An Innovative Photoreactor, FluHelik, To Promote UVC/H2O2 Photochemical Reactions: Tertiary Treatment of an Urban Wastewater

1 An Innovative Photoreactor, FluHelik, To Promote UVC/H2O2 1 Photochemical Reactions: Tertiary Treatment of an Urban 2 Wastewater 3 Jonathan C. Espíndola1,2, Raquel O. Cristóvão1,*, Sara R.F. Araújo1, Teresa Neuparth3, 4 Miguel M. Santos3,4, Rosa Montes5, José B. Quintana5, Rosario Rodil5, Rui A.R.5 Boaventura1, Vítor J. P. Vilar1,* 6 7 1Laboratory of Separation and Reaction Engineering - Laboratory of Catalysis and Materials 8 (LSRE-LCM), Department of Chemical Engineering, Faculty of Engineering, University of 9 Porto, Rua Dr. Roberto Frias, 4200-465, Porto, Portugal 10 2CNPq - National Council for Scientific and Technological Development, Brazil 11 3CIMAR/CIIMAR - LA, Interdisciplinary Centre of Marine and Environmental Research, 12 Avenida General Norton de Matos S/N, 4450-208 Matosinhos, Portugal 13 4FCUP – Department of Biology, Faculty of Sciences, University of Porto, Rua do Campo 14 Alegre, 4169-007 Porto, Portugal. 15 5Department of Analytical Chemistry, Nutrition and Food Sciences, IIAA—Institute for Food 16 Analysis and Research, Universidade de Santiago de Compostela, Constantino Candeira S/N, 17 15782 Santiago de Compostela, Spain 18 19 20 *Corresponding authors:21 R.O. Cristóvão: 22 Tel: +351 22 041 3606; E-mail: [email protected] 23 Vítor J. P. Vilar: 24 Tel: +351 91 825 7824; E-mail: v[email protected] 25 Fax: +351 22 508 1674 26 This is the postprint (accepted manuscript) version of the article published in Science of the Total Environment. http://dx.doi.org/10.1016/j.scitotenv.2019.02.335 © 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ Graphical Abstract *Graphical Abstract Highlights  The FluHelik photoreactor proved to enhance the OTC oxidation by UVC/H2O2;  The FluHelik design showed superior performance than conventional Jets photoreactor;  FluHelik reactor + UVC/H2O2 effectively reduced CECs complying with Swiss legislation;  FluHelik reactor + UVC/H2O2 effectively reduced CECs toxicity to zebrafish embryos;  The FluHelik scale-up proved to be feasible employing several reactors in series. *Highlights (for review : 3 to 5 bullet points (maximum 85 characters including spaces per bullet point) 2 Abstract 27 An innovative photoreactor, FluHelik, was used to promote the degradation of 28 contaminants of emerging concern (CECs) by a photochemical UVC/H2O2 process. 29 First, the system was optimized for the oxidation of a model antibiotic, oxytetracycline 30 (OTC), using both ultrapure water (UPW) and a real urban wastewater (UWW) 31 (collected after secondary treatment) as solution matrices. Following, the process was 32 evaluated for the treatment of a UWW spiked with a mixture of OTC and 10 different 33 pharmaceuticals established by the Swiss legislation at residual concentrations (CECs 34 < 660 g L-1). The performance of the FluHelik reactor was analyzed both at lab and 35 pre-pilot scale in multiple and single pass flow modes. 36 The efficiency of the FluHelik photoreactor, at lab-scale, was evaluated at different 37 operational conditions (H2O2 concentration, UVC lamp power (4, 6 and 11 W) and flow 38 rate) and further compared with a conventional Jets photoreactor. Both photoreactors 39 exhibited similar OTC removal efficiencies at the best conditions; however, the 40 FluHelik reactor showed to be more efficient (1.3 times) in terms of mineralization 41 when compared with the Jets reactor. Additionally, the efficiency of the UVC/H2O2 42 photochemical system using the FluHelik photoreactor in reducing the toxicity of the 43 real effluent containing 11 pharmaceuticals was evaluated through zebrafish (Danio 44 rerio) embryo toxicity bioassays. FluHelik scale-up from laboratory to pre-pilot to 45 promote UVC/H2O2 photochemical process proved to be feasible. 46 47 Keywords: FluHelik photoreactor; UVC/H2O2; CECs; Urban wastewater; Zebrafish 48 embryo toxicity test. 49 3 Introduction 50 UVC/H2O2 photochemical process is based in the homolytic cleavage of H2O2 51 molecules by UVC light, resulting in highly reactive species (HO), able to eliminate an 52 extensive variety of pollutants from water. However, due to the low values of molar 53 absorption coefficient of H2O2 at 254 nm, high H2O2 or UV dose is required to achieve 54 an efficient performance (Krishnan et al., 2017). On the other hand, urban wastewaters 55 composition can reduce significantly the system efficiency due to the presence of light 56 absorbing species (NOM, nitrite, etc.) (Diya’uddeen et al., 2011). The efficiency of 57 UVC/H2O2 photochemical process is also largely influenced by the reactor 58 hydrodynamics regime, which must promote an uniform UV fluence within the reactor 59 (Cambié et al., 2016). Here enters the importance of the reaction mixing conditions for 60 the treatment effectiveness (Karpel VelLeitner et al., 1997). Several commercial 61 reactors incorporate different mixing systems, such as static mixers and conical 62 dispersion components, to improve the degree of mixing inside the reactor, promoting 63 the contact between reagents/pollutants and the emitted UVC photons. Normally, the 64 irradiation source is located in the most turbulent zone of the reactor (Masschelein, 65 1992). Nowadays there is a great variety of photoreactors with diverse geometries 66 leading to different hydrodynamics. Generally, the photoreactor comprises a cylindrical 67 shell of stainless steel housing a concentric quartz sleeve filled with an UVC lamp and 68 the water to be treated flows between the concentric tubes (annular reactor). However, 69 the irradiation source may be also external, such as the parallel plate reactors or 70 cylindrical reactors reported by Noël (2017)). The mixing and irradiation conditions can 71 also be improved through introduction of a multi-lamp design (Boyjoo et al., 2014) or 72 by usage of rotating annular reactors (Subramanian and Kannan, 2010) or spinning disc 73 reactors (Yatmaz et al., 2001). Therefore, even with the same lamp type and intensity, 74 4 reagents/pollutants dosages and similar flow rates, the photons dissemination as well 75 and the pollutants removal may be completely different (Caris, 2011). Although 76 photochemical reactors have been already applied in water/wastewater treatment plants, 77 the process is not widely disseminated because of the inherent limitations that it 78 presents, namely in terms of energy costs and efficiency. Therefore, breakthrough 79 designs for photoreactors are required to achieve a cost-effective treatment solution (Su 80 et al., 2014). 81 The present work focuses on the application of an innovative photoreactor, FluHelik, in 82 the removal of contaminants of emerging concern (CECs) from urban wastewaters, as a 83 polishing step, using a UVC/H2O2 photochemical process. The FluHelik photoreator 84 comprises a cylindrical shell of stainless steel, internally polished, with inlet and outlet 85 pipes located perpendicularly to the fluid flow and tangentially to the shell in horizontal 86 plane and at the top in opposite sides. A concentric inner quartz sleeve houses an UVC 87 lamp. This configuration induces unique fluid dynamics (high degree of mixing) and 88 irradiation properties (a more homogeneous UV radiation distribution) by promoting a 89 helical motion of the fluid around the UVC lamp. First, synthetic solutions of OTC or 90 UWW fortified with OTC were used as reaction matrices. Two configurations of 91 photoreactors were employed: FluHelik and Jets (four inlet and four outlet pipes placed 92 in parallel with the fluid flow direction at the ends of the tube). Process efficiency was 93 evaluated as a function of several operational conditions, namely: (i) recirculation flow 94 rate, (ii) H2O2 concentration and (iii) UVC lamp power. In addition, the feasibility of 95 implementing the FluHelik reactor for the UVC/H2O2 process was tested both at 96 laboratory and at pre-pilot scale either in multiple or single pass flow mode. Finally, the 97 treatment of a real urban wastewater matrix spiked with a mixture of 11 CECs, at 98 5 residual concentrations, was evaluated using the FluHelik photoreactor and the 99 UVC/H2O2 system. 100 Considering that oxidation by-products might be more toxic and/or persistent than the 101 parent compounds, toxicological studies are needed to determine their deleterious 102 effects on ecosystems and human health. Zebrafish (Danio rerio) have been widely used 103 in Fish Embryo Toxicity (FET) Tests to assess the toxicity of several priority pollutants. 104 This embryonic bioassay has high sensitivity and low cost. Furthermore, zebrafish 105 embryos are translucent which allows for the monitoring of embryo development under 106 a stereomicroscope (Macedo et al., 2017; Zhang et al., 2015). This bioassay has recently 107 been proposed by the OECD as an alternative to classical acute fish toxicity tests 108 (Lammer et al., 2009), and is an appropriate tool to assess the decrease in toxicity after 109 treatment of wastewaters contaminated by emerging pollutants. 110 In this sense, embryo toxicity bioassays with zebrafish (Danio rerio) were used to 111 evaluate the initial effluent toxicity and possible attenuation of the toxic effect after the 112 UVC/H2O2 treatment by using the FluHelik reactor. 113 2. Materials and methods 114 2.1 Chemicals 115 Oxytetracycline hydrochloride (OTC, C22H24N2O9.HCl, 496.89 g/mol) was supplied by 116 Sigma-Aldrich and used as a model compound. Hydrogen peroxide (Fisher Chemical, 117 purity 49.5% (w/v)) was used as oxidant. Na2SO3 in a Na2SO3-to-H2O2 molar ratio of 118 1:1 was added to CECs and dissolved organic carbon (DOC) samples for H2O2 119 elimination (Jeong et al., 2010). Catalase (Sigma-Aldrich) was added to the samples to 120 eliminate residual H2O2 before performing ecotoxicological quality tests. Ammonium 121 monovanadate (Merck, p. a.) was used as colorimetric reagent to determine H2O2 122 concentration. Sulfuric acid (Pronalab, purity 96%, 1.84 g/cm3) and sodium hydroxide 123 6 (Merck) were used for pH adjustment. Ultrapure water was obtained from a Millipore® 124 Direct-Q system (18.2 MΩ cm resistivity at 25 °C). Real wastewater sample was 125 collected downstream from the secondary treatment of an urban WWTP from Northern 126 Portugal in September 2017. Its physicochemical characteristics, including the CECs 127 residual concentrations detected in the raw effluent, are summarized in Table 1. The 128 ultrapure water and the secondary effluent both spiked with 20 mg OTC L-1 were used 129 as feed solutions. Table 2 shows the 11 pharmaceutical compounds added to the real 130 wastewater. Tricaine (1000 mg g-1) used to anesthetize zebrafish larvae was purchased 131 from Pharmaq. Sodium hydrogen carbonate used as a buffer in the preparation of the 132 anesthetic was supplied by Merck KGaA. All the other chemicals supplied by VWR-133 Prolabo, Sigma-Aldrich, Panreac, Merck, Fisher Scientific and Pronalab were either of 134 HPLC grade or analytical grade. 135 Insert Table 1 136 Insert Table 2 137 2.2 Analytical determinations 138 OTC concentration was followed by HPLC using a VWR Hitachi ELITE LaChrom LC 139 fitted with a Merck LiChrosorb® RP-18 (5 µm) LiChroCART® 125-4 column at 25 °C 140 and a diode array detector (DAD). Low-molecular-weight carboxylic acids (LMWCA) 141 concentrations were determined by ion-exclusion HPLC using the VWR Hitachi ELITE 142 LaChrom LC fitted with a Phenomenex RezexTM ROA-Organic Acid H+ (8%) 300 143 mm × 7.8 mm column at room temperature (25 ºC). A detailed description of OTC and 144 LMWCA analysis is given in Supplementary Material. 145 H2O2 concentration was determined by the colorimetric (λ = 450 nm) metavanadate 146 method (Nogueira et al., 2005). Dissolved organic carbon (DOC), chemical oxygen 147 demand (COD), total dissolved nitrogen, total dissolved iron, total suspended solids 148 7 (TSS), volatile suspended solids (VSS), total phosphorous, pH, temperature and 149 turbidity, as well as inorganic anions and cations concentrations were assessed 150 according to the procedures already described by Moreira et al. (2016)). Conductivity, 151 dissolved oxygen and redox potential were determined by a HANNA Instruments HI 152 9828 Multiparameter meter. 153 CECs determination in water samples, at residual concentrations, was performed in an 154 Acquity UPLC® liquid chromatograph system from Waters (Milford, MA, USA). A 155 sample volume of 45 μL was directly injected into a Luna C18 100A column (50 mm × 156 2 mm, 3µm particle size) supplied by Phenomenex (Torrance, CA, USA) maintained at 157 a constant temperature of 30 °C. The target compounds were separated at a flow rate of 158 0.2 mL min-1 using 0.1% of formic acid in both, Milli-Q water (A) and MeOH (B) as 159 eluents. The applied gradient was as follows: 0–1 min, 0% B; 1–8 min, linear gradient 160 to 100% B; 7–13 min, 100% B and finally 13–20 min, 0% B. The system was interfaced 161 to a XEVO TQD® triple quadrupole mass spectrometer equipped with an electrospray 162 interface (ESI). Nitrogen was used as a nebulizing and drying gas and Argon was used 163 as collision gas. The analytes were determined in the electrospray (positive and negative 164 polarities) and multiple-reaction monitoring (MRM) mode of acquisition. Two MRM 165 transitions were used as quantifier and qualifier for each compound (see Table S1 for 166 detailed information). The method assured limits of quantification (LOQ) between 10 167 and 100 ng L−1 for all the compounds except for azytromicyn (LOQ 1.8 g L-1), see 168 supplementary material Table S1. Quantification was performed by the matrix matched 169 calibration method using standards prepared in treated wastewater in the 1-100 g L-1 170 (2-100 g L-1 in the case of azytromicyn) range (which was checked to be linear, 171 R2>0.99 for all the studied analytes). Repeatability of the determination was checked in 172 14 possible by H2O2 in the absence of radiation, attaining degradations of 20, 30 and 50% 322 after 180 min of reaction using 100, 300 and 500 mg L-1 of oxidant, respectively. As 323 expected, the combination of UVC light with H2O2 improved the OTC degradation 324 under all the studied irradiation intensities (Fig. 2). It is noteworthy that the 325 photochemical oxidation of OTC showed to follow a pseudo-first order kinetic model. 326 For all the UVC lamp intensities evaluated, the OTC removal rates increased with the 327 initial oxidant dose, attaining the highest removal rates (see pseudo-first order kinetic 328 constants, k, in Table 3) with 100 mg L-1 when using the 4 or the 6 W lamps (Fig. 2a 329 and 2b, respectively) and with 500 mg L-1 of H2O2 when the 11 W UVC lamp (Fig. 2c) 330 was employed (within the tested concentrations range). These results suggest an 331 increasing production of hydroxyl radicals (●OH) for growing H2O2 initial contents (H. 332 Baxendale and A. Wilson, 1957). For higher oxidant dosages, the reaction rates 333 remained constant and, for the 11 W lamp there was even a decrease of 1.3 times in the 334 kinetic constant when using 700 mg L-1 of H2O2 (Table 3). In fact, the H2O2 in excess 335 can act as an hydroxyl radicals scavenger (Muruganandham and Swaminathan, 2004). 336 This is supported by the growing H2O2 consumption for rising initial H2O2 doses (Fig. 337 2). 338 Insert Figure 2 339 Insert Table 3 340 Comparing the OTC removals by the UVC/H2O2 system under the best conditions for 341 each of the lamp powers studied (Fig. 2d), no significant differences between the 342 reaction rates (in terms of energy) were observed. For all the systems, an OTC removal 343 above 90% is already achieved with 0.4 kJ L-1. On the other hand, a different behaviour 344 was observed in relation to the OTC mineralization: DOC decays of 13, 50 and 45% 345 after 60 min of reaction were observed at the best conditions when using the 4, 6 and 11 346 15 W lamps, respectively. It is worth mentioning that, during this reaction period, similar 347 oxidant consumptions were noticed for the 4 and 6 W lamps (c.a. 35 mg H2O2 L-1); 348 however, when using the 11 W lamp a 5-fold increase in hydrogen peroxide 349 consumption (higher initial H2O2 dose) was observed. These results prove that the 350 higher H2O2 consumption associated to the 11 W lamp is not related with a higher 351 mineralization, but probably to parasite reactions. Therefore, the 6 W lamp provided the 352 most suitable photon flow for the experimental set-up used: the 4 W lamp showed to not 353 supply the necessary UV dosage and, in turn, using the 11 W UVC lamp a possible loss 354 of the emitted photons is occurring, probably due to the low molar absorption 355 coefficient of H2O2 at 254 nm, requiring higher amounts of H2O2 to absorb all those 356 photons. 357 As above-mentioned, although the total degradation of the parent compound (OTC) is 358 achieved in short reaction times, relatively low mineralization is observed. Actually, for 359 longer reaction times (360 min), a mineralization of 62% was reached (6 W UVC lamp; 360 100 mg H2O2 L-1), consuming 87 mg L-1 hydrogen peroxide, a higher value than the one 361 predicted by the reaction stoichiometry to completely mineralize 20 mg L-1 of OTC (77 362 mg H2O2 L-1). In fact, 38% of the residual DOC corresponds to low-molecular-weight 363 carboxylic acids (LMWCA) in solution, namely oxalic and oxamic acids. 364 The H2O2/UVC process led to higher OTC degradation rates with the increment of Q 365 from 50 to 100 L h-1 (Table 3), indicating a change in the hydrodynamic conditions 366 inside the photoreactor. A Re number of 446 (Q = 100 L h-1) allowed to a 2.3-fold 367 increase on OTC oxidation rate comparing with a Re of 223 (Q = 50 L h-1). At those 368 conditions, an OTC removal above 90% is achieved after 5 min of reaction with 0.4 kJ 369 L-1. 370 16 3.2 OTC degradation by an UVC/H2O2 photochemical system using an innovative 371 FluHelik photoreactor in multiple pass flow mode 372 An initial H2O2 dose of 300 mg L-1 led to a maximum OTC oxidation rate and 373 mineralization, corresponding to a 9-fold increment on reaction rate when compared to 374 direct photolysis (Table 3). A higher oxidant dose (400 mg L-1) led to a slightly decrease 375 in the OTC oxidation rate, due to hydroxyl radicals quenching by the hydrogen peroxide 376 molecule itself. This is also supported by the growing H2O2 consumption for rising 377 initial H2O2 doses (Fig. 3). Using 300 mg L-1 of oxidant, a 77% mineralization was 378 attained after 360 min with a final residual H2O2 concentration of 17 mg L-1. 50% of the 379 remaining DOC was from oxalic and oxamic acids. In turn, the nitrogen content of the 380 OTC compound was converted to nitrites, nitrates and ammonium, with ammonia 381 representing the largest fraction. The un-mineralized fraction of nitrogen proved to be 382 present as oxamic acid, as also observed by Pereira et al. (2013)). 383 Insert Figure 3 384 The OTC removal rate showed a 1.6-fold increase when the flow rate increased from 50 385 to 75 L h-1. A further increase on flow rate from 75 to 100 L h-1 resulted in an increment 386 on the reaction rate of only 1.2 times (see Table 3). This indicates that the 387 hydrodynamic conditions do not considerably change between 75 to 100 L h-1. Under 388 this condition (100 L h-1), an OTC removal >90% is reached after 5 min of reaction (0.4 389 kJ L-1 of accumulated energy), with a photonic efficiency (ξ) (number of OTC 390 transformed molecules divided by the number of incident photons) of 13.7%. 391 Comparing the degradation of the OTC molecule in UPW matrix by the two reactors 392 under study in the best conditions found for each one (Fig. 4), it was practically similar 393 in both reactors. On the other hand, FluHelik reactor showed to be more efficient (1.3 394 times) in terms of mineralization (77%) when compared with the Jets reactor (61%), for 395 17 the same accumulated UVC energy (14.4 kJUV L-1). This indicates that the limiting step 396 of the reaction is the by-products removal, which is improved by the unique fluid 397 dynamics and irradiation properties of FluHelik reactor. 398 Insert Figure 4 399 3.3 Effect of urban wastewater (UWW) matrix 400 OTC removal by the UVC/H2O2 photochemical system was also evaluated for an UWW 401 fortified with 20 mg OTC L-1. Fig. 5a shows an increment on OTC removal rate for 402 higher H2O2 doses using the Jets photoreactor. In fact, a 29-fold increase on OTC 403 oxidation rate is observed for the UVC/H2O2 system ([H2O2]0 = 500 mg L-1) when 404 compared to direct photolysis (Table 3). Likewise, it was found that the highest OTC 405 oxidation rate using the UWW and the FluHelik photoreactor was reached with the 406 highest oxidant concentration applied ([H2O2]0 = 500 mg L-1) (Fig 5b, Table 3). An 407 increase in the flow rate value from 50 to 100 L h-1, achieved a 1.2-fold improvement in 408 the OTC reaction rate (Table 3), associated with an higher degree of mixing inside the 409 system. Under those conditions (500 mg H2O2 L-1; 100 L h-1), 90% of OTC removal 410 was achieved after 7.5 min of reaction and using 0.6 kJUV L-1 and a final mineralization 411 of 71% was attained after 180 min of reaction and 14.4 kJUV L-1. Paralleling the reaction 412 rates in both matrices, a decrease of about 1.7 times was perceived when in the presence 413 of the UWW, using the same initial H2O2 dosage, mainly due to inner filter and 414 hydroxyl radicals scavenging effects (Wols and Hofman-Caris, 2012). Therefore, a 415 higher amount of oxidant is required to overcome those effects to obtain similar OTC 416 removals. In fact, when using the FluHelik photoreactor for both matrices, at the best 417 conditions, similar photochemical space time yield (PSTY) were observed (0.50 and 418 0.53 m3water m-3reactor day-1 kW-1 using the UPW and UWW, respectively). This shows 419 the ability of the FluHelik reactor design to overcome matrix effects due to its unique 420 18 characteristics. It should be noted that the high H2O2 concentrations used in these 421 experiments were due to the low photon flows provided by the available UVC lamps. If 422 higher photon flows are provided (able of overcoming the wastewater inner filter 423 effects), lower initial doses of oxidant would be required to achieve the same OTC 424 oxidation rates. 425 Insert Figure 5 426 The FluHelik photoreactor showed a better performance than the conventional Jets 427 reactor during the OTC oxidation when in the presence of the UWW matrix at the same 428 operating conditions (500 mg L-1 of H2O2 and 100 L h-1) (Fig. 6, Table 3). In this case, 429 the helical movement of the fluid around the radiation source allows a more 430 homogeneous UV radiation distribution, enhancing the reaction rate. In fact, the Jets 431 reactor presents a solution flow pattern parallel to the radiation source, and 432 consequently, the liquid streams at higher distance from the light source receive a less 433 UV dose. FluHelik’s unique fluid hydrodynamics also provided a more efficient oxidant 434 homolytic cleavage, allowing further removal of the remaining by-products. In fact, a 435 1.4 times higher hydrogen peroxide consumption was observed when using the 436 FluHelik reactor (Fig. 6), reaching a mineralization of 71% instead of 56% for the Jets 437 reactor, using the same accumulated UVC energy (14.4 kJUV L-1). 438 Insert Figure 6 439 The FluHelik photoreactor was also evaluated for the treatment of an UWW matrix 440 spiked with OTC and 10 additional CECs (described in Table 2) from the 12 established 441 by Swiss legislation (Hochstrat et al., 2015) at residual concentrations (60 μg L-1) using 442 the UVC/H2O2 photochemical process. The Swiss legislation establishes 80% removal 443 for 5 out of 12 indicator compounds (11 pharmaceuticals and 1 biocide) (Hochstrat et 444 al., 2015). An OTC removal of more than 80% can be achieved using only 10 mg L-1 of 445 19 H2O2 (Table 2) after 30 min. In order to comply with the Swiss legislation, an oxidant 446 amount of 250 mg L-1 is required to achieve 80% removal of 5 compounds after 30 min. 447 However, using a H2O2 dose of 500 mg L-1, after 30 min of reaction, a removal 448 efficiency of 80% is obtained for all the 11 CECs. 449 3.4 CECs removal by an UVC/H2O2 photochemical system using FluHelik/Jets 450 photoreactors in single pass flow mode 451 In order to estimate the efficiency in a real scale implementation, tests were performed 452 in single pass flow mode (one passage through the reactor) instead of recirculating the 453 solution between the reactor and the feed tank (multiple passage). In this way it is also 454 ensured that only the hydrodynamic effect of the reactors is evaluated, excluding the 455 additional mixture promoted by the recirculation. Fig 7a shows an improvement on the 456 OTC removal from an UWW matrix, at the steady state conditions (5 times the 457 residence time), by the UVC/H2O2 process (500 mg H2O2 L-1; 6 W; 100 L h-1), using the 458 FluHelik photoreactor (18% OTC removal) instead of the Jets reactor (15% OTC 459 removal). In fact, the longer residence time of FluHelik reactor (0.4 min) when 460 compared to that of the Jets reactor (0.3 min), along with the higher accumulated energy 461 (in a single passage) and with the lower dead volume zones due to the FluHelik helical 462 movement of the fluid contributed to the higher OTC removal. 463 Table 2 shows that much smaller removals of all the 11 CECs spiked in the UWW were 464 achieved by operating the FluHelik reator in single passage mode (low residence time): 465 none of the compounds reach the removal imposed by Swiss legislation. The design of 466 the FluHelik photoreactor strongly favors the implementation of various reactors in 467 series, promoting its application in industry. Therefore, two FluHelik photoreactors 468 associated in series were tested for the removal of 20 mg L-1 of OTC under the same 469 conditions previously tested with only one reactor. However, it was noticed that when 470 20 using the two reactors in series an OTC removal of only 31% was obtained after 471 reaching the steady state (Fig. 7b), a lower value than the one expected (36% - twice the 472 one achieved with only one reactor). In fact, when using 2 FluHelik reactors in series, 473 different velocity profiles can be found in the each reactor due to an increase in fluid 474 energy dissipation (pressure drop). Therefore, a new test was carried out doubling the 475 flow rate (200 L h-1), corresponding to a residence time of 0.4 min. At these conditions, 476 an OTC removal of 36% was attained at steady state conditions. Therefore, when using 477 2 FluHelik reactors in series there is a minimum flow rate value to be used to achieve 478 fluid velocities profiles inside both photoreactors similar to when using only one 479 FluHelik reactor. 480 Insert Figure 7 481 Finally, a pilot-scale FluHelik reactor (95 W UVC lamp) under multiple pass flow mode 482 was also evaluated for the OTC removal using either UPW and UWW as solution 483 matrices. This system was operated at a flow rate of 7500 L h-1, attaining a turbulent 484 regime inside the photoreactor (Re = 15000). Fig. 8a and 8b show that the highest OTC 485 oxidation rate was reached when using 500 mg L-1 of H2O2 for both reaction matrices. 486 However, a 1.5-fold decrease in the OTC kinetic rate (Table 3) was obtained for the 487 UWW when compared with UPW, as well as a slightly lower mineralization (52% 488 instead of 58%) using 4.6 kJUV L-1. It should be noted that, at the best conditions, when 489 using UPW, a higher photochemical space time yield (PSTY) at pre-pilot scale (0.85 490 m3water m-3reactor day-1 kW-1) when compared with the one at lab scale (0.50 m3water 491 m-3reactor day-1 kW-1) was noticed. This dissimilarity is mainly associated to the different 492 flow rates and UV fluence inside the reactors (distinct path length and UVC lamp 493 power). In addition, when using UWW, similar PSTY were observed at both scales 494 (0.57 and 0.53 m3water m-3reactor day-1 kW-1 at the pre-pilot scale and lab scale, 495 21 respectively). These data are in agreement with the results obtained by Moreira et al. 496 (2019)) when comparing the degradation of a model compound, 3-amino-5-497 methylisoxazole, using the FluHelik at the lab and pre-pilot scale; and indicate the 498 feasibility of scaling-up the FluHelik reactor. 499 Insert Figure 8 500 3.5 Toxicity 501 The percentage of embryo mortality at the end of the bioassays was similar among 502 treatments and remained at low levels, below 3% (data not shown). The total 503 abnormalities, the length and the yolk sac perimeter observed on zebrafish embryos 504 exposed to the initial and treated effluents are presented in Fig. 9. A significant increase 505 of total abnormalities (sum of tail abnormalities, lordosis anomalies and pericardial 506 oedema) was observed in embryos exposed to the UWW + 11 CECs, with 12.5% of 507 abnormal embryos in comparison with 1.4% of the control (p < 0.05). These 508 abnormalities were significantly reduced after the FluHelik photochemical treatment 509 (UWW + 11 CECs + Treatment), with values similar to the control (4.7% - p > 0.05). 510 The exposure to the UWW without the 11 CECs addition did not cause significant 511 abnormalities in the embryos (p > 0.05). 512 Insert Figure 9 513 It was also verified that the exposure to the UWW + 11 CECs significantly decreased 514 the length of the larvae and increase the yolk sac perimeter when compared with control 515 (p < 0.05). These endpoints return to control levels after the FluHelik photochemical 516 treatment (UWW + 11 CECs + Treatment). The yolk sac perimeter was also 517 significantly increased by the UWW without the 11 CECs. 518 When comparing the toxicity effects of the UWW + 11 CECs with those of the UWW + 519 11 CECs + Treatment (Fig. 9), it is evident that the UVC/H2O2 photochemical system 520 22 using the FluHelik photoreactor led to a significant decrease of the toxicity on zebrafish 521 embryos. In this sense, the treated wastewater had no significant effects on the total 522 abnormalities incidence, in the length of the larva and in the yolk sac perimeter. Thus, it 523 may be assumed that the degradation of pollutants present in the UWW fortified with 11 524 CECs by the UVC/H2O2 photochemical system with the FluHelik photoreactor did not 525 result in toxic transformation products to zebrafish embryos. 526 4. Conclusions 527 The FluHelik photoreator showed to be an interesting system for UVC/H2O2 528 photochemical process applied to the removal of CECs from urban wastewaters, as a 529 polishing step, being able to to comply with the Switzerland legislation and to 530 effectively reduce CECs toxicity to zebrafish embryos. Due to its unique configuration, 531 the FluHelik promotes an helical movement of the fluid around the irradiation source, 532 providing a more homogeneous UV radiation distribution (each fluid particle receives a 533 similar UVC radiation dosage), being able to overcome matrix effects in wastewaters 534 with low to moderate transmissibility (inner filter effects). Another advantage of this 535 reactor technology is its easy scalability, through its very simple and compact 536 arrangement in series, strongly promoting its use in industrial applications. 537 Acknowledgments 538 This work was financially supported by: Associate Laboratory LSRE-LCM - 539 UID/EQU/50020/2019 - funded by national funds through FCT/MCTES (PIDDAC). 540 V.J.P. Vilar acknowledges the FCT Investigator 2013 Programme (IF/00273/2013). 541 J.C.A. Espíndola acknowledges CNPq (Brazil) for his scholarship (205781/2014-4). R. 542 Montes, R. Rodil and J.B. Quintana acknowledge the financial support of Spanish 543 “Agencia Estatal de Investigación” (ref. CTM2017-84763-C3-R-2) and Xunta de 544 Galicia (ref. ED431C2017/36), both confounded by FEDER/ERDF. 545 23 References 546 Barros S, Montes R, Quintana JB, Rodil R, André A, Capitão A, et al. Chronic 547 environmentally relevant levels of simvastatin disrupt embryonic development, 548 biochemical and molecular responses in zebrafish (Danio rerio). Aquatic 549 Toxicology 2018; 201: 47-57. 550 Boyjoo Y, Ang M, Pareek V. CFD simulation of a pilot scale slurry photocatalytic 551 reactor and design of multiple-lamp reactors. Chemical Engineering Science 552 2014; 111: 266-277. 553 Cambié D, Bottecchia C, Straathof NJW, Hessel V, Noël T. Applications of 554 Continuous-Flow Photochemistry in Organic Synthesis, Material Science, and 555 Water Treatment. Chemical Reviews 2016; 116: 10276-10341. 556 Caris CHM. New concepts of UV/H2O2 oxidation: KWR, 2011. 557 Diya’uddeen BH, Daud WMAW, Abdul Aziz AR. Treatment technologies for 558 petroleum refinery effluents: A review. Process Safety and Environmental 559 Protection 2011; 89: 95-105. 560 H. Baxendale J, A. Wilson J. The Photolysis of Hydrogen Peroxide at High Light 561 Intensities. Vol 53, 1957. 562 Hochstrat R, Scharer M, Bleny H. Elimination of micropollutants – the Swiss approach. 563 TAPES Final Conference, Brussels, 2015. 564 Jeong J, Song W, Cooper WJ, Jung J, Greaves J. Degradation of tetracycline antibiotics: 565 Mechanisms and kinetic studies for advanced oxidation/reduction processes. 566 Chemosphere 2010; 78: 533-540. 567 Karpel VelLeitner N, Le Bras E, Foucault E, Bousgarbies JL. A new photochemical 568 reactor design for the treatment of absorbing solutions. Water Science and 569 Technology 1997; 35: 215-222. 570 Krishnan S, Rawindran H, Sinnathambi CM, Lim JW. Comparison of various advanced 571 oxidation processes used in remediation of industrial wastewater laden with 572 recalcitrant pollutants. IOP Conference Series: Materials Science and 573 Engineering 2017; 206: 012089. 574 Kuhn HJ, Braslavsky SE, Schmidt R. Chemical actinometry (IUPAC Technical Report). 575 Pure and Applied Chemistry. 76, 2004, pp. 2105. 576 Lammer E, Carr GJ, Wendler K, Rawlings JM, Belanger SE, Braunbeck T. Is the fish 577 embryo toxicity test (FET) with the zebrafish (Danio rerio) a potential 578 alternative for the fish acute toxicity test? Comparative Biochemistry and 579 Physiology Part C: Toxicology & Pharmacology 2009; 149: 196-209. 580 Leblebici ME, Stefanidis GD, Van Gerven T. Comparison of photocatalytic space-time 581 yields of 12 reactor designs for wastewater treatment. Chemical Engineering and 582 Processing: Process Intensification 2015; 97: 106-111. 583 Liu Y, He X, Duan X, Fu Y, Dionysiou DD. Photochemical degradation of 584 oxytetracycline: Influence of pH and role of carbonate radical. Chemical 585 Engineering Journal 2015; 276: 113-121. 586 Macedo S, Torres T, Santos MM. Methyl-triclosan and triclosan impact embryonic 587 development of Danio rerio and Paracentrotus lividus. Ecotoxicology 2017; 26: 588 482-489. 589 Masschelein WJ. Ultraviolet disinfection of water. In : Unit Processes in Drinking 590 Water Treatment. Vol Chap. 4, 1992. 591 Moreira FC, Bocos E, Faria AGF, Pereira JBL, Fonte CP, Santos RJ, et al. Selecting the 592 best piping arrangement for scaling-up an annular channel reactor: An 593 30 Figure 3 700 701 702 703 704 705 706 0 2 4 6 8 10 12 14 16 18 20 0.0 0.2 0.4 0.6 0.8 1.0 [OTC]/[OTC]0 Time (min) 020 40 60 0 50 100 150 200 Consumed H 2O2 (mg L-1) Time (min) 31 Figure 4 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 0.0 0.4 0.8 1.2 1.6 0.0 0.2 0.4 0.6 0.8 1.0 0 4 8 12 16 20 24 28 0 50 100 150 200 250 300 QUV (kJ L-1) Consumed H 2O2 (mg L-1) 0 20 40 60 80 100 Mineralization (%) jet flu [OTC]/[OTC]0 QUV (kJ L-1) 32 Figure 5 726 727 (a) (b) 728 729 020 40 60 0.0 0.2 0.4 0.6 0.8 1.0 020 40 60 0 25 50 75 100 Consumed H 2O2 (mg L-1) Time (min) [OTC]/[OTC]0 Time (min) 020 40 60 0.0 0.2 0.4 0.6 0.8 1.0 020 40 60 0 50 100 150 200 Consumed H 2O2 (mg L-1) Time (min) [OTC]/[OTC]0 Time (min) 33 Figure 6 730 731 732 0.0 0.4 0.8 1.2 1.6 2.0 2.4 0.0 0.2 0.4 0.6 0.8 1.0 04812 0 50 100 150 200 250 300 350 400 450 QUV (kJ L-1) Consumed H 2O2 (mg L-1) 0 20 40 60 80 100 Mineralization (%) [OTC]/[OTC]0 QUV (kJ L-1) 34 Figure 7 733 734 (a) (b) 735 736 737 738 739 740 741 742 0.0 0.1 0.2 0.3 0.4 0.80 0.84 0.88 0.92 0.96 1.00 [OTC]/[OTC]0 QUV (kJ L-1) 0 1 2 3 4 5 0.64 0.68 0.72 0.76 0.80 0.84 0.88 0.92 0.96 1.00 [OTC]/[OTC]0 Time (min) 35 Figure 8 743 (a) (b) 744 0 5 10 15 20 25 30 35 40 45 0.0 0.2 0.4 0.6 0.8 1.0 [OTC]/[OTC]0 Time (min) 020 40 60 0 10 20 30 40 50 60 70 Consumed H 2O2 (mg L-1) Time (min) 0 5 10 15 20 25 30 35 40 45 0.0 0.2 0.4 0.6 0.8 1.0 020 40 60 0 10 20 30 40 50 Consumed H 2O2 (mg L-1) Time (min) [OTC]/[OTC]0 Time (min) 36 Figure 9 745 746 (a) (b) 747 0 20 40 60 80 100 Control UWW UWW + 11 CECs UWW + 11 CECs + Treatment Total abnormalities (%) a b a a 0 20 40 60 80 100 120 140 Control UWW UWW + 11 CECs UWW + 11 CECs + Treatment Perimeter (%) a b a a + * # + a a b a 37 Table 1. Main physicochemical characteristics of the real urban wastewater collected 748 after secondary treatment. 749 Parameter (units) Values Color Pale/Yellow Odor n.d.a pH 6.5 Temperature (°C) 24.7 Turbidity (UNT) 1.0 Conductivity (µS cm-1) 883 Dissolved oxygen (mg L-1) 3.8 Redox potential (mV) -10 Total dissolved carbon (mg L-1) 51 Dissolved inorganic carbon (mg L-1) 33 Dissolved organic carbon (mg L-1) 18 Chemical oxygen demand (mg O2 L-1) 56 Total dissolved iron (mg L-1) 0.26 Absorbance at 254 nm (AU) 0.21 Total suspended solids (mg L-1) 1.7 Volatile suspended solids (mg L-1) 1.7 Total dissolved nitrogen (mg L-1) 3.9 Total dissolved organic nitrogen (mg L-1) 2.7 Ammonium - N-NH4+ (mg L-1) 1.1 Nitrite - N-NO2- (mg L-1) <0.02 Nitrate - N-NO3- (mg L-1) 0.09 Bromide - Br- (mg L-1) 0.1 Chloride - Cl- (mg L-1) 174 Phosphate - PO43- (mg L-1) 12 Sulfate - SO42- (mg L-1) 76 Calcium - Ca2+ (mg L-1) 57 Lithium - Li+ (mg L-1) <0.02 Magnesium - Mg2+ (mg L-1) 9.1 Potassium - K+ (mg L-1) 29 Sodium Na+ (mg L-1) 136 Total phosphorous - P (mg L-1) 4.8 Atenolol (μg L-1) 1.1 Carbamazepine (μg L-1) 4.8 Diclofenac (μg L-1) 2.8 Metformin (μg L-1) 2.1 Sulfamethoxazole (μg L-1) 2.4 Trimethoprim (μg L-1) 3.7 a n.d. - Not detected. 750 38 Table 2. Effect of H2O2 initial concentration in the removal of 11 pharmaceuticals spiked in a real urban wastewater by UVC/H2O2 751 photochemical system using a FluHelik photoreactor in multiple or single pass flow mode. 752 Name Chemical Formula Company % Removal of CECs in multiple pass flow mode (after 30 min) % Removal of CECs in single pass flow mode [H2O2]0 10 mg L-1 25 mg L-1 50 mg L-1 250 mg L-1 500 mg L-1 500 mg L-1 Azytromicyn C38H72N2O12 TCI 9.3 19 23 > 95 > 95 24 Naproxen C14H14O3 AlfaAesar 60 68 79 > 99 > 99 26 Atenolol C14H22N2O3 AlfaAesar 27 42 62 > 99 > 99 35 Metformin NH2C(=NH)NHC( =NH)N(CH3)2.HCl AlfaAesar 1.9 6.0 6.7 72 90 15 Bezafibrate C19H20ClNO4 AlfaAesar 40 51 66 > 99 > 99 18 Ibuprofen C13H18O2 AlfaAesar 30 38 63 > 99 > 99 35 Trimethoprim C14H18N4O3 AlfaAesar 19 29 43 > 99 > 99 17 Carbamazepin C15H12N2O ACROS organics 25 30 45 > 99 > 99 19 Sulfamethoxazole C10H11N3O3S TCI 98 99 > 99 > 99 > 99 25 Oxytetracycline C22H24N2O9.HCl AppliChem Panreac 92 96 98 > 99 > 99 32 Diclofenac C14H10ClN.NaO2 Sigma Aldrich 96 98 97 > 99 > 99 31 753 39 Table 3. Pseudo-first order kinetic constants along with the corresponding coefficient of determination (R2) and residual variance (S2r), photonic 754 efficiencies ( and photochemical space time yields (PSTY) for degradation of 20 mg L-1 of OTC at pH 7.5 and 25 ºC. 755 Experiment [OTC]0 (mg L-1) [H2O2] (mg L-1) Q (L h-1) pH k×101 (min-1) k (L kJ-1) R2 S2r (mg L-1)2 r0×102 (mg L-1 min-1) ×102 PSTY (m3water m-3reactor day-1 kW-1) Jets photoreactor - UVC lamp of 6 W 1.1 5.1 0 100 4.5 0.17 ± 0.01 0.21 ± 0.01 0.989 0.02 0.018 ± 0.001 0.11 0.02 1.2 5.0 0 100 7.5a 0.30 ± 0.02 0.37 ± 0.03 0.963 0.08 0.032 ± 0.002 0.19 0.03 1.3 5.0 0 100 7.5 1.1 ± 0.1 1.4 ± 0.1 0.976 0.07 0.12 ± 0.01 0.70 0.12 Jets photoreactor - UVC lamp of 4 W 2.1 21.3 0 100 7.5 0.41 ± 0.04 1.2 ± 0.1 0.963 2.0 0.19 ± 0.02 2.6 0.10 2.2 20.9 20 100 7.5 1.3 ± 0.1 4.0 ± 0.3 0.976 1.3 0.6 ± 0.1 8.5 0.34 2.3 21.8 50 100 7.5 1.5 ± 0.2 4.6 ± 0.5 0.963 2.0 0.7 ± 0.1 10.3 0.40 2.4 19.9 100 100 7.5 2.7 ± 0.1 7.8 ± 0.4 0.988 0.6 1.2 ± 0.1 16.1 0.69 2.5 20.0 200 100 7.5 2.8 ± 0.1 8.2 ± 0.3 0.999 0.1 1.23 ± 0.04 17.2 0.73 Jets photoreactor - UVC lamp of 6 W 3.1 20.9 0 100 7.5 0.52 ± 0.05 0.65 ± 0.06 0.961 2.0 0.24 ± 0.02 1.4 0.06 3.2 22.0 20 100 7.5 2.4 ± 0.3 3.1 ± 0.3 0.968 2.1 1.2 ± 0.1 6.9 0.26 3.3 22.9 50 100 7.5 4.0 ± 0.3 5.0 ± 0.4 0.972 1.6 2.0 ± 0.2 11.7 0.43 3.4 20.2 100 100 7.5 5.0 ± 0.2 6.3 ± 0.2 0.998 0.1 2.2 ± 0.1 13.0 0.54 3.5 21.0 200 100 7.5 4.7 ± 0.2 5.9 ± 0.3 0.996 0.3 2.1 ± 0.1 12.6 0.51 3.6 20.8 100 50 7.5 2.2 ± 0.1 2.8 ± 0.2 0.992 0.5 1.0 ± 0.1 5.9 0.24 3.7 21.4 100 75 7.5 3.0 ± 0.3 3.7 ± 0.4 0.982 0.8 1.4 ± 0.1 8.2 0.32 Jets photoreactor - UVC lamp of 11 W 4.1 21.5 0 100 7.5 0.58 ± 0.04 0.61 ± 0.05 0.962 1.8 0.27 ± 0.02 1.3 0.05 4.2 21.0 20 100 7.5 2.0 ± 0.2 2.1 ± 0.2 0.983 1.2 0.9 ± 0.1 4.4 0.18 4.3 22.0 50 100 7.5 3.5 ± 0.2 3.6 ± 0.2 0.974 1.2 1.7 ± 0.1 8.2 0.32 4.4 21.3 100 100 7.5 5.18 ± 0.02 5.40 ± 0.02 0.999 0.1 2.40 ± 0.01 11.8 0.47 4.5 21.3 200 100 7.5 6.4 ± 0.1 7 ± 1 0.986 0.9 2.94 ± 0.03 14.4 0.57 4.6 20.6 500 100 7.5 7.2 ± 0.2 7.5 ± 0.2 0.999 0.1 3.2 ± 0.1 15.9 0.65 4.7 20.0 700 100 7.5 5.7 ± 0.3 5.9 ± 0.3 0.996 0.2 2.5 ± 0.1 12.1 0.51