances not removed in the

                                      treatment process, including low MCRT activated sludge


It is important to ensure that a comprehensive sampling programme is undertaken of the effluent from the wastewater treatment plant, so as to identify minimum UV transmittances. UV transmittance depends on the concentration of suspended solids in the treated effluent, which will be affected by the operation of the secondary treatment process, inversely to the variation of the SVI parameter of the mixed liquor. This is because high dissolved oxygen concentrations lead to greater excretion of exocellular polysaccharides and cause compact floc. This leads to strong floc and reduced straggler particles in the clarified effluent. Low or varying dissolved oxygen concentrations will lead to reduced polysaccharides and higher effluent particle concentrations and thus higher turbidity (reference needed).

Seasonal effects from trade wastes (increased organic loadings) leading to increased turbidity and seasonal colour from the colour of seasonal potable water supplies, needs to also be quantified. Sample on different days of the week to take into account weekly production variation (e.g. trade shutdowns on Fridays to Sundays).

For these reasons an effluent sampling programme over 1 to 2 years, is recommended for the purpose of identifying the minimum UV transmittance. Generally there is not a significant diurnal variation in UV transmittance, unless affected by trade wastes. It is recommended that this is checked.

When carrying out a source analysis programme of trade wastes, use Beer’s Law to represent the contributions of n different wastes to the UV absorbance (this does not work for UV transmittance calculations – only those in absorbance (base 10 and base e units)):


Equation 16

Or for bulk absorbance measurements of m different dischargers:


Equation 17


            I/I0       = UV absorbance of the mixture of wastewaters

            ei          = the extinction coefficient for the pure compound at concentration

                           at 1 mg/L or 1 M and 254 nm.

            ci          = the concentration of species i in the wastewater mixture

            Ci j       = the concentration of species i in substream j in the similar units to

                           the ei

            Aj         = the absorbance of stream j (base 10)

            qj         = flowrate of substream j in the same units as Qtot

            Qtot      = flowrate of total substreams or total flowrate of total wastewater mix

It is best to measure the colour or absorbance of the waste after activated sludge treatment, as the activated sludge process will change the colour. A laboratory or large treatment plant could set up a mixer/settler treatment plant to simulate these processes.

As a starting point assume (based upon Eastern Australian observations) the colour of the potable water supply is doubled in the wastewater effluent and 1 Pt/Co unit in the wastewater equates to an absorbance of up to A= 0.00208 [a.u./Pt/Co unit, 254 nm, 1 cm], for the background UV absorbance. Contributions from other continents appreciated.

Use of iron salts in the wastewater treatment process, will add UV absorbing dissolved iron in the effluent, although may precipitate colour. Ferric ions are more UV absorbing than ferrous ions. Alum is not significantly UV absorbing (USEPA, 2006b). Permanganate, hypochlorite, ferrous ions and sulfite are also strong to moderate UV absorbers respectively.

UV transmittance samples should not be taken as a sub-sample of the microbiological samples, if sodium thiosulfate is present as a dechlorination agent. Sodium thiosulfate is a strong UV absorber and will result in false (low percent transmittance) results. 

Typical UV Transmittances are detailed below:

Table 2: Typical absorbances and UV transmittance values from wastewaters (Used with permission from T Asano & Metcalf & Eddy, Water Reuse (Metcalf & Eddy, 2007d)). Not to be amended.

Type of Wastewater

[base 10, cm-1]

UV% Transmittance
[%, cm-1]

Primary effluent

0.55 - 0.30

28 – 50

Secondary effluent

0.35 – 0.15

45 – 70

Nitrified effluent

0.25 – 0.10

56 – 79

Filtered nitrified effluent

0.25 – 0.10

56 – 79

Microfiltration and MBR

0.10 – 0.04

79 – 91

Reverse Osmosis

0.05 – 0.01

89 – 98

McGraw Hill makes no representations or warranties as to the accuracy of this information, including warranties of merchantability or fitness for purpose. McGraw Hill will have no liability to any party for special, incidental, tort or consequential damages arising from the use of this information. Copyright ©2007, McGraw Hill Companies, Inc.                  

Filtered Versus Unfiltered Transmittance

This is the classical definition of the UV transmittance coefficient and is measured on the actual sample. Some equipment suppliers suggest that the filtered UV transmittance should be used, which will achieve a higher UV transmittance and a lower absorbance. Their explanation is that particulate and colloidal substances scatter light from the transmitted beam and do not absorb light.

A number of studies have been carried out to investigate this phenomenon, by bioassay (Qualls, 1983) and actinometry (Linden, 1998) The unfiltered UV transmittance underestimates the true UV transmittance and the filtered UV transmittance overestimates the UV transmittance. The figure below shows the typical contributions of soluble absorbance, particle absorbance and scatter. Improvements can be achieved by using a spectrometer with an integrating sphere to measure all scattered light (Linden, 1998). If such an instrument is not available, it is best to use the unfiltered UV transmittance, as it is closer to the true transmittance, although it is conservative for sizing disinfection units.


Figure 16: Diagram showing contributions of dissolved, particulate and scattered light to UV absorbance.

Classical UV Dose Calculation

Assume we have an in-channel UV disinfection system. The wastewater flows along the channel, with the top water surface level controlled by an outlet extended length weir. The tubular UV lamps are, enclosed in tubular quartz sleeves, arranged in a regular grid (e.g. on a square layout arrangement, with multiple rows and columns) with lamps extending from the floor to the top WSL and across the full width of the channel. We will look at the UV dose provided by one bank of UV lamps. The lamps are oriented with the lamp axis parallel to the flow direction. The UV% Transmittance is x%. From a suppliers chart of average UV intensity versus UV% Transmittance we read off the intensity (Inom) at x% transmittance. This value is then adjusted by the lamp percent UV output (UV%P), adjusted by varying power input, actual fractional fouling level (Ft) and the actual lamp UV power output at the current lamp operating age or hours run (Fp). The design case is the worst case dose or fluence delivered, under the worst case transmittance (say x%), the worst case design fouling level (say Ft) and the lamp fractional output at the end of the lamps’ lives (Fp)


Equation 18



Equation 19

The irradiated volume of the channel is the area where the lamps are placed, which is the channel width x channel wetted depth x lamps UV emitting length (the length between the electrodes – termed the “arc length”). The wetted volume is this irradiated volume, less the volume taken up by the lamps in the quartz sleeves


Equation 20

The irradiated contact time, per bank, is:


Equation 21

The UV dose thus delivered by each bank, under worst-case conditions is thus:


Equation 22

This is the classical method for determining the minimum delivered UV dose.

The Scheible Dose Calculation

The following form of equation should be used when backmixing is significant. This occurs when the lamps are not oriented parallel o the flow lines in the reactor or are not long.

The first equations to model the ultraviolet inactivation of coliforms in treated wastewater effluent were proposed by Scheible (Scheible, 1987), (USEPA, 1986b).


Equation 23


N         = final coliform concentration

No        = initial coliform concentration

u          = wastewater velocity = x /(Ve/Q)

x          = reactor characteristic length

            = the avg length traveled by water while UV exposed

Ve        = reactor liquid side irradiated volume

Q         = wastewater flowrate

E          = reactor dispersion coefficient

            = accounts for non-plug-flow behavior in the reactor    

a          = UV inactivation constant

Inom     = Nominal intensity in the reactor for new lamps and clean sleeves

Fp        = fractional lamp UV output or UV power at end of lamp life compared

                 with initial lamp output

Ft         = fraction of lamp UV output transmitted through the quartz sleeves 

b          = experimental constant

c          = experimental  constant

SS        = wastewater suspended solids concentration

m         = experimental constant

Whilst this is a complex equation, the first grouped term incorporates factors to allow for the dispersion (or back mixing) of the wastewater flowing through the reactor.

The derivation of the back mixing terms is detailed in USEPA (USEPA, 1986c), or generally from a text such as Levenspiel (Levenspiel, 1972).

1.      Prepare a Residence Time Distribution (RTD) for the reactor, say by addition of a step change in some additive, such as a salt solution being run through the unit.

2.      From the trace of the salt in the outlet derive a pulse output curve by differentiating the step curve data.

3.      Calculate the theoretical detention time T=V/Q and the reactor characteristic length

4.      From the pulse data or the curve, compute the following columns of data:

·        ti – individual time measurements of the pulse curve

·        dci/dt – individual concentration slope terms

·        ti.dci/dt – product of the two columns of data

·        ti2.dci/dt – product of time squared and concentration slope curve

5.      Now the following can be generated from the sum of each column and compute the following variables:

·        θ = Σtidci/dt / Σdci/dt

·        σ2 = Σti2dci/dt / Σdci/dt

·        σD2 = σ2 / θ2

·        E/ux ≈ σD2/2 – this is an approximation that applies when E/ux is < 0.12 and the RTD distribution is approximately symmetrical

6.      These values can then be used in the equation above.

The Reduction Equivalent Dose (RED)

The classical method of determining the UV dose has been described for a simple relatively uniform reactor. It is more difficult to calculate the UV dose for other reactors which have non-uniform velocity profiles and non-uniform UV intensity fields.

Bioassay and chemical actinometric methods are the other methods available for determining the delivered UV dose. Bioassay has become popular. Refer to USEPA (USEPA, 2006c) for full details of the method.

The method for bioassay is to take a batch of typical wastewater and spike it with a microorganism or use indigenous organisms. The spiked water is run through the UV reactor at the desired operating conditions, with inlet and outlet samples taken and assayed for microbiological reduction.

If the organism employed is the targeted organism and the required microorganism log reduction known, then the disinfection capacity of the unit can be stated as a number of log reductions. Given the species of microorganism, the dose delivered by the UV reactor can be calculated.

A more complex, but more accurate, method of determining the delivered dose is the Reduction Equivalent Dose (or RED) method. Doses computed by this method are referred to as Reduction Equivalent Doses (REDs), instead of labeling them simply as the delivered dose.

UV reactors measured by this method are said to be validated.

Two definitions:

Target Organism – the organism which the UV reactor is to inactivate, under operating conditions. There may be more than one target organism.

Challenge Organism – the organism used to measure the inactivation performance of the UV reactor, by a performance trial. The challenge organism is often a bacteriophage, or bacillus subtilis spores (refer USEPA Table 5.2 for the usual range of challenge organisms and their UV inactivation doses).

The RED procedure is to carry out parallel trials on the UV reactor and with the Collimated Beam device. The collimated beam tests are carried out, in the proposed wastewater, spiked with the challenge organism, at approximately six UV doses and a curve developed of delivered dose (calculated by the collimated beam formula) as a function of log inactivation of the challenge organism.

The UV reactor trial is carried out using the proposed wastewater(s), at the proposed operating conditions (flowrate, UV%T and banks operating ranges), spiked with the challenge organism. Dosing with coffee, lignin sulfonic acid, or humic acid concentrates is employed, to adjust the UV transmittance. The challenge organism log reduction at each case is determined by microbiological assay.

The full scale UV unit log reductions are used with the collimated beam calibration to determine the RED, at each of the full scale UV unit operating conditions. The range of REDs are submitted for log-log least squares regression, in equations of the form:

The Calculated Dose approach:

            RED = A x UV%Tb x UV Power Inputc x 1/Flowrated

            Log Red = log A + b.log UV%T + c.log UV% Power Input + d.log (1/Flowrate)

The UV Intensity Setpoint approach:

            RED = A x UV%Tb x  (UVSensor signal/So)c x (1/Flowrate)d

            Log RED = log A + b.log (UV%T) + c.log (S/So) + d.log (1/Flowrate)

Adjustments for RED bias, end of lamp life, and sleeve fouling are then made.

The following RED protocols are available for this validation process:

  • USEPA “Ultraviolet Disinfection Guidance Manual for the Long Term 2 Enhanced Surface Water Treatment Rule”, November 2006. (potable water standard)[1]
  • NWRI/AwwaRF Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse, 2nd Edition, May 2003.
  • NSF, USEPA, “Environmental Technology Verification Protocol, Verification Protocol for Secondary Effluent and Water Reuse Disinfection Applications”, October 2002.[2]

The following three standards define measured flow rate, UV intensity, and lamp status for a Bacillus subtilis RED of 40 mJ/cm2 in potable water.

  • DVGW W294 (2006)[3].
  • ÖNORM M 5873-1, “Plants for the disinfection of water using ultraviolet radiation - Requirements and testing - Low pressure mercury lamp plants”, March 2001.
  • ÖNORM M 5873-2 “Plants for the disinfection of water using ultraviolet radiation - Requirements and testing - Part 2: Medium pressure mercury lamp plants”, August 2003.[4]

For discussion of validation details, please refer to the chapter “Practicalities- Validation” at the end of the Practicalities sections.

Non-Ideal Reactors: Adjustments to the RED Due to Differences in UV Sensitivity

All UV reactors are non-ideal; in that

  • there are regions of dead space;
  • there are regions of non-uniform velocity;
  • there is absence of complete dispersion perpendicular to the flow direction; and
  • there is a non-uniform UV intensity field, through the reactor.

Conventional wastewater UV reactors, such as the in-channel, horizontal lamp, types, have a relatively uniform velocity and UV intensity profile, whereas the Medium Pressure closed pipe reactors have velocity and UV intensity distributions which exhibit substantial variation. In the year 2000, a paper was presented by Wright and Lawryshyn (Wright, 2000a), which explained that non-ideal flow through real reactors resulted in differences in observed RED, when the challenge and target organism were of differing UV sensitivities (UV sensitivity = dose for 1 log10 reduction). See also (USEPA, 2006g).

<Content to be inserted when permissions received>

The fundamental lesson from this paper is that:

When carrying out a Reduction Equivalent Dose determination for a reactor, it is best to use a challenge microorganism that has a similar UV sensitivity to the target microorganism.

For example, when sizing a reactor for a target organism of Cryptosporidium (D10=3 mJ/cm2) use say the T7 phage (D10=3.6 mJ/cm2) as challenge organism for bioassay, instead of say MS2 phage (D10=16 mJ/cm2), which is readily employed for this sort of work.

Adjustments for Polychromatic Medium Pressure Lamps

Contributions appreciated.

Adjustments must be made to dose-response relationships, because medium pressure lamps emit light at a number of wavelengths. The general approach has been to apply a factor relative to the maximum, or 254 nm, absorbance of light by DNA, for each of the wavelengths in MP lamp spectrum (Bircher, 2000), (Blatchley III, 2000).

Wright (2000b) published the following formula for describing the effective output of a medium pressure lamp, used in a collimated beam experiment:


Equation 24



Equation 25


Wright found that for a broad range of wastewaters (UV%T: 25% to 77% cm-1) it was only necessary to compute the Morowitz factor at 254 nm, and not as a function of wavelength. This introduced an error of only 1.3% at a UV%T of 25 to 35%. He also noted that spectral changes in lamps, due to differences in the mercury content do not introduce significant differences in the MP factor of individual lamps. Thus he concluded the average spectrum for a particular lamp model can be used.

A similar approach was taken in the USEPA-UVDGM (USEPA, 2006d). See also (Wright, 2002). The application resulted in the presentation of two tables for the Bpoly or the polychromatic factor for drinking water applications, when Non-germicidal sensors are employed. The polychromatic effect comprises the following four factors:

  • The action spectra of the Challenge Organism and the Target Organism;
  • UV absorbance of the water used for the validation and at the operating site;
  • UV output of the MP UV lamps at the validation site and at the operating site;UV transmittance of the quartz sleeves used for the validation and at the operating site.

USEPA presents a table of overall action spectra differences for an assumed MP UV lamp, based on the formula:


Equation 26


            PG        = Germicidal output of the MP UV lamp [W/cm],

            P(λ)     = Lamp output at wavelength λ nm [W/nm],

            G(λ)     = Relative UV sensitivity at wavelength λ nm, [cm-1]

            Δλ        = Wavelength increment [nm]

The table is based on the spectra of Cryptosporidium as the target organism.  The document notes that, quantitatively, MS2 and B. subtillis did not deviate significantly from the Cryptosporidium, however phage φx174 did differ by 16%. Currently there is insufficient data to determine this parameter for other target organisms. The following alternative formula is presented for computing the factor for different challenge organisms:


Equation 27


            CFas     =  Correction factor for action spectra [dimensionless]

            k          =  Inactivation rate constant (slope of the Inactivation vs Dose graph),

                               for the MP and LP lamp collimated beam devices and for the

                               challenge organism and MS2 respectively [cm2/W]

The RED produced by the validation should be divided by this correction factor.

Adjustments for the UV absorbing water comprise an adjustment between the validation of synthetic water and operation under true drinking water quality conditions. The difference is the difference between the absorption spectrum of the synthetic UV transmittance additive and the natural water. Corrections are provided for using coffee or lignin sulfonic acid. The difference depends on the pathlength between the Non-germicidal UV sensor and the lamp. Only a pathlength of 1 cm provides a true calculation through the UV absorbance. The table in the manual provides for a number of lamp to sensor spacings, for drinking water applications (UV transmittances). No adjustment is needed for a germicidal sensor, as it takes into account the germicidal spectrum and the correction is thus minor.

Bircher (Bircher, 2000) investigated the inactivation of E. coli using a MP collimated beam device. They used an interference filter and a sheet of Pyrex glass to restrict the MP lamp to individual MP radiation bands. Intensity was checked by radiometer and by actinometry. They confirmed previously derived ratios of the action spectra at 230:254:280 nm (normalized to 254 nm) of 0.53:1.0:0.74, as previously measured by Gates (Gates, 1930). They noted a better UV transmittance at the wavelengths above 235 nm, and that the relative effectiveness of the light above 254 nm to 280 nm was greater than below 254 nm.

Linden (Linden, 2006) found that between wavelengths 250 and 275 nm, inactivation of Cryptosporidium parvum, by these wavelengths were essentially equally germicidal.

Computational Fluid Dynamics Models and UV Dose Calculation

Contributions appreciated regards CFD for determining the UV Dose.Put Hydraulic issues in the section, “Practicalities – Hydraulics”, below.

N.B.  This article is divided into 2 parts. Please Click Here to read Part 2


This is part 1 of Wastewater Ultraviolet Disinfection.  All references to this article relate to both part 1 and  2. For a full list of references , please Click Here

Related Publications

Wastewater Treatment: Advanced Processes and Technologies - DG Rao, R Senthikumar, J Anthony Byrne, S Feroz
 Publication Date: Jul 2012 - ISBN - 9781780400341

Evolution of Sanitation and Wastewater Technologies through the Centuries - Andreas N. Angelakis, Peter A. Wilderer and Joan Bray Rose
 Publication Date: Mar 2014 - ISBN - 9781780404844

Related Articles


Wastewater Ultraviolet Disinfection 2

Wastewater Disinfection

Wastewater Lagoon Disinfection

Wastewater Chlorine Disinfection

Wastewater Ozonation

Wastewater Chlorine Dioxide Disinfection

Other Wastewater Disinfectants

Featured Articles

WaterWiki Newsletter

Share the WaterWiki

Share the WaterWiki!

Add this button to your site to share the WaterWiki with your users. Just copy and paste the HTML code below into your website.