Influence of hydraulic loading rate and recirculation on oxygen transfer in a vertical flow constructed wetland

Science of the Total Environment 668 (2019) 988–995

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Influence of hydraulic loading rate and recirculation on oxygen transfer in a vertical flow constructed wetland Samara T. Decezaro a,⁎, Delmira B. Wolff b, Catiane Pelissari c, Rolando J.M.G. Ramírez b, Thiago A. Formentini b, Janaína Goerck b, Luiz F. Rodrigues d, Pablo H. Sezerino c a

Department of Engineering and Environmental Technology, Federal University of Santa Maria, Frederico Westphalen, Brazil Department of Sanitary and Environmental Engineering, Federal University of Santa Maria, Santa Maria, Brazil GESAD - Decentralized Sanitation Research Group, Department of Sanitary and Environmental Engineering, Federal University of Santa Catarina, Florianópolis, Brazil d Institute of Petroleum and Natural Resources, Pontifical Catholic University of Rio Grande do Sul, Porto Alegre, Brazil b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Gas tracer to assess the oxygen transfer rate (OTR) in a vertical flow wetland • OTR much higher than oxygen consumption rates (OCR) estimated by stoichiometry • OTR decrease with hydraulic loading rate increase

a r t i c l e

i n f o

Article history: Received 2 January 2019 Received in revised form 20 February 2019 Accepted 4 March 2019 Available online 06 March 2019 Editor: Paola Verlicchi Keywords: Propane Oxygen transfer rate Dissolved oxygen Gas tracer method Aeration Subtropical climate

a b s t r a c t The oxygen transfer rate (OTR) has a significant impact on the design and operation of vertical flow constructed wetlands (VFCWs) intended for organic matter removal and nitrification. Despite its key role, the information on real oxygen input in VFCWs is limited, being usually estimated by mass balance (stoichiometry), through which is calculated only the oxygen consumption rate (OCR). In this study, for the first time, the gas tracer method was applied to evaluate the oxygen transfer capacity of a real-scale VFCW (24.5 m2) applied to the treatment of domestic wastewater. Propane was used as tracer. The OCR and the OTR were evaluated in VFCW under hydraulic loading rates (HLR) of 60, 90, and 120 mm d−1 corresponding to recirculation rations of 0%, 50%, and 100%. The OTR in standard conditions (20 °C) ranged from 120 to 176 g O2 m−2 d−1. The highest OTR was found for the lowest HLR. For the operating conditions tested, the OTR obtained with gas tracer were higher than the OCR calculated by stoichiometry in VFCW, which ranged from 20.6 to 27.8 g O2 m−2 d−1. Besides, the OTR were sufficient to satisfy the VFCW oxygen demand for organic matter removal and nitrification. These results show that the gas tracer method for OTR determination may allow advances on the understanding of treatment processes and on the design of new VFCWs since its treatment performance requires aerobic conditions. © 2019 Published by Elsevier B.V.

⁎ Corresponding author. E-mail address: [email protected] (S.T. Decezaro).

https://doi.org/10.1016/j.scitotenv.2019.03.057 0048-9697/© 2019 Published by Elsevier B.V.

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995

1. Introduction

2. Material and methods

Constructed wetlands (CW) are wastewater treatment systems widely used under diverse technological arrangements, operational conditions, and purposes (Stefanakis et al., 2014). Vertical subsurface flow constructed wetlands (VFCW) are applied to promote the removal of carbonaceous organic matter and nitrification (Saeed and Sun, 2012). This treatment performance is achieved due to i) the great oxygen supply provided by the intermittent feeding and ii) the control of the hydraulic loading rate (HLR). The oxygen transfer into VFCWs occurs naturally, i.e., with no need for external sources of energy. This transfer occurs primarily by atmospheric reaeration, convection, diffusion, intermittent feeding, and secondarily by plants (Platzer, 1999; Tanner and Kadlec, 2003; Kadlec and Wallace, 2009; Liu et al., 2016). Ye et al. (2012) found that more than 99.9% of the dissolved oxygen in a VFCW with Phragmites came from atmospheric reaeration, whereas less than 0.1% came from plants release and inflow with influent. Nivala et al. (2013) reported oxygen transfer by plants in the range of 0.005 to 12 g O2 m−2 d−1. Despite its key role, information on oxygen transfer in VFCWs is scarce and, when available, discrepancies are observed. This occurs because, in most studies, the oxygen transfer rates (OTR) are deduced (by mass balance - stoichiometry) by water quality data (Brix et al., 2002; Cooper, 2005; Platzer, 1999). In this case, the system oxygen consumption rate (OCR) is estimated, rather than the real OTR. Hence, the term “oxygen transfer” should only be used when referring to the total amount of oxygen that is physically transferred from the atmosphere to the subsurface of a CW (Nivala et al., 2013). A gas tracer is, ideally, a chemically and biologically inert gas whose transfer rate is proportional to the oxygen transfer rate (Tyroller et al., 2010). Gas tracers have been used to assess the OTR in wastewater treatment systems such as trickling filters (Vasel and Schrobiltgen, 1991; Vieira, 2013) and rotating biological contactors (Boumansour and Vasel, 1998). Nevertheless, the gas tracer method has not been used to assess the OTR in VFCWs. The application of this method in VFCWs, which unlike other CW modalities are reactors with unsaturated media, may provide information about the real aeration capacity of these systems. This study aimed at evaluating the oxygen transfer capacity of a realscale VFCW applied to the treatment of domestic wastewater under different operational conditions. To the best of our knowledge, this is the first time that the gas tracer method is applied to assess the OTR in VFCWs. This knowledge may allow advances to the design of new VFCWs based on the oxygen balance since its treatment performance (organic matter removal and nitrification) is related to the maintenance of aerobic conditions.

2.1. Wastewater treatment plant

989

The domestic wastewater treatment plant was installed in 2015 in Santa Maria city, in southern Brazil, under subtropical climate (latitude: −29.7175; longitude: −53.7132). It was composed by a septic tank (working volume of 4.7 m3) operating as primary treatment followed by a VFCW (Fig. 1). The VFCW was 7.0 m long, 3.5 m wide (surface area = 24.5 m2), and 0.75 m deep, with gravel (19 mm and 25 mm) as bed media. The influent distribution piping in the VFCW was drilled and arranged in three parallel and equidistant lines, installed 50 cm above the gravel surface to facilitate the pipes clean (Figs. 1 and 3). The drainage piping was also drilled and arranged in three parallel and equidistant lines, installed 5 cm from the bottom of the VFCW. Ventilation pipes were connected to the drainage piping with one pipe end open to the atmosphere. The VFCW was operated over two years in three operational periods with different recirculation rates (RR): 0% RR in period I, 50% RR in period II, and 100% RR in period III. These resulted in HLR from 60 to 120 mm d−1 and organic loading rates (OLR) from 19 to 47 g COD m−2 d−1 (Fig. 2). Throughout the three periods, the VFCW was operated at intermittent flow, consisting of eight pulses per day, from 8:30 a.m. to 10:30 p.m. with a two-hour interval between each pulse. The samples were taken from the influent and from the VFCW effluent. All samples were collected at 8:30 a.m. Physicochemical parameters (chemical oxygen demand – COD and total Kjeldahl nitrogen – TKN) were analyzed according to the Standard Methods for the Examination of Water and Wastewater (APHA, 2012). 2.2. Estimation of oxygen consumption rate (OCR) using stoichiometry An oxygen mass balance was performed taking into account the HLR, OLR, and TKN loading rates applied in the VFCW. The oxygen consumption was estimated based on the energy amount required to degrade the carbonaceous organic matter and to oxidize the ammonia in the three operational periods. The OCR for periods I, II and III was calculated using Eq. (1) (Platzer, 1999). Stoichiometrically, 4.3 kg of oxygen is necessary for the complete nitrification of 1 kg of TKN, and a factor of 0.7 is applied to the COD values to estimate the biodegradable COD, thus suggesting that 0.7 kg of oxygen is consumed for each kg of COD removed. The oxygen release due to denitrification was not considered in the calculation, neither was the oxygen demand for sulfide oxidation (H2S) to sulfate (SO42−).

OCR ¼

½0:7ðΔM COD Þ þ 4:3ðΔMTKN Þ A

Fig. 1. Diagram of the wastewater treatment plant.

ð1Þ

990

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995

Fig. 2. Operational conditions applied in periods I, II and III.

where OCR is the oxygen consumption rate in g m−2 d−1; ΔM is the load removed for COD or TKN (=(Qin × Cin) − (Qout × Cout)) in g d−1; Qin is the influent flow in m3 d−1; Qout is the effluent flow in m3 d−1; Cin is the influent COD or TKN concentration in mg L−1 = g m−3; Cout is the effluent COD or TKN concentration in mg L−1 = g m−3; and A is the surface area of the VFCW in m2.

2.3. Determination of the oxygen transfer rate (OTR) using gas tracer After operational periods I, II and III were concluded, determination of the OTR in the VFCW was conducted using the same operational conditions of each period. At the time of the OTR experiments, the VFCW filtering mass presented porosity at 49%, and the plants (Canna spp.) were well developed, with the roots extended through the gravel. The experiments were performed after feeding the VFCW

with clean water for 24 h to obtain an endogenous state (Vasel and Schrobiltgen, 1991). Propane gas (99.5% pure) was used as the gas tracer. It was obtained and used in a 45-kg cylinder equipped with a two-stage pressure regulator, flowmeter (rotameter), and a silicone rubber diffuser hose (Fig. 3). The propane was injected into a 500-L tank containing clean water until the dissolved oxygen (DO) concentration was steadily close to zero. The water containing propane was then pumped into the VFCW, thus simulating a wastewater pulse. The samples for propane determination were collected using 10-mL syringes coupled to a needle and a three-way stopcock to avoid gas loss. The water containing propane was collected in different points of VFCW: (i) directly from the distribution piping (point 2, Fig. 3); (ii) in the surface layer of the gravel (point 3, Fig. 3); and (iii) in the drainage pipe (point 4, Fig. 3), at four different percolation times (at the maximum flow rate and after one, three, and six average percolation

Fig. 3. Oxygen transfer evaluation in the VFCW. 1 - DO and temperature determination in water tank; 2 - sampling of influent directly from the distribution pipe (propane and DO determination); 3 - sampling in the surface layer of the gravel (propane and DO determination); 4 - sampling of effluent in the drainage pipe (propane, DO and temperature determination).

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995 Table 1 Operational conditions in which the experiments were performed, referring to the application of a pulse in the VFCW in different HLRs. Experiment E1 E2 E3

HLR (mm d−1)

HLR (mm min−1)

Feeding time (min)

Batch thickness (mm)

60 90 120

7.6 5.8 5.1

1 2 3

7.6 11.5 15.3

times). Samplings were made in triplicate. The volume collected with the syringes was transferred to 20-mL amber glass vials with screw caps. Samples were immediately cooled to 4 °C and sent to the laboratory for propane determination. Dissolved oxygen concentrations and temperature in points 1, 2, 3 and 4 (Fig. 3), as well as the air temperature, were monitored during the gas tracer experiments. Dissolved oxygen concentrations were determined using the Winkler Method (APHA, 2012), after collecting the samples into 300-mL sealed bottles. Periodic flow measurements were took at the VFCW inlet and outlet. Each experiment lasted 120 min, which correspond to a drainage period of a wastewater pulse in the VFCW. Three experiments (denoted E1, E2, and E3) were conducted to evaluate the OTR, corresponding to the operational conditions of periods I (60 mm d−1), II (90 mm d−1) and III (120 mm d−1), respectively (Table 1). The propane determination was made using headspace gas chromatography (Rodrigues et al., 2014), using a gas chromatograph (GC-2014 Shimadzu) equipped with FID detector which is used to quantify mainly hydrocarbons. The samples were shaken vigorously to release dissolved gas and homogenize the headspace, then left standing until they reached room temperature. An aliquot (3 mL to 5 mL) was collected from the vial using the polypropylene syringe and injected into the equipment to fill the 100-μL loop. The temperature program started the column at 150 °C, maintained for 3 min, then heated (10 °C min−1) until 220 °C and maintained for 25 min, totaling 35 min of analysis. The retention time of propane was 10.9 min. The propane concentration in the headspace of the samples was obtained using high purity propane standards. These concentrations were then converted into propane percentages in the headspace and thereafter the dissolved propane concentrations in the samples were calculated (Vieira, 2013).

991

One-minute resolution propane concentrations in the effluent were calculated using the propane concentrations in the four samples collected in each experiment (corresponding to the maximum flow rate and after one, three, and six average percolation times). A firstorder kinetic equation [C = Co · exp(−K · t)] and Solver (Microsoft Excel) were used for the simultaneous estimation of K and Co by nonlinear regression. Then, the propane effluent loads were determined. Secondly, the oxygen mass transfer coefficients (KLa, O2) were calculated using the KLa, P previously obtained. The calculation was performed using Eq. (3). K L a; O2 ¼ R  K L a; P

ð3Þ

where R is the dimensionless ratio between oxygen and propane mass transfer coefficients (1.39; Rathbun et al., 1978) and KLa,O2 and KLa,P are the oxygen and propane mass transfer coefficients in h−1, respectively. Then, the calculated KLa,O2 were standardized to 20 °C using Eq. (4) (Metcalf and Eddy, 2016): K L a ðT Þ; O2 ¼ K L a ð20° CÞ; O2  ð1024ÞðT−20Þ

ð4Þ

The oxygen transfer rates in standard conditions (OTRstandard) were calculated using the standardized coefficient - KLa (20 °C), O2 by Eq. (5), which requires the saturation concentration at 20 °C and at the sea level (Csat = 9.1 mg L−1) (Metcalf and Eddy, 2016). This standard condition corresponds to the higher oxygen deficit (Ct = 0 mg L−1), and relates KLa,O2 with the system dimensions. The OTRstandard values were obtained in g O2 m−2 h−1 and subsequently in g O2 m−2 d−1. OTRstandard ¼ ½K L að20° CÞ; O2  ðC sat −C t Þ  V =A

ð5Þ

where V is the real wetland volume (considering the porosity in the filter material) in m3 and A is the wetland surface area in m2. The OTRs for the real conditions of the study (OTRfield) were calculated using Eq. (6). We used the KLa,O2 from Eq. (3) and the Csat in the VFCW's real altitude in the liquid average temperature recorded during the tests and adjusted to the wastewater. It was assumed that the Csat in wastewater is about 95% of that of clean water (β = 0.95; Von Sperling and Chernicharo, 2005). The Ct value assumed for each experiment, in the study field, was the average between the DO concentration in the inlet (gravel surface) and the outlet of the VFCW. OTRfield ¼ ½K L a; O2  ðC sat −C t Þ  V =A

2.3.1. Calculation of oxygen transfer rates (OTR) Firstly, the mass transfer coefficient for propane (KLa, P) was calculated for experiments E1, E2, and E3 using the first order kinetics equation for piston flow (Eq. (2)):

ð6Þ

3. Results and discussion 3.1. Estimated oxygen consumption rate (OCR) using stoichiometry

  P out 1 K L a; P ¼ − ln  t P in

ð2Þ

where Pout is the propane load dissolved into the effluent in mg pulse−1; Pin is the propane load dissolved into the influent in mg pulse−1; and t is the percolation time of a wastewater pulse in hours (hydrodynamic tests).

The OCRs in the VFCW during operating periods I, II and III ranged from 20.6 to 27.8 g m−2 d−1 (Table 2). These results are within the reported range (7.9 to 49.8 g m−2 d−1) for VFCW filled with gravel in Germany (Nivala et al., 2013). There was no direct relationship between OCRs and HLRs. The higher OCR was observed at an HLR of approximately 90 mm d−1. This coincided with the highest organic load applied in the

Table 2 Average oxygen consumption rates (OCR) for operating periods I, II and III of the system over two years of monitoring. Period

HLR (mm d−1)

Loads applied (g m−2 d−1) COD

I II III

60 90 120

19.1 47.0 40.0

Loads removed (g m−2 d−1) TKN 4.3 7.0 8.4

COD 12.5 22.8 12.2

TKN 2.8 2.7 2.8

Efficiencies (%)

Oxygen consumption rate – OCR (g m−2 d−1)

COD

TKN

COD degradation

Nitrification

Total

65 49 30

65 39 33

8.8 16.0 8.5

12.1 11.8 12.1

20.9 27.8 20.6

992

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995

Table 3 Oxygen consumption rates (OCRs) in constructed wetlands reported in the literature, based on mass balances. CW typea HFCW HFCW SFCW VFCW VFCW VFCW VFCW VFCW VFCW VFCW VFCW VFCWb

OCR (g m−2 d−1) 5.86 3.87–11.66 1.66–5.27 23–64 28–35 55 80 5.77–18.45 24–213 7.90–49.80 20.60–27.80 450

References Brix (1990) Gasiunas (2011) Gasiunas (2011) Platzer (1998) cited in Cooper (1999) Weedon (2003) Kayser and Kunst (2005) Kantawanichkul et al. (2009) Gasiunas (2011) Saeed et al. (2012) Nivala et al. (2013) This study Wu et al. (2011)

Table 4 Results calculated to evaluate the oxygen transfer rate (OTR) of the VFCW in experiments E1, E2 and E3 carried out with HLR variation. Parameter

Percolation time (h−1) Pin (mg)a Pout (mg) Pout/Pin KLa, P (h−1) KLa, O2 (h−1) KLa, O2 (20 °C) (h−1) OTRfield (g m−2 d−1) OTRstandard (g m−2 d−1)

Experiment E 1 (pulse of 7.5 mm/HLR of 60 mm d−1)

E 2 (pulse of 11 mm/HLR of 90 mm d−1)

E 3 (pulse of 15 mm/HLR of 120 mm d−1)

2 1509 46 0.030 1.74 2.42 2.19 98 176

2 1100 68 0.062 1.39 1.93 1.71 71 137

2 1095 95 0.087 1.22 1.70 1.50 63 120

a Constructed wetland (CW) type: surface flow (SFCW), horizontal flow (HFCW), and vertical flow (VFCW); b Tidal flow.

a The influent propane loads (Pin) considered in the calculations were those found in the surface layer of the gravel (point 3).

VFCW (47 g m−2 d−1). Increasing the HLR from 90 to 120 mm d−1, the COD removal was lower. However, there was a negative correlation between removal efficiencies and HLRs (Table 2). COD removal efficiencies were 65%, 49%, and 30% for HLRs of 60, 90, and 120 mm d−1, respectively. Similarly, TKN removal efficiencies were 65%, 39%, and 33% for HLRs of 60, 90, and 120 mm d−1, respectively. High HLRs may negatively interfere in the organic matter removal since they provide less contact time of the wastewater with the microorganisms (Torrens et al., 2009). Furthermore, it has already been reported that the application of excessively high HLRs (higher than 0.6 m d−1 for single stage French VFCW, for example) may decrease oxygen input in the CW (Millot et al., 2016). High organic loads might be detrimental to nitrification in VFCWs (Wu et al., 2011) due to the competition for oxygen between autotrophic (nitrifying) and heterotrophic bacteria responsible for the carbonaceous organic matter degradation (Saeed and Sun, 2012; Li et al., 2014). In this study, the OCR for nitrification was roughly the same (i.e. about 12 g m−2 d−1) for the operating periods I, II and III. This shows that the nitrification velocity was not affected by the variation of the OLRs and the HLRs. They were low in this study; therefore, higher HLRs and OLRs must be tested to assess the limits of the system. Additionally, it suggests that the nitrification was limited by the nitrogen adsorption capacity in the VFCW. During the VFCW feeding periods (pulses), the NH4+ ions are adsorbed by the negatively charged filter material surface. During the drainage, the air goes through the material's pores, and the nitrification of the adsorbed NH4+ occurs. In the next pulse, NO3− and NO2− are desorbed in the water flow. In this study, the N-NO3− and N-NO2− effluent values were 1.8, 1.9, and 3.4 g m−2 d−1 for the operating periods I, II, and III, respectively. These values correspond to

about 30% to 40% of the influent TKN (Table 2). Likewise, it has already been demonstrated, in a French VFCW system, that 1/3 of the influent TKN is adsorbed during the feeding period and later nitrified during the pause periods (Morvannou et al., 2014). Furthermore, in “tidal flow” systems, nitrification has been demonstrated during the drainage periods, which theoretically reduce the competition for oxygen within biofilms without impairing nitrification (Austin and Nivala, 2009). Table 3 shows a comparison between the OCRs reported in the literature and those observed in this study. The OCR in CW depends on several factors, such as the oxygen transfer of these systems, the organic and nitrogen loads applied, the contact time between the wastewater and microorganisms, and the characteristics of the filtering material. VFCWs have the highest OCRs. If there is enough oxygen as higher carbon and nitrogen loads are applied, the greater the oxygen consumption is in the system. However, this consumption is limited to the contact time of the effluent with the treatment microorganisms, which in turn depends on the HLR as well as the porosity and permeability of the filter material. An OCR of 213 g m−2 d−1 (i.e. about eight times higher than those observed in this study) was reported by Saeed et al. (2012) in a VFCW filled with coco-peat and applied to tannery wastewater treatment (high organic loading rate - 690 g COD m−2 d−1). The OCR was related to the high porosity of the coco-peat media, of about 50%, which may have allowed the larger oxygen input. According to Platzer (1999), the oxygen diffusion rate changes according to the type of filtering soil/material used since it is related to physical parameters such as effective diameter (d10) and porosity. Moreover, a high OCR (450 g m−2 d−1) was reported in the study by Wu et al. (2011) in a “tidal flow” VFCW

Fig. 4. Propane and dissolved oxygen concentrations in different points of VFCW: directly from the distribution piping (point 2); in the surface layer of the gravel (point 3); and in the drainage pipe (point 4).

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995

Fig. 5. Gas tracer curves obtained in experiments E1, E2 and E3.

(known as intensified CW), whose operation mode theoretically provides elevated oxygen input. 3.2. Determination of the oxygen transfer rate (OTR) using gas tracer The propane concentrations in the VFCW influent (point 2, Fig. 3), measured in the distribution piping, ranged from 5 to 10 mg L−1 (Fig. 4). These values were similar to those obtained by Vieira (2013) and Tyroller (2008) in tests using propane as a tracer, and below the propane saturation concentration, which is approximately 74 mg L−1 (Tyroller et al., 2010). The propane concentrations in the gravel layer located on the VFCW surface (point 3, Fig. 3) were lower than those found in the distribution pipes (point 2, Fig. 3). The opposite was observed for the dissolved oxygen (Fig. 4), whose concentrations on the VFCW surface (point 3, Fig. 3) were higher compared to those in the pipes (point 2, Fig. 3). This means that there was an effective oxygenation in the liquid's path from the distribution piping to the VFCW surface. Thereby, the raised distribution piping demonstrated being a significant way of transferring oxygen in VFCW. On the other hand, the concentrations of propane and dissolved oxygen in the VFCW effluent (point 4, Fig. 3) were lower than in the influent, thus suggesting the mass transfer of these gases from the water to the pores of the filtering material and to biofilm. The OTRs for the field and standard conditions at 20 °C calculated for operating periods I, II and III are presented in Table 4. The OTRs assessed in standard conditions were 176, 137, and 120 g O2 m−2 d−1 for the HLRs of 60, 90, and 120 mm d−1, respectively. It is well known that oxygen transfer rate may be far higher as compared to the consumption rate (Kadlec and Wallace, 2009; Nivala et al., 2013). This study corroborates with this assertion since the OTRs were more than five times higher as compared to OCRs. The propane loads dissolved into the influent (Pin) were 1509, 1100, and 1095 mg pulse−1 for E1, E2, and E3, respectively (Table 4). The highest amount of dissolved propane in the influent water was in experiment 1, when the VFCW received the lowest HLR (60 mm d−1). The resulting gas tracer curves are presented in Fig. 5. The total propane loads dissolved into the effluent (Pout) were 46, 68 and, 95 mg pulse−1

993

for E1, E2, and E3, respectively (Table 4). The Pout/Pin ratios suggest the oxygen transfer capacity of the VFCW. The lower this ratio is, the higher the OTRs in a reactor for the same hydraulic detention (or percolation) time according to Eq. (2). In this study, the Pout/Pin ratio was lower the lower the HLR (0.030 for HLR of 60 mm d−1, 0.062 for HLR of 90 mm d−1 and 0.087 for HLR of 120 mm d−1; Table 4). Vasel and Schrobiltgen (1991), Pinheiro et al. (2012), and Vieira (2013) found the same trend. However, this fact also can be related to the higher propane amount in the influent on smaller HLRs, given that the gas dissolution in water sometimes is not so efficient with the application of larger flows. The propane mass transfer (KLa, P) had a small variation in the three experiments, with values of 1.74 h−1 for E1, 1.39 h−1 for E2 and 1.22 h−1 for E3 (Table 4). The equation used to calculate KLa, P (Eq. (2)) was the first-order kinetics equation for the plug flow reactor, similar to the used by Tyroller et al. (2010) for evaluating OTR in HFCW. This equation results in coefficient values lower than those calculated by dispersed flow (more complex) and complete mix equations. Yet, in a study carried out with a trickling filter, Vasel and Schrobiltgen (1991) evaluated different equations for calculating oxygen transfer and found similar results for the plug-flow and axial dispersed models. As the assumed percolation time (2-hour intervals between pulses) was the same for the three HLRs, the KLa values varied according to the Pout/Pin ratios (Table 4). The KLa, O2 (20 °C) values determined in the three experiments were 2.19 h−1, 1.71 h−1 and 1.50 h−1 for E1, E2 and E3, respectively. These values were within the range reported for a rotating biological contactor (Boumansour and Vasel, 1998; Chavan and Mukherji, 2008), as shown in Table 5. The highest value of KLa, O2, and consequently the highest OTR, was found in experiment 1 when the VFCW received the smallest pulse volume (7.5 mm) regarding the HLR daily fractionation (60 mm d−1) into eight pulses (Table 4). This behavior was the opposite of the expected. It had already been experimentally shown that high volumes quickly applied to the VFCW surface force the atmospheric oxygen to move downwards penetrating the filter material (Stefanakis et al., 2014). According to Platzer (1999), when applying pulses of short duration periods, the oxygen transfer rate by convection is proportional to the HLR. In this study, the highest OTR obtained in experiment 1 may have occurred due to the higher instantaneous inflow (7.6 mm min−1). This provided more water onto the filter and thus, higher infiltration velocity (higher infiltration rates) and turbulence. A comparison of the OTRstandard obtained in this study (176, 137, and 120 g O2 m−2 d−1 for E1, E2 and E3, respectively; Table 4) with those obtained in other CW is shown in Table 6. This study's VFCW demonstrated higher OTR values than the HFCW and SFCW. This may be explained by the way VFCWs work, with intermittent flow, which theoretically results in larger OTRs. In this type of flow, the alternation of wet and dry periods provides greater diffusion of atmospheric oxygen into the filtering material (Saeed and Sun, 2012). Schwager and Boller (1997) used the gas tracer method to evaluate oxygen transfer in a sand filter operated with intermittent flow. The sand filter evaluated showed an OTR of 56 g O2 m−2 d−1. In a gravel VFCW, a higher OTR is expected due to the larger pore size of the material and the presence of plants (Table 6). Besides, the recirculation in operating periods II and III may be led to an enhancement of the natural re-aeration as

Table 5 Values of KLa, O2 obtained in different studies. Reactor type

KLa,O2 (h−1)

Method

References

Trickling filters Rotating biological contactors (RBC) RBC RBC HFCW VFCW

9–70 2–16 1,68 0.66–4.0 0.013–0.132 1.50–2.19

Gas tracer Gas tracer OD measurement OD measurement Gas tracer Gas tracer

Vasel and Schrobiltgen (1991) Boumansour and Vasel (1998) Kubsad et al. (2004) Chavan and Mukherji (2008) Tyroller et al. (2010) This study

994

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995

Table 6 OTR in different CW types and sand filter. Systema

Methodb

OTR (g m−2 d−1)

CC GT CC GT GT

6.47–7.92 0.3–3.2 1.36–1.76 56 120–176

HFCW HFCW SFCW Sand filter VFCW

References Wu et al. (2001) Tyroller et al. (2010) Wu et al. (2001) Schwager and Boller (1997) This study

a System: Surface flow constructed wetland (SFCW), horizontal flow constructed wetland (HFCW), and vertical flow constructed wetland (VFCW). b OTR measurement method: closed chamber method (CC) and gas tracer method (GT).

suggested by other studies (Foladori et al., 2014; Foladori et al., 2013; Brix et al., 2002). The calculated OTRs for operating conditions I, II and III were enough to meet the oxygen demand for organic matter degradation plus nitrification. The total oxygen demands were on average 32, 63, and 64 g O2 m−2 d−1 for periods I, II and III, respectively. Thus, this study corroborates others that report that VFCWs are effective in oxygen transfer due to the intermittent feeding, primarily by the convection and diffusion processes (Platzer, 1999; Ye et al., 2012). Furthermore, the results suggest that the form of construction also had a relevant role in the OTR of the system because the elevated wastewater distribution piping (50 cm above the surface of the filter material) provided an additional oxygen input for the VFCW. Therefore, for HLRs varying from 60 to 120 mm d−1, oxygen was not the limiting element for the carbonaceous organic matter degradation and nitrification. Therefore, it is likely that the magnitude of these processes was more influenced by the contact time of wastewater and microorganism in the VFCW than by oxygen availability. 4. Conclusions For the operational conditions evaluated in this research, with HLRs of 60, 90, and 120 mm d−1 and daily HLR fractionation into eight pulses per day (two-hour intervals between the pulses), the VFCW showed an oxygen transfer capacity in standard conditions of 120 to 176 g O2 m−2 d−1. The OTRs obtained were higher than the consumption rates verified in VFCW (20.6 to 27.8 g O2 m−2 d−1), besides being sufficient to meet the oxygen demand for carbonaceous organic matter degradation and nitrification (32 to 64 g O2 m−2 d−1). The results of this research suggest that OTRs in VFCWs may be higher than the oxygen consumption calculated by mass balance. This indicates that, in many VFCWs, the oxygen consumption may be limited not by the low availability of this element in the environment but by construction and operation aspects that determine the contact time between pollutants and microorganisms. The gas tracer method was demonstrated to be a promising tool for evaluating the oxygen transfer capacity of VFCWs. Acknowledgements The authors would like to thank the Funding Authority for Research and Projects (FINEP) and the Coordination for the Improvement of Higher Education Personnel (CAPES) for funding this research. References APHA, 2012. Standard Methods for the Examination of Water and Wastewater. 22th ed. American Public Health Association, Washington, DC, USA. Austin, D., Nivala, J., 2009. Energy requirements for nitrification and biological nitrogen removal in engineered wetlands. Ecol. Eng. 35, 184–192. https://doi.org/10.1016/j. ecoleng.2008.03.002. Boumansour, B.-E., Vasel, J.-L., 1998. A new tracer gas method to measure oxygen transfer and enhancement factor on RBC. Water Res. 32 (4), 1049–1058. https://doi.org/ 10.1016/S0043-1354(97)00324-2.

Brix, H., 1990. Gas exchange through the soil-atmosphere interphase and through dead culms of Phragmites australis in a constructed reed bed receiving domestic sewage. Water Res. 24 (2), 259–266. https://doi.org/10.1016/0043-1354(90)90112-J. Brix, H., Arias, C. A., Johansen, N. H., 2002. BOD and nitrogen removal from municipal wastewater in an experimental two stage vertical flow constructed wetland system with recycling. In: Proceedings of the Eighth International Conference on Wetland Systems for Water Pollution Control, Arusha, Tanzania, 16–19 September 2002. Chavan, A., Mukherji, S., 2008. Dimensional analysis for modeling oxygen transfer in rotating biological contactor. Bioresour.Technol. 99, 3721–3728. https://doi.org/ 10.1016/j.biortech.2007.07.021. Cooper, P., 1999. A review of the design and performance of vertical-flow and hybrid reed bed treatment systems. Water Sci. Technol. 40 (3), 1–9. https://doi.org/10.1016/ S0273-1223(99)00414-X. Cooper, P., 2005. The performance of vertical flow constructed wetland systems with special reference to the significance of oxygen transfer and hydraulic loading rates. Water Sci. Technol. 51 (9), 81–90. https://doi.org/10.2166/wst.2005.0293. Foladori, P., Ruaben, J., Ortigara, A.R., 2013. Recirculation or artificial aeration in vertical flow constructed wetlands: a comparative study for treating high load wastewater. Bioresour. Technol. 149, 398–405. https://doi.org/10.1016/j.biortech.2013.09.099. Foladori, P., Ruaben, J., Ortigara, A.R.C., Andreottola, G., 2014. Batch feed and intermittent recirculation to increase removed loads in a vertical subsurface flow filter. Ecol. Eng. 70, 124–132. https://doi.org/10.1016/j.ecoleng.2014.04.006. Gasiunas, V., 2011. Oxygen transfer rate and BOB removal efficiency of constructed wetlands with dependence on construction. 8th International Conference on Environmental Engineering (ICEE) (Selected papers). Kadlec, R. H., Wallace, S., 2009. Treatment Wetlands. second ed. Boca Raton, Florida, EUA. Kantawanichkul, S., Kladprasert, S., Brix, H., 2009. Treatment of high-strength wastewater in tropical vertical flow constructed wetlands planted with Typha angustifólia and Cyperus involucratus. Ecol. Eng. 35, 238–247. https://doi.org/10.1016/j.ecoleng.2008.06.002. Kayser, K., Kunst, S., 2005. Processes in vertical-flow reed beds: nitrification, oxygen transfer and soil clogging. Water Sci. Technol. 51 (9), 177–184. https://doi.org/ 10.2166/wst.2005.0314. Kubsad, V., Chaudhari, S., Gupta, S.K., 2004. Model for oxygen transfer in rotating biological contactor. Water Res. 38, 4297–4304. https://doi.org/10.1016/j.watres.2004.08.016. Li, F., Lu, L., Zheng, X., Ngo, H.H., Liang, S., Guo, W., Zhang, X., 2014. Enhanced nitrogen removal in constructed wetlands: effect of dissolved oxygen and step-feeding. Bioresour. Technol. 169, 395–402. https://doi.org/10.1016/j.biortech.2014.07.004. Liu, H., Hu, Z., Zhang, J., Ngo, H.H., Guo, W., Liang, S., Fan, J., Lu, S., Wu, H., 2016. Optimizations on supply and distribution of dissolved oxygen in constructed wetlands: a review. Bioresour. Technol. 214, 797–805. https://doi.org/10.1016/j.biortech.2016.05.003. Metcalf & Eddy., 2016. Tratamento de efluentes e recuperação de recursos, fifth ed. Porto Alegre: AMGH. Millot, Y., Troesch, S., Esser, D., Molle, P., Morvannou, A., Gourdon, R., Rousseau, D.P., 2016. Effects of design and operational parameters on ammonium removal by single-stage French vertical flow filters treating raw domestic wastewater. Ecol. Eng. 97, 516–523. https://doi.org/10.1016/j.ecoleng.2016.10.002. Morvannou, A., Choubert, J.M., Vanclooster, M., Molle, P., 2014. Modeling nitrogen removal in a vertical flow constructed wetland treating directly domestic wastewater. Ecol. Eng. 70, 379–386. https://doi.org/10.1016/j.ecoleng.2014.06.034. Nivala, J., Wallace, S., Headley, T., Kassa, K., Brix, H., van Afferden, M., Müller, R., 2013. Oxygen transfer and consumption in subsurface flow treatment wetlands. Ecol. Eng. 61P, 544–554. doi:https://doi.org/10.1016/j.ecoleng.2012.08.028. Pinheiro, A., Faht, G., da Silva, M.R., 2012. Determinação do coeficiente de reaeração em rios através do uso do traçador gasoso GLP. Eng. Sanit. Ambient. 17 (1), 107–116. Platzer, C., 1998. Entwicklung eines Bemessungsansatzes zur Stickstoffelmination In Pjlantzenklaranlagen. Ph.D Thesis. Technische Universitat Berlin, Germany. Platzer, C., 1999. Design recommendation for subsurface flow constructed wetlands for nitrification and denitrification. Water Sci. Tech. 40 (3), 257–263. https://doi.org/ 10.1016/S0273-1223(99)00420-5. Rathbun, R.E., Tai, D.Y., Shultz, D.J., Stephens, D.W., 1978. Laboratory studies of gas tracers for reaeration. J. Environ. Eng. 104 (2), 215–229. Rodrigues, L.F., Laroque, D., Lourega, R., Heemann, R., Ketzer, J.M., 2014. An alternative gas chromatography setting for geochemical analysis. J. Chem. Eng. Process Technol. 5 (6), 1–4. https://doi.org/10.4172/2157-7048.1000208. Saeed, T., Sun, G., 2012. A review on nitrogen and organics removal mechanisms in subsurface flow constructed wetlands: dependency on environmental parameters, operating conditions and supporting media. J. Environ. Manag. 112, 429–448. https://doi. org/10.1016/j.jenvman.2012.08.011. Saeed, T., Afrin, R., Al Muyeed, A., Sun, G., 2012. Treatment of tannery wastewater in a pilot-scale hybrid constructed wetland system in Bangladesh. Chemosphere. 88 (9), 1065–1073. https://doi.org/10.1016/j.chemosphere.2012.04.055. Schwager, A., Boller, M., 1997. Transport phenomena in intermittent filters. Water Sci. Technol. 35 (6), 13–20. https://doi.org/10.1016/S0273-1223(97)00090-5. Stefanakis, A.I., Akratos, C.S., Tsihrintzis, V.A., 2014. Vertical Flow Constructed Wetlands: Eco-engineering Systems for Wastewater and Sludge Treatment. Newnes. Tanner, C.C., Kadlec, R.H., 2003. Oxygen flux implications of observed nitrogen removal rates in subsurface-flow treatment wetlands. Water Sci.Technol. 48 (5), 191–198. https://doi.org/10.2166/wst.2003.0317. Torrens, A., Molle, P., Boutin, C., Salgot, M., 2009. Impact of design and operation variables on the performance of vertical-flow constructed wetlands and intermittent sand filters treating pond effluent. Water Res. 43, 1851–1858. https://doi.org/10.1016/j. watres.2009.01.023. Tyroller, L., 2008. Method for the Determination of Oxygen Transfer in a Subsurface Flow Constructed Wetland. Minor thesis. Polytechnic University of Catalonia. Barcelona, Spain.

S.T. Decezaro et al. / Science of the Total Environment 668 (2019) 988–995 Tyroller, L., Rousseau, D.P., Santa, S., García, J., 2010. Application of the gas tracer method for measuring oxygen transfer rates in subsurface flow constructed wetlands. Water Res. 44, 4217–4225. https://doi.org/10.1016/j.watres.2010.05.027. Vasel, J.L., Schrobiltgen, P., 1991. Oxygen transfer in trickling filters. Water Res. 25 (1), 53–60. https://doi.org/10.1016/0043-1354(91)90098-B. Vieira, P. de C., 2013. Estudo do comportamento de filtro biológico percolador com laterais abertas aplicado ao pós-tratamento de efluente de reator UASB. Tese (Doutorado) – Programa de Pós-Graduação em Saneamento, Meio Ambiente e Recursos Hídricos, Universidade Federal de Minas Gerais, Belo Horizonte. Von Sperling, M., Chernicharo, C.A. de L., 2005. Biological Wastewater Treatment in Warm Climate Regions. v. IWA Publishing, London, p. 1.

995

Weedon, C.M., 2003. Compact vertical flow constructed wetland systems – first two years performance. Water Sci.Technol. 48 (5), 15–23. https://doi.org/10.2166/wst.2003.0269. Wu, M., Franz, E.H., Chen, S., 2001. Oxygen fluxes and ammonia removal efficiencies in constructed treatment wetlands. Water Environ. Res. 73 (6), 661–666. Wu, S., Jeschke, C., Dong, R., Paschke, H., Kuschk, P., Knöller, K., 2011. Sulfur transformations in pilot-scale constructed wetland treating high sulfate-containing contaminated groundwater: a stable isotope assessment. Water Res. 45 (20), 6688–6698. https://doi.org/10.1016/j.watres.2011.10.008. Ye, J., Wang, L., Li, D., Han, W., Ye, C., 2012. Vertical oxygen distribution trend and oxygen source analysis for vertical-flow constructed wetlands treating domestic wastewater. Ecol. Eng. 4, 8–12. https://doi.org/10.1016/j.ecoleng.2011.12.015.