The effect of CO₂ addition and hydraulic retention time on pathogens removal in HRAPs

Pathogenic microorganisms and inadequate sanitation are responsible for the spread of numerous diseases and bacterial infections in the world (WHO 2018). The main route of entry of these microorganisms into the aquatic environment is through direct discharge of untreated or inadequately treated wastewater. Despite increasing coverage of basic sanitation services for the world's population, there is still significant inequality, especially in developing countries, and in 2017, 673 million people were still practicing open defecation, facilitating the dissemination of waterborne diseases (UNICEF & WHO 2019). Even where sewage is collected, it is often discharged into water bodies without treatment, as is the case, for instance, with 44% of sewage in Brazil and 78% of sewage in India. Furthermore, the intensive use of antibiotics, and their release into the environment, has accelerated the development of antibiotic resistant bacteria (ARBs) and antibiotic resistant genes (ARGs) (Bouki et al. 2013). These multiresistant organisms increase the risks to human and animal health because they make it difficult to control infections (Rizzo et al. 2013).

Considering that pathogens are always present in the natural aquatic environment, even though often only in small amounts, and that detection is not always simple to realize, usually indicator organisms, which are present in faecal material and with characteristics that may be related to pathogens in general, are used to detect contamination. Escherichia coli is considered the universal indicator due to its high prevalence, and is related to various gastrointestinal infections that may even result in chronic sequelae (Aw 2018). Other common inhabitants of the intestinal tract of warm-blooded organisms, and used as indicator organisms, are Enterococcus faecalis species, which have the aquatic environment as their natural habitat but are very resistant, making their detection related to fecal contamination (Motlagh & Yang 2019). Other sanitary-relevant bacteria may also be used as alternative indicators in monitoring, like for instance Pseudomonas, Clostridium and Staphylococcus (Coronel-Olivares et al. 2011). These bacteria are nowadays frequently resistant to antibiotics and, when resistant, can cause serious infections that were considered ‘under control’, such as dysentery (Motlagh & Yang 2019). The conventional technologies used in wastewater treatment plants (WWTPs) are not efficient in removing pathogenic bacteria, and in fact WWTPs are hotspots for the propagation of multi-resistant organisms, as in these plants optimal conditions for the appearance of new, resistant variants are present due to high bacterial density, availability of nutrients, and the possibility for gene exchanges (Bouki et al. 2013; Rizzo et al. 2013).

Microalgal-bacterial systems, used in process units such as high rate algal ponds (HRAPs), constitute a sustainable, low-cost and efficient alternative for the treatment of effluents (Posadas et al. 2015). Recently, satisfactory results were shown using HRAPs for the inactivation of pathogenic bacteria (Ruas et al. 2018). The operating conditions of such reactors, such as HRT, pH and concentration of CO₂, can be changed according to the pollutants present and the required removal rates. For example, in the majority of domestic wastewaters, carbon is limiting, and thus the addition of CO₂, decreasing this limitation, can improve the removal of organic matter and of nutrients, resulting in increased productivity (Park & Craggs 2010). In this case, the necessary CO₂ can be easily obtained from a preceding anaerobic treatment of the wastewater, or from anaerobic digestion of excess sludge at the WWTP itself, or even from fossil fuel combustion for power generation, if a power plant is located nearby.

However, there are still few studies evaluating the pathogen removal performance of HRAPs, and understanding the removal mechanisms and the effects of operating conditions on this removal is critical for determining the best configuration for these systems, in which nutrient removal can then be combined with disinfection and biomass production in a single optimal configuration. Ruas et al. (2018) investigated the influence of CO₂ addition on pathogen removal in an HRAP. However, in that study only a single HRAP was operated, and thus the different stages were characterized by different environmental conditions. Simultaneous operation of at least two reactors would greatly improve the comparative analysis of the impact of CO₂ addition and other parameters on pathogen removal. Thus, the aim of this work was to study, in parallel reactors, the influence of CO₂ addition and hydraulic retention time (HRT) on the removal of Pseudomonas aeruginosa, C. perfringens, Staphylococcus sp., Enterococcus sp. and E. coli from raw domestic wastewater (RDW).

The experimental setup consisted of three 21 L pilot HRAPs with 0.13 m² surface area and 16 cm cultivation broth depth (called HRAP-1, HRAP-2 and HRAP-3). A liquid recirculation velocity of 20 cm s⁻¹ was maintained in each reactor, by means of a submerged pump with a nominal flow rate of 650 L h⁻¹ (Sarlo Better B650, Brazil). The effluent of each reactor passed through a 1 L settler with a hydraulic retention time (HRT) of 6 h for HRAP-1 and HRAP-2, and 8.4 h for HRAP-3 (Figure 1).

The reactors were inoculated with 1.3 L microalgae culture (≈97% Chlorella vulgaris) with 0.9 g L⁻¹ of total suspended solids (TSS) and with 0.28 L of a nitrifying-denitrifying active sludge (4.7 gTSS L⁻¹) from a WWTP in Campo Grande-MS, Brazil. Microalgae were collected and acclimated to the RDW prior to inoculation of the HRAPs, according to Serejo et al. (2015).

RDW was collected once a week from a WWTP in Campo Grande-MS (Brazil) and stored in a 300 L agitation cooling tank (Implemis, Brazil) at 4 °C, in order to prevent sedimentation of suspended solids and degradation of organic matter. The initial concentrations of soluble chemical oxygen demand (COD), total organic carbon (TOC), inorganic carbon (IC), total organic nitrogen (TN), ammonium ion (N-NH₄⁺), nitrite (N-NO₂⁻), nitrate (N-NO₃⁻), phosphorus (P), and non-soluble P. aeruginosa, C. perfringens, Staphylococcus, Enterococcus sp. and E. coli, as well as pH and TSS of the RDW, are summarized in Table 1. A synthetic gas mixture (SGM) composed of CO₂ (30%) and N₂ (70%) (White Martins, Brazil) was used as source of CO₂ for the pathogen removal efficiency experiments in HRAP-1, passing 2.5 ± 0.4 mL CO₂ min⁻¹. In the HRAP-2 and in the HRAP-3, no CO₂ was added.

The three HRAPs were continuously fed with RDW using peristaltic pumps (Watson Marlon 505 U, UK) applying an HRT of 5 days in HRAP-1 and HRAP-2 and an HRT of 7 days in HRAP-3, and with (HRAP-1) or without (HRAP-2, HRAP-3) using a synthetic gas mixture, in order to study the influence of CO₂ dosing and HRT on the removal efficiency of P. aeruginosa, C. perfringens, Staphylococcus, Enterococcus sp. and E. coli. During the experiment, three times a week, at 10:00 a.m., 300 mL liquid samples were withdrawn from the influent and the effluent (settler output, Figure 1) to determine the concentrations of pathogens, TSS, and (filtration through 0.45 μm glass fibre filters prior to analysis) soluble COD, TOC, IC, TN, N-NH₄⁺, NO₂⁻, NO₃⁻ and P. Also, 100 mL liquid samples were withdrawn from the HRAPs to determine the concentrations of pathogens and TSS in the cultivation broth. Light intensity, pH, dissolved oxygen (DO), and ambient and cultivation broth temperatures were monitored daily. The evaporation rate was measured from the difference between measured influent and effluent flows (Serejo et al. 2015). The experiment was conducted at the Effluents Laboratory of the Federal University of Mato Grosso do Sul (UFMS) in Campo Grande-MS (Brazil), for 40 days, at a temperature of ≈26 °C.

The TOC, IC and TN were determined using a total organic carbon analyzer (Vario TOC Cube, Elementar, Germany). N-NH₄⁺ was measured using an Orion Five Star multiparameter meter (Thermo Scientific, The Netherlands), while the pH was measured with a Hanna pH meter HI2211 (Hanna Instruments, Brazil). DO and temperature in the HRAPs were measured using a Jenway 9500 DO2 Oximeter (Jenway, UK). NO₂⁻, NO₃⁻, and P-PO₄³⁻ were analyzed using a Dionex UltiMate ICS 1100 ion chromatography system with an IonPac AG19/AS19 column (Thermo Scientific, USA). P. aeruginosa, C. perfringens, Staphylococcus sp., Enterococcus sp. were determined using the membrane filtration method with M-PA Agar Base – M1121, M-CP Agar Base M1354, Baird Parker Agar Base M043 (HIMEDIA, India) and Agar Base M-Enterococcus K25-610134 (KASVI, Brazil), respectively. E. coli was determined using Colilert^® kits (IDEXX Laboratories, USA). The light intensity (PAR – Photosynthetically Active Radiation) was recorded using a Quantum meter MQ-200 (Apogee Instruments, USA). The identification of microalgae was conducted by microscopic examination (Olympus BX41, USA) of samples fixed with 5% Lugol's solution and stored at 4 °C prior to analysis, according to Sournia (1978).

The results were evaluated using an analysis of variance (ANOVA) with a Fisher's least significant difference (LSD) test applying a 95% confidence level, using Microsoft Excel.

The C:N ratio of the RDW was 1.2 (Table 1), lower than the ≈3:0 C:N ratio generally found in RDW (Posadas et al. 2015), and far below the ideal C:N ratio for the growth of microalgal biomass of C:N ≈5.6 (Borowitzka & Borowitzka 1988), suggesting a carbon limitation during the treatment. The habits and lifestyle of the population directly influence the wastewater composition and consequently the C:N. The tropical climate and unlimited access to water may have influenced the low C:N ratio found. When there is carbon limitation (a low C:N ratio), the biomass growth rate is low, and a relatively high pH may be reached. Consequently, the removal of nutrients may be incomplete, since at higher pHs (>9) ammonia removal occurs by volatilization and phosphorus removal by precipitation. In addition, at this pH, inhibition of microalgal and bacterial growth may occur due to the presence of free ammonia. That is why numerous studies have endeavored to increase the C:N ratio, present in domestic wastewater, through the direct addition of CO₂ (Park & Craggs 2010).

During the 40 days of operation, the temperatures of the cultivation broths inside the three HRAPs were similar (≈22 °C), and within the range of between 20 and 30 °C considered optimal for the growth of most microalgal species. At the same time, mean ambient temperature was 25.6 ± 4.7 °C and light intensity was 725 ± 157 μmol m⁻² s⁻¹. Evaporation rates were similar in HRAP-1, HRAP-2 and HRAP-3, at 0.9 ± 0.7, 0.7 ± 0.7 and 0.8 ± 0.7 L m⁻² d⁻¹, respectively (Table 2), somewhat lower than the evaporation rates of ≈1.3 L m⁻² d⁻¹ reported by Guieysse for HRAPs under outdoor conditions (Guieysse et al. 2013).

The pH and DO in HRAP-2 were 8.7 ± 0.3 and 8.5 ± 1.4 mg O₂ L⁻¹, respectively. Higher results were recorded in HRAP-3 (9.5 ± 0.2 and 9.2 ± 1.6 mg O₂ L⁻¹). This higher pH in HRAP-3 was attributed to the higher HRT (7 days) in combination with the low buffer capacity of the RDW, due to low IC concentrations (Posadas et al. 2015). On the other hand, the pH and DO in HRAP-1 of 7.8 ± 0.8 and 7.8 ± 1.3 mg O₂ L⁻¹, respectively, were lower than those found in HRAP-2. The CO₂ addition in the HRAP-1 thus showed to promote an acidification of the cultivation broth, which in its turn led to better conditions for nitrifying–denitrifying sludge (Park & Craggs 2010). In all reactors, DO was above 2.0 mg O₂ L⁻¹, thus aerobic conditions were maintained (Posadas et al. 2014a).

The removal efficiencies of COD and TOC were higher in HRAP-3 (67 ± 6 and 69 ± 5%, respectively) than in HRAP-1 (41 ± 5 and 45 ± 3%) and in HRAP-2 (34 ± 4 and 36 ± 6%), due to higher bacterial activity at higher HRT, as also observed by Posadas et al. (2014b). Likewise, IC removal from HRAP-3 (80 ± 3%) was also higher than from HRAP-1 (67 ± 8%) and HRAP-2 (64 ± 2%), as probably the higher HRT permitted higher C stripping from HRAP-3 (Posadas et al. 2014b). The average removal of TN and TP in HRAP-3 were 90 ± 1% and 87 ± 7%, respectively, higher than removals as recorded in HRAP-1 and HRAP-2, where removals of ≈79% and 58% were recorded for TN and TP, respectively (Table 3).

The highest biomass productivity, 3.2 ± 0.2 g m⁻² d⁻¹, was recorded in HRAP-1, where CO₂ was added, followed by HRAP-3 (2.6 ± 0.2 g m⁻² d⁻¹) and HRAP-2 (2.0 ± 0.2 g m⁻² d⁻¹) (Table 3). The addition of CO₂ improved productivity by reducing the carbon deficiency in the RDW, increasing the C:N ratio (Posadas et al. 2015; Ruas et al. 2018). Biomass productivity was still somewhat lower though than the 5 g m⁻² d⁻¹ obtained by Posadas et al. (2014b) in the co-treatment of domestic wastewater with fish farm wastewater in a 180 L outdoor HRAP, with a 7-day HRT and a much higher C:N ratio of ≈100:25. The best settleability, of ≈ 49%, was obtained in HRAP-1. This is in line with the observation by Park & Craggs (2010) that CO₂ addition in HRAPs may promote aggregation/bioflocculation of the microalgae with bacterial flocs, increasing biomass settling. At the beginning of the experiment, the HRAPs were inoculated with a consortium formed mainly by Chlorella sp. Gradually, Scenedesmus sp., which is commonly found in outdoor HRAPs (Posadas et al. 2015), replaced the original biomass, obtaining 100% dominance in HRAP-1, 99% in HRAP-2 and 90% in HRAP-3.

The removal efficiency in units log (log-RE) of P. aeruginosa was statistically similar in the three HRAPs, between 3.4 and 3.8 log-RE (Figure 2), higher than the 1.6 and 2.5 log-RE reported by Ruas et al. (2018) when treating domestic wastewater in a 180 L HRAP with and without CO₂ addition, respectively. However, the first operational difference between the two studies was the way of adding CO₂: an absorption column (100 cm) was used in the study of Ruas et al. (2018), whereas in the present study the addition was performed by bubbling CO₂ at the bottom of the reactor (16 cm depth). The absorption column provides more contact time and a closed environment, increasing the transfer rate of CO₂ from gaseous to the liquid phase, while bubbling is limited in terms of CO₂ dissolution in the cultivation broth (Serejo et al. 2015), and thus had no significant effect on the removal of P. aeruginosa in the present study. Additionally, the second operational difference is the light intensity, which in our current study (at ≈725 μmol m⁻² s⁻¹) was almost double the maximum recorded by Ruas et al. (2018). Light intensity (photo-oxidative damage) is one of the main factors in the inactivation of P. aeruginosa and other bacteria, which may explain the higher removal efficiency in the present study.

Furthermore, each microalgae produces different toxins, antibiotic substances and photosynthetic pigments, so they can affect pathogenic bacteria differently (Evangelista et al. 2008). In the earlier experiments of Ruas et al. (2018), different microalgae were present in each stage: Chlorella sp. and Microscopora sp., in stages I and II, respectively, possibly caused by the CO₂ addition, and the author indicated this as one of the possible causes of different P. aeruginosa-RE. Under the current experimental conditions, in spite of the CO₂ addition, the microalgae population remained constant in all HRAPs, which could explain the similar P. aeruginosa-RE under all conditions. However, further studies are needed to understand the effect of each microalgae genus on the removal of pathogenic bacteria in HRAPs.

Bahlaoui et al. (1997) reported ≈ 1.1 log-RE of P. aeruginosa treating raw domestic wastewater at 8 days' HRT in a 16,800 L HRAP, conducting experiments for 2 years, testing different HRTs. The authors concluded that the behaviour of opportunistic pathogen bacteria (Pseudomonas sp. and Aeromonas sp.) is different from that of faecal-indicator bacteria, and thus the removal mechanisms may be different as well, in terms of acting factors and complexity. The authors also emphasize that light intensity and temperature are important in disinfection in HRAPs, since the best results are obtained in summer, and the HRT might need to be increased in winter as a strategy to improve removal under conditions of lower light intensity and temperature.

When the treated wastewater is intended for agricultural reuse, C. perfringens must be monitored together with E. coli, because it is resistant to disinfection due to its ability to form spores when under adverse environmental conditions. Moreover, it is an indicator of enteric viruses and protozoa in treated wastewater (Gutiérrez-Alfaro et al. 2018). HRAPs are able to remove C. perfringens better than maturation ponds (García et al. 2008), because in maturation ponds the exposure to UV radiation and high temperatures is more intense, and this is one of the conditions that start the formation of spores in C. perfringens (Raju et al. 2007), which makes it difficult to inactivate.

The C. perfringens-RE was 2.2 ± 0.2 log-RE in HRAP-2 and in HRAP-3, and 2.6 ± 0.2 log-RE in HRAP-1; these three removal efficiencies were not statistically different (Figure 2). García et al. (2008) reported lower C. perfringens removal efficiencies in a 461 L HRAP (10 d HRT) and a 15 m² maturation pond (20 d HRT) of 1.7 and 1.2 log-RE, respectively, both treating RDW in summer (≈27 °C). When testing two wetland configurations, García et al. (2018) obtained higher removals, of 2.63 and 3.42 units log, in a free-water surface wetland and a sub-surface flow wetland, respectively, and UV radiation in the system decreased the C. perfringens removal, due to spore formation. On the other hand, Gutiérrez-Alfaro et al. (2018) compared three systems for C. perfringens removal: HRAP, solar disinfection (SODIS) and dissolved air flotation (DAF), finding lower removal efficiencies, of 0.1, 0.9 and 1.7 log, respectively. In this case, DAF can remove the bacteria even in spore form. Also, the size of the bacteria (diameter ∼1 μm and length ∼5 μm) facilities the filtration process of this system. However, the conditions of cultivation in the HRAP was not show, it cannot be inferred why the C. perfringens removal was so low compared to the present study. Therefore, more studies need to be conducted to determine which mechanisms and parameters act in the C. perfringens removal.

The addition of CO₂ and the HRT variation had no statistically significant influence on Staphylococcus-RE; removal efficiencies of 1.7 ± 0.9, 1.3 ± 0.6 and 1.3 ± 0.6 units log were obtained in the HRAPs 1, 2 and 3, respectively (Figure 2). The removal efficiencies obtained were similar to those found by García et al. (2008) with Staphylococcus sp.-RE of ≈1.2 log-RE in winter and summer, and observed no statistically significant difference between a HRAP and a maturation pond, both treating the same wastewater. The main parameters affecting Staphylococcus sp. removal in synthetic medium and in lab conditions were the luminous intensity and the concentration of biodegradable organic matter (Nola et al. 2010), but both the potential and the main mechanisms of Staphylococcus-RE in HRAPs are not yet known, and thus more investigation is needed. It is important monitoring Staphylococcus sp. in treatment systems, understanding how it can be removed and thus improving the efficiency of the treatment, because this bacterium is pointed out as one of the antibiotic resistance bacteria, and its presence in the environmental can put at risk the health of people and animals (Bouki et al. 2013).

The Enterococcus-RE was 3.1 ± 0.6 log-RE in HRAP-1, 3.0 ± 0.2 log-RE in HRAP-2 and 2.6 ± 0.4 log-RE in HRAP-3, statistically similar removal efficiencies (Figure 2). As previously reported by Ruas et al. (2018), CO₂ addition does not directly influence removal of Enterococcus sp. Similar Enterococcus sp. removal (≈2.7 log-RE) was obtained by Ruas et al. (2018) in a 180 L HRAP (5 d HRT), while lower RE of ≈ 2.0 and 1.1 unit log was found in a 0.7 m depth HRAP (7 d HRT) and 32 m² HRAP (6 d HRT), respectively, by Awuah et al. (2002) and Gutiérrez-Alfaro et al. (2018). This last work also recorded that Enterococcus sp. (mostly E. faecalis and E. faecium) was more effectively removed by the HRAP than by DAF (0.1 log-RE). According to Awuah et al. (2002), the main factor that can act in the Enterococcus sp. removal in microalgae systems was light and daily fluctuations of pH and DO.

Equal E. coli removal efficiencies were found in HRAP-2 and in HRAP-3 (3.3 ± 1.2 and 3.7 ± 0.2 log-RE, respectively), higher than that observed in HRAP-1 (2.5 ± 0.3 log-RE) (Figure 2). E. coli are Gram-negative bacteria, being more susceptible to pH, DO and light variations, when compared to Enterococcus sp. (Gram-positive) (Awuah et al. 2002). When the pH exceeds a value of 8.5, the kinetic constant of E. coli removal increases (Ouali et al. 2013). In both HRAPs-2 and -3, favourable conditions for E. coli removal were present throughout the experimental period: pH above 8.5 (≈8.7 and 9.5, in HRAP-2 and HRAP-3, respectively), and >8.5 mg O₂ L⁻¹. When the pH is closer to neutral, E. coli removal is reduced. For instance, Posadas et al. (2015) recorded lower E. coli removal of only ≈1.0 log in an 850 L HRAP with a controlled pH (pH ≈7.0) treating domestic wastewater (HRT 2.8 d), while Ruas et al. (2018) recorded ≈2.2 log-RE in two stages of operation, with pH of 7.7–6.8. Other parameters affect the E. coli removal including DO and light, where a lower intensity of these may have reduced the removal efficiencies.

The use of 30% CO₂ at 2.5 mL min⁻¹ in a 21 L HRAP for treatment of raw domestic wastewater with a low C:N ratio resulted in a strong increase in biomass productivity, from 2.0 to 3.2 g m⁻² d⁻¹, and in improvements in COD and TOC removal rates by 20 and 25%, respectively. On the other hand, the use of a 7-day HRT, when compared to a 5-day HRT, improved the removal of COD, TOC, TN and TP by 97, 91, 12 and 50%, respectively. In general, the removal efficiencies of pathogenic bacteria were similar under the tested conditions, but higher than those reported in previous studies, reaching 3.8, 3.7, 3.1, 2.6 and 1.7 units log for P. aeruginosa, E. coli, Enterococcus sp., C. perfringens and Staphylococcus sp., respectively.

The authors would like to acknowledge the support obtained from the following Brazilian institutions: Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq (Project number 429567/2016-2); Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES for the PhD grant awarded to Graziele Ruas (88882.458517/2019-01 and sandwich period 88881.190564/2018-01); Fundação de Amparo à Pesquisa do Estado de Minas Gerais – FAPEMIG; Instituto Nacional de Ciência e Tecnologia em Estações Sustentáveis de Tratamento de Esgoto – INCT ETEs Sustentáveis (INCT Sustainable Sewage Treatment Plants).