E. Nebot Sanza,, I. Salcedo Da´ vilaa, J.A. Andrade Balaob, J.M. Quiroga Alonsoa

Available at www.sciencedirect.com else V ti dr nm University of Ca´diz, Cadiz, Spain bAguas de Jerez, Empresa Municipal, Spain a r t i c l e i n f o & 2007 Elsevier Ltd. All rights reserved. recognized (Kashimada et al., 1996). isms n to light- dent This phenomenon therefore represents a potential disadvantage ARTICLE IN PRESS WAT E R R E S E A R C H 41 ( 2007 ) 3141 – 3151 �Corresponding author. Tel.: +34 956 016198. The inactivation of microorganisms by far-UV light (UV-C: 200–280nm) is effected through the formation of lesions in the genomic DNA of the organisms (Friedberg et al., 1995; Harm, 1980). The major lesion induced by germicidal UV-C light for UV-C disinfection methods in water reclamation (Weinbauer et al., 1997; Oguma et al., 2001; Liltved and Landfald, 1996). Photoreactivation is the phenomenon by which inactivated microorganisms recover activity through the repair of pyrimidine 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.04.008 E-mail address: enrique.nebot@uca.es (E. Nebot Sanz). disinfection to enable wastewater to be reused are widely (dark repair) mechanisms (Friedberg et al., 1995; Harm, 1980). promising alternative to chlorine for the disinfection of wastewater (Lindenauer and Darby, 1994). Reclaimed waste- water reuse for agricultural purposes and golf course irriga- tion is now expanding, and the advantages of applying UV and therefore result in inactivation of the microorgan (Oguma et al., 2001). However, many organisms are know possess the ability to repair their DNA damage by dependent (photoreactivation) as well as light-indepen 1. Introduction Ultraviolet (UV) light is increasingly being considered as a (254nm) is the formation of pyrimidine dimers (Harm, 1980; Rothman and Setlox, 1979; Tyrrell, 1973). The presence of lesions induced by UV-C would inhibit the normal replication of DNA Article history: Received 9 October 2006 Received in revised form 17 February 2007 Accepted 11 April 2007 Available online 25 May 2007 Keywords: Ultraviolet irradiation Photoreactivation Dark repair Wastewater Faecal microorganisms Model a b s t r a c t The increased use of UV radiation as a wastewater treatment technology has stimulated studies of the repair potential of microorganisms following treatment. In this study, samples of unfiltered secondary effluent were irradiated with seven levels of UV-C doses (50–200mWs/cm2) from six low-pressure lamps in an open-channel UV disinfection system. Following irradiation, samples were incubated at 20 1C under photoreactivating light or in darkness. Samples were analysed for 240min following incubation. The logistic model is proposed to explain the relation between photoreactivation and the UV-C dose received by the microorganisms. That model accurately fitted the data obtained in photoreactivation experiments, permitting interpretation of the estimated kinetic parameters: Sm and k2. In the experiments carried out in darkness, a slight reactivation is observed (o0.1%), followed by a decay period in which survival decreases. In order to model this last period, a modification was made to the logistic model by including a term of mortality that assumes a zero-order kinetic. The parameters Sm and k2, in both photoreactivation and darkness, show an exponential dependence on the UV-C inactivat- ing dose. It is possible to predict their values, and hence the reactivation curve, from the equations proposed in this work. journal homepage: www. Modelling of reactivation after U dose on subsequent photoreac E. Nebot Sanza,�, I. Salcedo Da´vilaa, J.A. An aDepartment of Chemical Engineering, Food Technologies and Enviro vier.com/locate/watres disinfection: Effect of UV-C vation and dark repair ade Balaob, J.M. Quiroga Alonsoa ental Technologies, Faculty of Marine and Environmental Sciences, ARTICLE IN PRESS No)�100, (%) S survival ratio at time t, (%) ( 2 dimers in the DNA under near-UV (UV-A) and visible light (310–480nm) with the enzyme photolyase and without excising the distorted region (Walker, 1984; Oguma et al., 2001; Liltved and Landfald, 1996). The dark repair mechanism, as its name suggests, can repair the damaged DNA without light. This mechanism is a multi-enzyme repair process involving the excision of dimers. Many studies have demonstrated the possibility for photo- reactivation or dark repair of UV-C damaged microorganisms, enabling regrowth of the microbial population under certain conditions, thus reducing the efficacy of UV-C inactivation (Chan and Killick, 1995; Lindenauer and Darby, 1994; Whitby and Palmateer, 1993; Eker et al., 1991; Harris et al., 1987; Levine and Thiel, 1987; Machida et al., 1986). Systematic quantitative study of photoreactivation, the more important of the two mechanisms, has suggested a two-step reaction scheme (Harm, 1980): Step 1: Formation of a complex between a photoreactivation enzyme (PRE) and the dimer to be repaired. This step does not require light, but is dependent on temperature, pH and ionic strength (Lindenauer and Darby, 1994). Step 2: Release of PRE and repaired DNA. The restoration of the dimer to its original monomerized form is absolutely dependent upon light energy intensity. The reaction occurs in less than a millisecond; consequently the limiting step of the whole reactivation process is the formation of the PRE–dimer complex. An extended period of exposure to photoreactivat- ing light would enable the release of PRE that would then be Nomenclature k1 reactivation first-order reaction rate constant, min�1 k2 reactivation second-order reaction rate constant, min�1 M mortality zero-order reaction rate constant, min�1 No concentration of microorganisms before disinfec- tion, FCU/100mL WAT E R R E S E A R C H 413142 available to form new complexes (Step 1). The effect of temperature on the reactivation phase is still little studied. Chan et al. (1995) investigated the effect of salinity and temperature on the reactivation of Escherichia coli in a marine environment. They found that the effect of salinity is greater than temperature, although an Arrhenius tendency was confirmed for the rate constants calculated. The factors that influence the rate and extent of reactivation are beginning to emerge through scientific research, but not enough is known at present to make quantitative predictions in most cases. It was showed (Kalisvaart, 2004) that medium- pressure UV lamps produce a broad, ‘‘polychromatic’’ spectrum of UV wavelengths that inflict irreparable damage not only on cellular DNA, but on other molecules, such as enzymes, as well. However low-pressure lamps emit a single wavelength peak which only affects DNA. Therefore, the microorganism reactiva- tion is more difficult with medium-pressure lamps. Several studies have noted that if reactivation is observed in a microbial species, the extent of reactivation is often inversely related to the applied UV-C dose (Baron and Bourbigot, 1996; Lindenauer and Darby, 1994). The repair was generally found to be higher at low doses (Hu et al., 2005). In an interesting recent review (Hijnen et al., 2006), it is explained that reactivation entails a lower inactivation kinetic, and itmeans that higher UV-C fluence are required to obtain the same level of inactivation. Quantita- tive data showed a 2.8–4.6 higher UV fluence requirement for 1–3 log inactivation of Legionella pneumophila (Knudson, 1985). Few papers focus on modelling the reactivation processes: most simply describe qualitative studies of the process (Harris et al., 1987; Lindenauer and Darby, 1994; Hassen et al., 2000; Oguma et al., 2004). Other authors have tried to model jointly the inactivation and reactivation phases (Tosa and Hirata, 1999; Beggs, 2002), by comparing experiments with and without reactivation. Kashimada et al. in 1996 proposed a model to predict independently the reactivation phase, but the experiments carried out were insufficient to be generalized. Therefore, it is pertinent to investigate new kinetic models that permit reactivation processes to be predicted, and provide a better understanding of the factors affecting this interesting phenomenon. Thus, in this study, photoreactivation and dark repair of three bacterial indicators, total coliforms (TE), faecal coliforms (FC) and faecal streptococci, were investigated in order to develop a kinetic model which allows prediction of their reactivation after UV disinfection depending on the UV-C dose Sm maximum survival ratio (Nm/No)�100, (%) Nd concentration of microorganisms after disinfec- tion (before reactivation), FCU/100mL Nm maximum concentration of microorganisms, FCU/100mL Nr concentration of microorganisms at time t after the beginning of the reactivation phase, FCU/ 100mL So survival immediately after UV disinfection (Nd/ 007 ) 3141 – 3151 applied. For this purpose, the relationship between the repair and the UV-C irradiation dose was particularly examined. The extent and rate of reactivation were also studied. 2. Materials and methods 2.1. UV-C irradiation UV-C irradiation treatment was performed with a 5.0m3/h horizontal-lamp open-channel UV disinfection system (Trojan Technologies, Spain, S.L.) and six low-pressure high- intensity mercury UV lamps (Philips 30W UV-C at 254nm). The UV channel received the water from the unfiltered secondary effluent of the Municipal Wastewater Treatment Plant of Jerez de la Frontera (Spain). ARTICLE IN PRESS (2 Bacteria were exposed to seven levels of UV-C fluence (50, 75, 100, 125, 150, 175 and 200mWs/cm2). The UV-C fluence (mWs/cm2) applied was calculated as a product of the average UV fluence rate in the reactor (mW/cm2) and the irradiation time (s). The average UV fluence rate was calculated by the Point Source Summation (PSS) method (Ho et al., 1998; Braunstein et al., 1996; USEPA, 1986, 1992; Qualls et al., 1989). The exposure time was calculated from the channel volume and the influent flow rate, after first ensuring that the plug flow condition existed in the channel. 2.2. Repair conditions After UV-C irradiation, the water sample was divided and transferred into two 500mL glass Erlenmeyer flasks (95% transparent for 360nm light). One of the two Erlenmeyer flasks was thermostated in a controlled-environment incubator (FOC 225E, Refrigerated Incubator, VELP Scientifica), which was equipped with one fluorescent lamp (3.7W, PHILIPS TLD), at six different temperatures: 5, 10, 15, 20, 25 and 30 1C (photoreactiva- tion). The range of the lampwavelengthswas 310–420nmwith a broad peak at 360nm. Irradiation periods were in the range of 30–240min (minimum andmaximum values). The UV-A fluence rate of the fluorescent lamp was 0.1mW/cm2 at 360nm at the sample surface, estimated by the PSS method and the distance between the samples and the lamp. The other Erlenmeyer flask was covered immediately with aluminium foil and incubated simultaneously at the same temperature for 240min (dark repair). Concentration of bacteria was measured every 30min taking samples from each Erlenmeyer flask with a pipette. 2.3. Enumeration of microorganisms Three bacterial indicators of microbiological contamination have been analysed: TC, FC and Streptococcus faecalis (SF). All the microorganisms were analysed according to Standard Methods for the Examination of Water and Wastewater (APHA et al., 1992) by using the membrane-filter technique. TC were cultured on MF-Endo agar and incubated at 35 1C for 24h. FC were determined on m-FC agar and 24h incubation at 44.5 1C. Finally, S. faecalis were grown on KF agar at 37 1C for 48h. After the incubation period, bacterial colonies were counted and the results calculated as colony forming units per 100mL of sample (CFU/100mL). For each microorganism and each experimental condition used, tests were repeated at least three times and mean values were obtained from these repeated experiments. Experiments were repeated three times independently for each bacterium and experimental condition used. Standard deviations of triplicates are not presented on the graphs, in the interests of clarity. When standard deviation was disproportionate (CV420%), data were rejected. The rates of reactivation were assessed by determining microorganism survival from microbial numbers before disinfection and after reactivation phenomenon. 2.4. Modelling the reactivation kinetics Reactivation is frequently expressed as a function of the WATER RESEARCH 41 survival ratio in respect of the initial microorganism concen- tration existing before the inactivating treatment. Therefore, the survival values were calculated using the following equation: S ¼ Nr No � 100, (1) where S is the survival ratio at time t, No is the concentration of microorganisms before disinfection and Nr is the concen- tration at time t after the beginning of reactivation. A typical inactivation–reactivation curve as a function of time is shown in Fig. 1. In that figure it is possible to differentiate the various phases of the process: exponential UV inactivation, reactivation process which includes an induction period, growth phase, stabilization phase and decay period. As stated, the reactivation can occur by two mechanisms according to the exposure of the samples to light or in darkness. 2.5. Photoreactivation kinetic Kashimada et al. (1996) proposed an asymptotic model, assuming the photoreactivation phenomenon follows a saturation-type first-order reaction, as: dS dt ¼ k1 � ðSm � SÞ, (2) where Sm is the maximum survival ratio and k1 the first-order reactivation rate constant. In the model, the term (Sm�S) acts as driving force for the reactivation. As the survival ratio, S, is reaching its maximum value (Sm), the process decelerates showing an asymptotic tendency. Reactivation curves have been obtained by means of the experiments described below. In Fig. 2a the typical asympto- tic-sigmoidal shape of these curves can be observed. In the induction period, the curve suggests imperceptible reactiva- tion; then a rapid exponential growth can be seen, and finally a stabilization period is reached when growth ceases. After the application of the Kashimada et al. (1996) model, we observed that it did not fit the data correctly, mainly at the beginning of the curve, when an induction period is observed (Fig. 2b). Therefore we decided to modify the model, but without increasing the number of parameters. The new model is represented by the following equation: dS dt ¼ k2ðSm � SÞ � S, (3) where k2 is the new growth second-order reactivation rate constant. This relationship is simply a combination of the second-order equation and the driving force concept em- ployed by Kashimada et al. (1996). The equation is really not new, because it coincides, in its mathematical form, with the logistic equation proposed by Verhulst in 1838 for interpreting biological population growth. Nevertheless, the originality of our work lies in its innovative application to microorganism reactivation prediction. The model has the advantage that both kinetic parameters: Sm and k2, have a clear physical significance. On the one hand, Sm is the maximum limit of the microorganisms’ survival by reactivation, and on the other hand, k2 represents the rate at which that value is reached. It 007) 3141– 3151 3143 can be seen in Fig. 2b that this proposed model correctly fits the experimental data. ARTICLE IN PRESS Fig. 2 – (a) Typical photoreactivation curve. (b) Curves from the model proposed by Kashimada et al. (1996), and the logistic model. Fig. 1 – A typical inactivation–reactivation curve as a function of time, where No is the concentration of microorganisms before disinfection, Nd, after disinfection but before reactivation and Nr at time t after the beginning of the reactivation phase. Nm is the maximum concentration of microorganisms reached by reactivation. WAT E R R E S E A R C H 41 ( 2007 ) 3141 – 31513144 By the integration of Eq. (3), the following is obtained: ln S So � Sm½ � So S� Sm½ � ¼ k2 � Sm � t, (4) where So is the survival immediately after UV disinfection (Nd/No). From Eq. (4), it is possible to express the variable S as a function of the kinetic parameters k2, Sm, So and time (Eq. (5)) and, as a result, we can easily obtain the two parameters, Sm and k2 by non-linear regression. S ¼ Sm 1þ Sm=So � 1 � � � e�k2 �Sm �t . (5) This Eq. (5) allows the photoreactivation curve over time to be simulated (Fig. 3). ARTICLE IN PRESS WATER RESEARCH 41 (2007) 3141– 3151 3145 Fig. 3 – Survival ratio versus time of exposure to photoreact (50–200mWs/cm2). Experimental data and prediction of the log ivating light for different inactivating UV-C doses applied istic model. 2.6. Dark repair kinetic After carrying out the dark repair experiments, it was observed that the curve of microorganisms’ survival versus time showed, after a low and brief reactivation period, a decay phase, not detected in photoreactivation experiments. We concluded that in darkness, the reactivation did occur but to a less extent than in illumination conditions, and then the survival commenced a decreasing tendency (Fig. 4). Therefore the model proposed (Eq. (5)) did not fit these data, and it was ARTICLE IN PRESS WAT E R R E S E A R C H 41 ( 2007 ) 3141 – 31513146 Fig. 4 – Survival ratio versus time of exposure to darkness for d Experimental data and prediction of the logistic model. ifferent inactivating UV-C doses applied (50–200mWs/cm2). apparently low percentage, for a No of 10 6 FCU/mL, would 4 ARTICLE IN PRESS le 1 – K in e ti c p a ra m e te rs o f th e lo g is ti c m o d e l a p p li e d to p h o to re a ct iv a ti o n e x p e ri m e n ts C d o se s s/ cm 2 ) T C F C S F S m (% su rv iv a l) k 2 (% m in )� 1 (o b s- p re d )2 (% su rv iv a l) 2 r2 S m (% su rv iv a l) k 2 (% m in )� 1 (o b s- p re d )2 (% su rv iv a l) 2 r2 S m (% su rv iv a l) k 2 (% m in )� 1 (o b s- p re d )2 (% su rv iv a l) 2 r2 5 0 0 .9 4 5 0 .0 3 9 6 .7 1 E -0 2 0 .9 6 2 0 .6 4 4 0 .0 3 1 1 .3 9 E -0 2 0 .9 7 6 0 .2 6 8 0 .1 5 1 1 .3 6 E -0 3 0 .9 8 4 7 5 0 .5 5 1 0 .1 0 7 1 .8 7 E -0 2 0 .9 7 2 0 .3 3 9 0 .1 0 2 3 .4 3 E -0 3 0 .9 8 2 0 .2 1 4 0 .2 2 0 7 .2 7 E -0 4 0 .9 9 0 1 0 0 0 .2 9 8 0 .1 3 4 4 .5 2 E -0 3 0 .9 7 5 0 .1 3 1 0 .2 0 4 4 .3 6 E -0 4 0 .9 8 2 0 .0 5 3 0 .6 8 6 9 .5 5 E -0 5 0 .9 6 4 1 2 5 0 .1 9 6 0 .1 6 2 2 .8 8 E -0 4 0 .9 9 7 0 .1 0 5 0 .1 4 6 2 .9 3 E -0 4 0 .9 6 9 0 .0 4 0 0 .7 9 1 3 .2 7 E -0 5 0 .9 7 9 1 5 0 0 .1 2 0 0 .8 5 9 1 .7 0 E -0 3 0 .9 3 7 0 .0 5 4 0 .4 3 5 1 .0 3 E -0 4 0 .9 6 8 0 .0 2 2 2 .1 7 4 2 .5 5 E -0 5 0 .9 3 9 1 7 5 0 .1 0 3 0 .2 1 5 6 .6 4 E -0 4 0 .9 6 1 0 .0 3 3 0 .8 5 8 5 .8 9 E -0 5 0 .9 4 4 0 .0 1 4 4 .9 5 5 1 .0 5 E -0 5 0 .8 5 6 2 0 0 0 .0 2 9 0 .8 3 0 2 .7 6 E -0 5 0 .9 7 7 0 .0 1 2 1 .3 0 1 3 .4 9 E -0 6 0 .9 3 2 0 .0 0 8 7 .7 0 3 2 .7 3 E -0 6 0 .8 0 7 (2007) 3141– 3151 3147 produce a reactivation of 10 colonies, which could cause serious health and environmental problems. On the other hand, if the value of k for the different microorganisms is compared (Table 1), it can be seen that TC and FC have values similar to each other but lower than SF. This means that SF reaches the maximum survival ratio sooner than TC and FC (see Fig. 3), although this maximum is lower than those reached by TC and FC. Therefore, k2 depends on particular biochemistry repair mechanisms that vary between the different microorganisms. In Table 2 and Fig. 5, the consistent behaviour of Sm and k2 with the UV-C dose can be seen. On the one hand, Sm (the asymptotic limit of the microorganism