Effects of fillers combined with biosorbents on nutrient and heavy metal removal from biogas slurry in constructed wetlands

Effects of fillers combined with biosorbents on nutrient and heavy metal removal from biogas slurry in constructed wetlands

Science of the Total Environment 703 (2020) 134788 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www...

705KB Sizes 0 Downloads 5 Views

Science of the Total Environment 703 (2020) 134788

Contents lists available at ScienceDirect

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

Effects of fillers combined with biosorbents on nutrient and heavy metal removal from biogas slurry in constructed wetlands Xiongfei Guo a,b, Xingyi Cui a, Huashou Li a,⇑ a b

College of Resources and Environmental Sciences, South China Agricultural University, Guangzhou 510642, PR China College of Environmental Science and Engineering, China West Normal University, Nanchong 637009, PR China

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

 Fillers combined with biosorbents

enhanced nutrients removal rates in constructed wetland.  The removal rates : As (35.38%– 83.89%), Zn (8.15%–23.69%) and Cu (0.32%–0.88%).  Biochar is effective in reducing As uptake by wetland plants.  The treated wetland plant water spinach could be used as soilages.

a r t i c l e

i n f o

Article history: Received 23 July 2019 Received in revised form 30 September 2019 Accepted 1 October 2019 Available online 3 November 2019 Editor: Huu Hao Ngo Keywords: Constructed wetland Biosorbents Fillers Heavy metal removal Nutrient removal

a b s t r a c t The performance of fillers (biochar and zeolite) and their combinations with biosorbents (compound microbial agent and chlorella) in nutrients and heavy metals removal from biogas slurry in constructed wetlands (CWs) planted water spinach (Ipomoea aquatica) and plant uptake of heavy metals was investigated. The results demonstrated that the removal rate of nutrients in CWs was all above 60%. COD removal efficiencies were not significantly affected by fillers and biosorbents, all above 80%. The removal rates of TN and NH+4-N were the highest when the two fillers and two biosorbents were added, and the combination of biochar and chlorella presented the optimal removal effect on TP. The efficiency of removing heavy metals from biogas slurry in CWs was As > Zn > Cu, and their removal rates were 35.38%– 83.89%, 8.15%–23.69% and 0.32%–0.88%, respectively. The removal efficiency of As by the combination of biochar and composite microbial agent was high. The combination of the two fillers and two biosorbents had the best effect on reducing Cu and Zn enrichment in the aboveground part of water spinach in each treatment, while biochar alone had the best effect on reducing As enrichment in the aboveground and underground parts of water spinach. This study can provide a basis for the application of fillers and biosorbents in the treatment of biogas slurry in livestock and poultry farms in wetlands. Ó 2019 Elsevier B.V. All rights reserved.

1. Introduction As a low-cost and simple water treatment method, constructed wetlands (CWs) have been widely applied in the treatment of ⇑ Corresponding author. E-mail addresses: [email protected] (X. Guo), [email protected] (X. Cui), [email protected] (H. Li). https://doi.org/10.1016/j.scitotenv.2019.134788 0048-9697/Ó 2019 Elsevier B.V. All rights reserved.

domestic sewage, industrial wastewater and urban storm runoff. Fillers are an important part of CWs, playing a major role in the removal of nitrogen, phosphorus and heavy metals in the water (Liu et al., 2015). In CWs, fillers can remove pollutants from the water by physical–chemical reactions such as adsorption, precipitation and filtration, and can also promote biological nitrogen and phosphorus removal by providing suitable conditions for microbial attachment and plant growth (Xu et al., 2019). Different

2

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

fillers have different abilities to remove pollutants because of their different properties. Biochar is a solid carbon-rich product produced by thermochemical conversion of biomass under oxygenlimited conditions and a good adsorbent due to its large specific surface area. At present, its application in water treatment has attracted wide attention (Wang et al., 2015). Zeolite is an ore mainly composed of CaCO3, has strong selective adsorption for NH+4-N (Xu et al., 2019) and can also be used to remove heavy metals from sewage (Merrikhpour and Jalali, 2013). Zeolite and biochar are both common fillers for CWs which have good treatment effect on nitrogen-containing wastewater and heavy metals (Merrikhpour and Jalali, 2013; Hina et al., 2015; Álvarez-Rogel et al., 2018). In addition, it is well known that biosorption is a cost-effective technology for removing heavy metals from wastewater. It uses organisms and their derivatives, such as fungi, algae and some cell extracts, as biosorbents to absorb heavy metal ions in the water (Neeraj et al., 2016), thus realizing the removal of heavy metal ions. Among the biosorbents, bacterial biosorbents, fungal biosorbents and algae biosorbents have been studied extensively. Microalgae grow fast and have strong environmental adaptability. Nutrients can be absorbed and utilized during their growth, so as to reduce main pollution indexes such as nitrogen and phosphorus and achieve biogas slurry treatment. The technology of applying microalgae to treatment of biogas slurry from swine manure has been confirmed by scholars at home and abroad (de Godos et al., 2009). Microbial agents are usually composed of various microorganisms. Functional microorganisms are added to enhance the degradation of pollutants in order to remove specific pollutants. Microorganisms are main undertakers of organic matters and nitrogen-phosphorus compounds in sewage purification in CWs. Denitrification-nitrification of microorganisms is the main way for nitrogen removal (Faulwetter et al., 2009). Further, biosorbents such as algae, bacteria, fungi and yeasts have been proved to be potential biological carriers for metals due to the solid solubility of metals, which can reduce the concentration of heavy metal ions in solutions (Nilanjana et al., 2008). As an important component of CWs, the selection of wetland plant species plays an important role in the effect of sewage purification (Maucieri et al., 2018). At present, there are certain limitations about the study on the removal of pollutants from wastewater by CWs. Firstly, most previous studies only considered the absorption of wetland plants, the adsorption and precipitation of fillers or the biological purification of algae and microorganisms, but no attention was paid to the synergistic effect of combined plant-filler-biosorbent systems. Secondly, previous studies mainly focused on heavy metals or nitrogen and phosphorus pollution (Bruch et al., 2011). However, during treatment, many kinds of pollutants exist in wastewater, especially compound wastewater containing heavy metals and nutrients, which limits the development of CWs to purify the compound polluted wastewater. Therefore, it is of great significance to combine plants, fillers and biosorbents to synergistically exert their respective purification functions specific to different pollutants, so as to improve the purification effect on compound polluted wastewater containing multiple pollutants. In this study, with zeolite and biochar as fillers, chlorella and compound microbial agent as biosorbents and water spinach as a hydroponic wetland plant, the purification of COD, TP, NH+4-N, As, Cu and Zn in biogas slurry in pig farms and the enrichment of As, Cu and Zn in water spinach were investigated using CWs when the two fillers were added alone and combined with the two biosorbents, aiming to explore the synergistic effect of different fillers and their combinations with biosorbents on pollutant removal in CWs. On this basis, the constructed wetland system with favorable purification effect on biogas slurry can be screened out, which lays a theoretical foundation for improving the overall treatment

capacity of the CWs for livestock and poultry wastewater containing biogas slurry in the farms, enriching and developing the sewage treatment process of the constructed wetland system, and achieving the standard discharge of biogas slurry in pig farms. 2. Materials and methods 2.1. Biogas slurry and wetland plants The experimental biogas slurry was from Desheng Agriculture and Animal Husbandry Development Co., Ltd. (Guangzhou, Guangdong). Wastewater discharged from pig farms was subjected to anaerobic fermentation after solid–liquid separation. The collected biogas slurry was preserved in sealed condition at room temperature (25 ± 5) °C until no gas production was observed. After natural precipitation for 2d, the supernatant was taken for experiment. The experiment was divided into two phases (see Section 2.4). The properties of biogas slurry in different phases were shown in Table 1. The experimental wetland plant was water spinach (Thailand narrow-leaved water spinach) which was purchased from the Vegetable Research Institute of Guangdong Academy of Agricultural Sciences. 2.2. Fillers Biochar and zeolite were used as fillers for CWs. Biochar was provided by Liaoning Biochar Engineering and Technology Research Center. It was prepared by pyrolysis of coconut shell under anoxic conditions. Ash content of the charcoal was removed using a 100-mesh sieve for further use. Zeolite was mined from Faku County, Liaoning Province, mainly containing clinoptilolite, a small amount of feldspar and other mineral components. The main proximate analysis and chemical properties of biochar and zeolite are shown in Table 2. 2.3. Biosorbents Chlorella and compound microbial agent were used as biosorbents. Purchased from Guangdong Fisheries Institute, the chlorella (1:10) was added to a nitrogen-free BG11 medium during the logarithmic growth period and cultured for 10 d till the logarithmic growth period under light intensity of 2700 Lx and temperature at 28 °C. Then, it was centrifugated in a 100-mL centrifuge tube at 3000 rpm and precipitation was collected for weighing. Compound microbial agent was a special strain for sewage treatment in Shenzhen Shipurui Biotechnology Co., Ltd., mainly containing Bacillus subtilis, yeasts, lactic acid bacteria, flocculating bacteria, etc. It was powdery, with the total number of viable bacteria 5.0  109 cfu g1. 2.4. Experimental design The experiment was carried out in transparent glass chambers with length  width  height = 0.5 m  0.3 m  0.4 m. The bottom layer of 4-cm river sand was used as the substrate to fix the roots of water spinach. Six plants of water spinach of the same size and growth (plant height, 15 cm) were cultured in each glass chamber, with plant spacing between two plants of 10 cm. In this study, the effects of fillers combined with biosorbents on nutrient and heavy metal removal from biogas slurry in constructed wetlands, as well as the uptake of heavy metals by wetland plants were discussed. The experiment lasted 30 d and was conducted in two phases, each for 15 d. The purpose of Phase 1 was to explore the purification effect of constructed wetlands on biogas slurry and the absorption of heavy metals by plants when wetland plants such as water spi-

3

X. Guo et al. / Science of the Total Environment 703 (2020) 134788 Table 1 Water quality of biogas slurry during different phases. Phase

TP/mgL1

NH+4-N/mgL1

COD/mgL1

pH

As/lgL1

Cu/mgL1

Zn/mgL1

1 2

24.72 19.40

416.91 358.32

11400.00 8627.45

8.40 8.4

107.64 71.75

7.29 7.20

0.84 0.62

Note: The water quality was assessed before the beginning of the experiment.

Table 2 The main proximate analysis and chemical properties of fillers. Fillers

pH

TC/gkg1

TN/gkg1

TP/gkg1

Ash content/%

Specific surface area/m2g1

CEC/cmolkg1

As/mgkg1

Cu/mgkg1

Zn/mgkg1

Biochar Zeolite

8.90 8.03

590.32 /

16.56 /

2.3 /

10.3 /

360 567.96

/ 160.8

56.32 0.03

9.7 0.27

55.6 0.78

nach were present. Phase 2 was to further study the purification effect of constructed wetlands on biogas slurry after harvesting the wetland plants. A total of 8 treatments (Table 3) and four transparent glass chambers with four replicates in each treatment were set up. Z, B, ZM, ZC, BM, BC, and ZBMC represented single applications of zeolite, single applications of biochar, combined applications of zeolite and microbial agents, combined application of zeolite and chlorella, combined applications of biochar and microbial agents, combined application of biochar and chlorella, and combined application of two kinds of fillers and two kinds of biosorbents. The control groups (CK) were treatment groups without any filler or biosorbent. Phase 1: In each glass chamber, 12 L water and 6 L biogas slurry were added in the proportion of water: biogas slurry = 2:1. During the experiment, running water was used to supplement for water evaporated naturally in each glass chamber every 3 d to maintain the constant volume of water sample in each glass chamber. After treatment for 15 d, the aboveground part of water spinach was harvested along the second node above the water surface. The concentrations of COD, NH+4-N, TN and TP in the water were measured on the 1st, 3rd, 5th, 7th, 10th and 15th d. The concentrations of Cu, Zn and As in the water were measured on the 1st and 15th d, and the concentrations of Cu, Zn and As in the aboveground part of plants were measured on the 15th d. Phase 2: On the 15th d in Phase 1, after harvesting the aboveground part of water spinach, 12 L biogas slurry was added (Table 3), and the rest treatments remained unchanged. The experiment lasted 15 d, during which running water was used to supplement for water evaporated naturally in each glass chamber every 5 d to maintain the constant volume of water sample in each glass chamber. The concentrations of COD, NH+4-N, TN and TP in the water were measured on the 1st, 5th, 10th and 15th d. The concentrations of Cu, Zn and As in the water were measured on the 1st

Table 3 Experimental design. Treatment CK Z B ZM ZC BM BC ZBMC

Zeolite (g)

Biochar (g)

Compound microbial agent (g)

Chlorella (g)

100 100 100 100

100

1 3 100 100 100

1 1

3 3

Note: Z, zeolite; B, biochar; ZM, zeolite + microorganism; ZC, zeolite + chlorella; BM, biochar + microorganism; BC, biochar + chlorella; ZBMC, zeolite + biochar + microorganism + chlorella. The same below.

and 15th d, and As concentration in the aboveground and underground parts of plants were measured on the 15th d. 2.5. Sample analysis and definition The concentrations of COD, NH+4-N, TN and TP were determined according to references (APHA, 1995). The concentrations of Cu and Zn in biogas slurry and plant samples were determined using atomic absorption spectrometry, and As concentration was determined by double-channel atomic fluorescence spectrometry. The removal efficiency of pollutants was calculated based on Formula 1

R ¼ ð1  Ci =C0 Þ R indicates removal efficiency, C0 and Ci are the initial and final concentrations of pollutants, respectively. 2.6. Statistical analysis Data were statistically analyzed using SPSS 20.0. One-way ANOVA was used to analyze the significant variance of the statistical results of each measured item. The significant differences between different treatments were analyzed by Duncan test. 3. Results and analysis 3.1. Effect of nutrient removal 3.1.1. Effect of COD removal As shown in Fig. 1, on the 1st d during Phase 1, COD concentration in each treatment group was lower than that in CK, and COD concentration under BM, BC and ZBMC was significantly lower than that under other treatments (P < 0.05). After 15 d, COD concentration under CK was still the highest, being 747.71 ± 44.67 mg L1 (mean ± SE), while that under ZBMC was the lowest, 603.92 ± 23.53 mg L1 (Fig. 1a). However, no significant differences were found in COD concentration between different treatments (P > 0.05). In Phase 2, biogas slurry was added on the basis of Phase 1. Therefore, on the 1st d of Phase 2, the nutrient concentrations under different treatments were different. In the first 10 d of Phase 1 and Phase 2, COD concentration decreased rapidly before reaching a slow decrease in the next 5 d. In accordance with Phase 1, COD concentration under BM, BC and ZBMC were significantly lower than that under other treatments on the 1st d (P < 0.05). After 15 d, COD concentration was the highest and lowest under CK and BC, being 115.38 ± 12.82 mg L1 and 25.64 ± 0.00 mg L1 (Fig. 1b), respectively. As seen in Table 4, in the two-phase experiment, the removal rates of COD were treatment groups > CK group, Z > B, ZM and ZC > Z, BM and BC > B, respectively. However, the removal rate of COD showed no signif-

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

COD concentration (mg/L)

5000

Phase 1

CK Z B ZM ZC BM BC ZBMC

4000 3000 2000 1000

a

0 1

3

5

7

10

3500

COD concentration (mg/L)

4

Phase 2

CK Z B ZM ZC BM BC ZBMC

3000 2500 2000 1500 1000 500 0

b 1

15

5

10

15

Experimental Time (days)

Experimental Time (days)

Fig 1. COD concentrations under different combinations of fillers and biosorbents in CWs over time at the first and second phases: (a) Phase 1 and (b) Phase 2.

icant differences between different treatments (P > 0.05). The removal rate of COD in Phase 2 was always higher than that in Phase 1. The highest removal rate of COD during Phase 1 and Phase 2 were 84.11% ± 2.49 (mean ± SE) and 98.77% ± 6.23, respectively (Table 4). These results suggested that the removal effect of the CWs of water spinach on COD was satisfactory, and the purification effect of each treatment presented no great differences. Zeolite as an additive had a slightly better removal effect on COD than biochar. Compared with zeolite and biochar alone, the removal efficiency of the two biosorbents combined with zeolite and biochar on COD was improved to some extent, but such improvement was insignificant. The causes may be as follows. The existing forms of organic compounds in sewage can be divided into solubility and insolubility. CWs can intercept insoluble organic pollutants by mechanical retention of fillers such as filtration, precipitation and adsorption (Haarstad et al., 2012), and absorb dissolved organic pollutants directly through cell walls or membranes of bacteria, algae and plants (Pilon-Smits, 2005). In this experiment, the concentration of organic pollutants was high, and the provided carbon source could be used as the energy source of organisms, resulting in massive growth of algae (Bi et al., 2019). Therefore, the removal efficiency COD was high in the early phase. CWs mainly rely on microbial metabolic activities to decompose and remove organics. As a place for microbial metabolism, fillers’ direct adsorption and other effects showed small impact on organic removal, mainly indirectly affecting microorganisms, thus changing the removal effect of organics (Ladu et al., 2014). Bacterial degradation and the respiration of algae need oxygen. With the continuous con-

sumption of oxygen in biogas slurry, the oxygen content was more and more low, which had a certain inhibitory effect on algae and microorganisms. 3.1.2. Effect of NH+4-N and TN removal As shown in Fig. 2a and 2c, during the first 15 d of Phase 1, the concentrations of NH+4-N and TN in the treatment group were continuously lower than those in the CK group. In the first 5 d, NH+4-N concentration decreased rapidly, which slowed down in the later 5–10 d except for ZM and ZC. No significant differences were found in TN concentration under CK, Z and B (P > 0.05), but significant differences were detected between CK and other treatments (P < 0.05). On the 15th d, the concentrations of NH+4-N and TN under ZBMC reached the lowest level. As seen in Fig. 2b and 2d, on the 1st d of Phase 2, NH+4-N concentration showed no significant differences under different treatments (P > 0.05). The highest TN concentration was found under CK in 1–15 d. On the 15th d, NH+4-N concentration showed significant differences under CK, Z and other treatments except for B (P < 0.05). NH+4-N concentration was the highest under CK and the lowest under ZC and ZBMC. The concentration of NH+4-N did not reach the detection limit. TN concentration was the lowest under ZBMC, but no significant differences were found in TN concentration under different treatments (P > 0.05). Table 4 presented that the removal rates of NH+4-N and TN in the treatment group were significantly higher than those in the CK group during the two phases (P < 0.05). The removal rates of NH+4-N in Phase 1 were Z > B, ZM > ZC > Z, and BM > BC > B, and

Table 4 Removal rate of nutrients using different fillers and their combinations with biosorbents. Treatment

COD removal rates (%)

NH+4-N removal rates (%)

TN removal rates (%)

TP removal rates (%)

Phase 1 CK Z B ZM ZC BM BC ZBMC

80.32 ± 5.16b 82.25 ± 1.37b 80.81 ± 5.04b 82.46 ± 4.52b 82.46 ± 4.12b 83.08 ± 7.68b 82.56 ± 3.52b 84.11 ± 2.49b

88.01 ± 6.77b 95.39 ± 3.45a 93.24 ± 5.71a 99.29 ± 1.21a 97.11 ± 4.34a 98.07 ± 3.48a 97.38 ± 7.96a 99.55 ± 1.52a

79.69 ± 4.56c 88.25 ± 4.98ab 84.95 ± 3.17b 91.49 ± 5.76ab 92.30 ± 3.78a 96.27 ± 4.43a 94.94 ± 2.19a 98.14 ± 5.67a

71.60 ± 5.78c 72.33 ± 4.80c 82.16 ± 4.76b 82.16 ± 2.25b 80.95 ± 3.98b 84.22 ± 2.79ab 92.11 ± 1.45a 88.47 ± 1.76a

Phase 2 CK Z B ZM ZC BM BC ZBMC

96.13 ± 5.64a 95.19 ± 4.35a 95.10 ± 5.23a 97.67 ± 2.35a 96.99 ± 1.57a 98.59 ± 7.34a 98.77 ± 6.23a 96.34 ± 3.25a

69.67 ± 2.04d 72.22 ± 4.39d 81.66 ± 5.84c 96.32 ± 5.09a 100.00 ± 2.34a 96.26 ± 6.34a 97.86 ± 1.25a 100.00 ± 4.24a

62.27 ± 5.36e 64.31 ± 3.6de 66.4 ± 4.69d 71.13 ± 8.23d 77.95 ± 4.36c 73.84 ± 2.96 cd 78.87 ± 5.64c 80.02 ± 2.36bc

67.51 ± 1.57d 64.13 ± 1.98d 65.54 ± 2.17d 68.23 ± 4.09 cd 82.82 ± 4.55b 69.75 ± 3.79c 82.03 ± 3.22b 81.19 ± 4.36b

Note: values in Table 4 were means ± SD. Different lowercase letters in the same line indicate significant difference (p < 0.05).

5

140

Phase 1

CK Z B ZM ZC BM BC ZBMC

120 100 80 60 40 20 0

a 1

3

5

7

10

NH4+-N concentration (mg/L)

NH4+-N concentration (mg/L)

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

70

40 30 20 10 0

15

1

3

5

7

10

15

Experimental Time (days)

100

TN concentration (mg/L)

TN concentration (mg/L)

c 1

b 5

10

15

Experimental Time (days) CK Z B ZM ZC BM BC ZBMC

Phase 1

CK Z B ZM ZC BM BC ZBMC

50

Experimental Time (days) 160 140 120 100 80 60 40 20 0

Phase 2

60

Phase 2

CK Z B ZM ZC BM BC ZBMC

80 60 40 20

d

0 1

5

10

15

Experimental Time (days)

Fig. 2. NH+4-N and TN concentrations under different combinations of fillers and biosorbents in CWs over time at the first and second phases: (a) Phase 1 and (b) Phase 2.

those of TN were Z > B, ZC > ZM > Z, BM > BC > B. In Phase 2, the removal rates of NH+4-N and TN were basically the same (B > Z, ZC > ZM > Z, BC > BM > B). It indicated that the removal efficiency of NH+4-N and TN in CWs applied with zeolite was higher than that with biochar in Phase 1, but the result was opposite in Phase 2. The cause may be that zeolite has high CaCO3 content and strong selective adsorption of NH+4-N, contributing to rapid removal of NH+4-N from the water (Xu et al., 2019). Some studies have demonstrated that the adsorption of NH+4-N by biochar can be divided into two steps: rapid surface adsorption and speed-controlled intramolecular diffusion (Tang et al., 2013). In Phase 1 of this experiment, the removal efficiency of NH+4-N was B > Z in the first 7 d, especially on the 1st d. Then, their difference was gradually narrowed until Z > B. However, in Phase 2, the removal efficiency of NH+4-N was B > Z. These suggested that on the 1st d, the adsorption rate on biochar surface was higher than that of zeolite, but with the intramolecular diffusion of biochar, the adsorption rate decreased gradually and then became lower than that of zeolite. Moreover, biochar is alkaline, which leads to the gradual rightward movement of the equilibrium NH+4+OH– NH3H2O in solution, and reduced adsorption effect of biochar on NH+4. Therefore, the removal effect in Phase 1 was Z > B, which was consistent with the results obtained by Hina et al. (2015). However, some studies have shown that the amount of NH+4 adsorbed by biochar increases with its concentration (Gai et al., 2014). In Phase 2 of this experiment, biogas slurry was supplemented, which increased the initial concentration of NH+4 and promoted the adsorption of NH+4 by biochar. The study of Li et al. (2016) also concluded that increasing the initial concentration of NH+4-N could improve its removal rate by biochar. One possibility is that, with increasing initial concentration of NH+4-N in solution, more NH3, NH3H2O, and NH+4 surrounded the surface of biochar while most adsorption sites reached adsorption saturation. NH+4 in solution formed a large mass concentration difference with the surface of biochar, which increased the migration power of NH+4 to the internal parts of biochar adsorbent, and made the adsorption reaction more thorough (Ahmad et al. 2013). The final

adsorption amount of biochar was larger than that of zeolite. Therefore, the removal effect in Phase 2 was B > Z. In Phase 1, the removal effect of NH+4-N and TN by the combinations of the two biosorbents with zeolite and biochar was superior to that by zeolite and biochar alone, suggesting that the combinations of the two biosorbents and fillers play a synergistic role in the removal of nitrogen. For NH+4-N, the strengthening effect of compound microbial agent was better than that of chlorella. However, in Phase 2, the strengthening effect of the combined use of chlorella and the two fillers in NH+4-N and TN removal was improved compared to that of combined microbial agent. NH+4-N was mainly removed by plant absorption, substrate adsorption, volatilization, nitrification and denitrification. Nitrogen in biogas slurry of pig farms mainly existed in the form of NH+4-N. TN and NH+4 in the CWs decreased with time and at a declining rate in the early phase. This may have been caused by the fact that the removal of nitrogen by the constructed wetland occurred through the rapid adsorption and interception of suspended nitrogen by zeolite and biochar, followed, in decreasing order, by the effects of plants, chlorella, and microorganisms on dissolved nitrogen. In addition, in the twostage trial in our study, the initial COD concentration was high (Table 1). Because of dissolved nitrogen by the slow decomposition of suspended organic pollutants adsorbed and intercepted by fillers with the action of microorganisms, the subsequent removal effect was reduced. The application of compound microbial agent enhanced the removal of nitrogen by CWs in this experiment, especially in Phase 1. Moreover, zeolite and biochar with high porosity and huge specific surface area can provide a storage place for NH+4N while facilitating the adhesion of microorganisms. The biofilm formed after stable operation of the system is also helpful for microorganisms to remove nitrogen by nitrification and denitrification. Therefore, adsorption combined with nitrification and denitrification of NH+4-N by fillers increased the denitrification efficiency of CWs, which is consistent with previous results (Yalcuk and Ugurlu, 2009; LUEDERITZ et al., 2001). Further, chlorella can remove nitrogenous and phosphoric nutritious salts effi-

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

ciently (Wang and Ren, 2014). NH+4-N can be removed from the culture medium in two ways. One is that the growth of algae mainly consumes NH+4-N which is easily absorbed to synthetic organic matters for its own cells. The other is that, during photosynthesis of microalgae, the pH value of solution increases, leading to the volatilization of NH+4-N (Vogel et al., 2010). This is the reason why the combination of zeolite and the two biosorbents as well as the combination of biochar and the two biosorbents present better effect on NH+4-N removal than the two fillers alone. The removal rates of NH+4-N and TN were the highest with the combined application of the two biosorbents and two fillers in both the first and second phases. One cause is that the combination of biosorbents and fillers could improve the removal rate of nitrogen’, and ‘On the other hand, the simultaneous application of chlorella and microbial agent also led to the formation of a microalgalbacterial symbiosis system. This improved the removal rate of nutrients in water through algal photosynthesis and microbial metabolism and respiration (Liu et al., 2017). 3.1.3. Effect of TP removal Fig. 3a presented that TP concentration under CK was the highest in the first 15 d of Phase 1, and that under Z was slightly lower than that under CK, showing no significant difference (P > 0.05). Further analysis showed that TP concentration under ZBMC was the lowest in 1–3 d, and that under BC was the lowest in 5–15 d, with the highest final removal rate of 92.11% (Table 4). Therefore, the removal effect of TP by biochar combined with chlorella in the CWs was the optimal. As seen in Fig. 3b, in 1–10 d of Phase 2, TP concentration was the lowest under BM. TP concentration was the highest under Z in 5–15 d, with the lowest removal rate of 64.13% ± 1.98 (Table 4), indicating that zeolite used as adsorbent alone had a short-term effect. On the 15th d, TP concentration was the lowest under BC, which was 1.81 ± 0.69 mg L1 (Fig. 3b). No significant differences were found in TP concentration between different treatments (P > 0.05). As shown in Table 4, in Phase 1, the removal rate of TP in the treatment group was higher than that in the CK group. Except for Z, significant differences were detected in the removal rate of TP between other treatments and CK (P < 0.05). The removal rates were B > Z, ZM > ZC > Z and BC > BM > B. In Phase 2, except for Z and B, the removal rate of TP under other treatments was higher than that under CK. The removal rates were B > Z, ZC > ZM > Z and BC > BM > B. The results demonstrated that zeolite as adsorbent alone had no obvious enhancement effect on TP removal in CWs, and the removal effect of biochar was slightly better than that of zeolite. A potential cause may be that zeolite as a substrate in CWs has a weak ability to adsorb and remove total phosphorus, and it is mainly used to treat nitrogen-containing wastewater (Miller et al., 2011; Xu et al., 2019).

TP concentration (mg/L)

8 7

Phase 1

CK Z B ZM ZC BM BC ZBMC

6 5 4 3 2 1

a

0 1

3

5

7

10

Experimental Time (days)

15

In CWs, main approaches for total phosphorus removal include the uptake of algae and plant, transformation and absorption of microorganisms, absorption and sedimentation of fillers. The amount of phosphorus removal by microorganisms and plants is relatively small (Seo et al., 2005). Reddy et al. (1998) have found that 7%–87% phosphorus in the wetland system is removed by adsorption or precipitation. The removal mechanism mainly works by the adsorption and precipitation reaction between soluble phosphate and A13+, Fe3+ and Ca2+ in the substrate. A13+ and Fe3+ in the substrate mainly react with phosphorus at pH < 7, while Ca2+ is prone to react with phosphorus at pH > 7. Moreover, compared with Mg, Al, Fe and Ca, the adsorption of Ca and P has the highest correlation (Zhu et al., 1997). It is consistent with the results of this experiment which proved that the removal of TP by alkaline biochar was more effective. In this experiment, the two biosorbents combined with zeolite and biochar had certain advantages in TP purification compared with zeolite and biochar alone, and finally chlorella as a biosorbent showed the best enhancement effect. Usually, there are two main ways for microalgae to remove phosphorus: one is that, with the growth of microalgae, phosphorus elements in the culture medium is assimilated and nutrients of microalgae themselves is synthesized (Martinez et al., 1999). The other is that the photosynthesis of algae increases the pH value of the solution, and phosphorus precipitates in the form of insoluble salts (Vogel et al., 2010). Whether in Phase 1 or Phase 2, the combination of biochar and chlorella could maintain a good removal effect of TP in the wetland system of water spinach. Nevertheless, the reproduction of chlorella has strict requirements for light and temperature. Therefore, in practical application, a suitable environment should be available for chlorella to reproduce normally. 3.2. Effect of heavy metal removal In Phase 1, it can be seen from Fig. 4a that on the 1st and 15th d, Cu concentration was the highest under ZM and the lowest under ZBMC. After treatment for 15 d, no significant differences were found in Cu concentration between ZBMC and CK, Z and BC (P > 0.05), but the findings were contrary between ZBMC and other treatments (P < 0.05). The order of the removal rates of Cu under each treatment was ZBMC > BC > Z > CK > ZC > BM > B > ZM. In Phase 2, on the 1st d, Cu concentration was the highest under CK, which was 2.4163 ± 0.0012 mg L1 (Fig. 4b). On the 15th d, the highest and lowest concentrations of Cu were found under B and ZC, being 2.4062 ± 0.0026 mg L1 and 2.3905 ± 0.0138 mg L1 (Fig. 4b), respectively. Cu concentration showed no significant differences between different treatments (P > 0.05). The order of the removal rate of Cu under each treatment was CK > ZC > ZM > ZBMC > BC > BM > Z > B (Table 5). In both phases, the removal rate

16

TP concentration (mg/L)

6

14

Phase 2

CK Z B ZM ZC BM BC ZBMC

12 10 8 6 4 2

b

0 1

5

10

15

Experimental Time (days)

Fig. 3. TP concentrations under different combinations of fillers and biosorbents in CWs over time at the first and second phases: (a) Phase 1 and (b) Phase 2.

7

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

Phase 1

CK Z B ZM ZC BM BC ZBMC

2.42 2.41 2.40

a

2.39 1day

2.43

Cu concentration mg/L

Cu concentration mg/L

2.43

Phase 2

2.42 2.41 2.40

b

2.39

15days

1day

Experimental treatment Phase 1

CK Z B ZM ZC BM BC ZBMC

0.22

0.21

c

0.20 1day

0.30

Phase 2

0.26 0.24 0.22

d

0.20

15days

1day

Phase 1

CK Z B ZM ZC BM BC ZMBC

30 25 20 15 10 5

e

0 1day

15days

Experimental treatment 40

As concentration mg/L

As concentration mg/L

35

CK Z B ZM ZC BM BC ZBMC

0.28

Experimental treatment 40

15days

Experimental treatment

Zn concentration mg/L

Zn concentration mg/L

0.23

CK Z B ZM ZC BM BC ZBMC

35

Phase 2

CK Z B ZM ZC BM BC ZBMC

30 25 20 15 10 5

f

0 1day

15days

15days

Experimental treatment

Experimental treatment

Fig. 4. Cu, Zn and As concentrations under different combinations of fillers and biosorbents in CWs over time at the first and second phases: (a) Phase 1 and (b) Phase 2.

of Cu under zeolite treatment was slightly higher than that under biochar treatment, which is consistent with the results obtained by Peng et al. (2018). The removal rate of Cu in water under each treatment was<1%, suggesting poor removal effect. As can be learnt from Fig. 4c, after culture for 15 d in Phase 1, the lowest and highest concentrations of Zn were found under ZBMC and BC, respectively. No significant differences were found in Zn concentration between different treatments (P > 0.05). In Phase 1, the removal rate of Zn was between 24.04% ± 1.09 and 27.04% ± 3.71, with slight differences observed (Table 5). In Phase 2, Zn concentration presented no significant differences between different treatments on the 1st d (P > 0.05). After treatment for 15 d, CK and BC resulted in the highest and lowest concentrations of Zn, being 0.2535 ± 0.0120 mg L1 and 0.2133 ± 0.0008 mg L1 (Fig. 4d), respectively. Except for ZC and BC, Zn concentration showed no significant differences between CK and other treatments (P > 0.05). The removal rates of Zn under each treatment ranged from 8.15% ± 3.66 to 23.69% ± 4.02 (Table 5). The order of the removal rates was BC > ZC > ZBMC > B > Z > ZM > BM > CK (Table 5). In Phase 2, the removal rate of Zn under ZC, BC and ZBMC was significantly higher than that under other treatments, which indicated that chlorella had a good enhancement effect on the

Table 5 Removal rate of Cu, Zn and As under different combinations of fillers and biosorbents. Treatment

Cu removal rates (%)

Zn removal rates (%)

As removal rates (%)

Phase 1 CK Z B ZM ZC BM BC ZBMC

0.55 ± 0.012d 0.58 ± 0.021 cd 0.51 ± 0.017d 0.49 ± 0.034de 0.53 ± 0.031d 0.53 ± 0.041d 0.60 ± 0.011c 0.75 ± 0.049b

25.54 ± 2.76a 26.50 ± 1.95a 24.18 ± 1.38a 26.54 ± 2.94a 24.57 ± 1.09a 25.25 ± 3.91a 24.04 ± 1.09a 27.04 ± 3.71a

76.64 ± 4.56b 74.61 ± 6.22b 81.30 ± 5.23a 77.09 ± 4.33b 81.91 ± 3.11a 83.89 ± 5.22a 80.99 ± 6.19a 82.72 ± 3.28a

Phase 2 CK Z B ZM ZC BM BC ZBMC

0.88 ± 0.029a 0.37 ± 0.016f 0.32 ± 0.042f 0.73 ± 0.052b 0.87 ± 0.022a 0.45 ± 0.011e 0.46 ± 0.032e 0.64 ± 0.029c

8.15 ± 3.66b 9.83 ± 2.87b 10.37 ± 1.24b 8.71 ± 2.09b 23.49 ± 2.11a 8.27 ± 1.01b 23.69 ± 4.02a 22.38 ± 3.11a

35.58 ± 2.22e 37.18 ± 3.21e 40.73 ± 3.12e 35.63 ± 3.56e 53.85 ± 2.89d 68.94 ± 3.01c 35.38 ± 2.12e 74.09 ± 4.90b

Note: Values in Table 5 were means ± SD. Different lowercase letters in the same line indicate significant difference (p < 0.05).

8

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

removal of Zn from water in the CWs of water spinach. Zn is a necessary trace element for the growth and metabolism of chlorella. The growth of chlorella can be promoted by Zn at appropriate concentration. Research has found that the adsorption capacity of immobilized chlorella to Zn2+ is 33.20 mg/L, and that of suspended chlorella to Zn2+ is 28.50 mg/L (Maznah et al., 2012). As shown in Fig. 4e, on the 1st and 15th day of Phase 1, the highest and lowest As concentrations were detected under Z and BM, respectively. There were no significant differences in As concentration among CK, Z and B (P > 0.05), but there were significant differences between CK, Z and B and other treatments (P < 0.05). On the 15th d of the experiment, the highest and lowest concentrations of As were 9.11 ± 0.58 lgL1 and 5.78 ± 0.48 lgL1 (Fig. 4e), respectively. As concentration exhibited significant differences between Z and BM and ZBMC (P < 0.05), but no significant differences were found from other treatments (P > 0.05). The removal rate of As in each treatment ranged from 74.61% ± 6.22 to 83.89% ± 5.22 (Table 5), and its order was BM > ZBMC > ZC > B > BC > ZM > CK > Z. Hence, the addition of microbial agent and chlorella to the CWs of water spinach could purify As quickly. The removal effect of the combination of zeolite and the two biosorbents on As was better than that of zeolite alone. The removal effect of the combination of microbial agent and biochar on As was also higher than that of biochar alone. On the 1st day of Phase 2, the highest and lowest As concentrations were found under Z and BM, being 37.95 ± 1.41 lg L1 and 17.87 ± 0.46 lg L1, respectively. On the 15th d, the highest and lowest concentrations of As were detected under Z and ZBMC, 23.84 ± 2.76 lg L1 and 4.83 ± 0.41 lg L1 (Fig. 4f), respectively. As concentration showed significant differences between different treatments. As shown in Table 5, the removal rate of As in water under each treatment ranged from 35.58% ± 2.22 to 74.09% ± 4.90, and its order was ZBMC > BM > ZC > B > Z > ZM > CK > BC. ZBMC, BM and ZC presented the best removal effect on As in the two phases. In summary, the removal effect on heavy metals by the CWs of water spinach was As > Zn > Cu, and the removal effect on As in Phase 1 was obviously better than that in Phase 2. Main ways of removing heavy metal ions in CWs include adsorption and precipitation of substrate or soil, and absorption and enrichment of plants (Mays and Edwards, 2001). The effect of removing heavy metals by plant harvesting is not obvious, and the substrate is the main place for heavy metal accumulation (Lesage, et al., 2007; Mustapha et al., 2018). Therefore, sedimentation, adsorption and filtration are considered the most important ways for metal removal (Yadav et al., 2010). Causes for the low removal efficiency of Cu and Zn under each treatment in this study may be as follows. Cu and Zn concentrations in biogas slurry are not high, and the removal effect of each treatment on Cu and Zn is not obvious. The CW system has a high water level, and after planting water spinach, the substrate at the bottom of the wetland pond is in a hypoxic environment for a long time, with low dissolved oxygen and high organic content, resulting in a low redox potential for the substrate. Moreover, water may contain many organic ligands that are easy to form stable complexes with iron and manganese, which will release iron and manganese from the solid phase, leading to the release of various combined trace heavy metals into the water phase to cause secondary pollution (Wang et al., 2015). In CWs, Cu and Zn mainly exist in the exchangeable state with high leachability and bioavailability, and can lead to the release of Cu and Zn (Zhou et al., 2019). In addition, competitive adsorption exists between Cu and Zn. Cu is the most preferred heavy metal to be retained and has a great ability to inhibit the adsorption of other heavy metals (Sheikhhosseini et al., 2013). Galletti et al. (2010) found that the removal rates of Cu and Zn in horizontal subsurface CWs for the treatment of municipal domestic sewage were also low, being 3%–9% and 25% respectively,

which is consistent with the removal effects of Cu and Zn in our study. Research has shown that wetland system has a good removal effect on As. Islam et al. (2013) have found that the removal rate of As in the constructed wetland of Micranthemum brosum ranged from 79.3% to 89.5%. In Phase 1, the removal rates of As under different treatments were not significantly different, but the removal rates by adding biochar were relatively high. Previous studies have revealed that biochar has a certain effect on the removal of As from wastewater (Yadav et al., 2010). Mondal et al. (2008) have showed that the relative removal rates of As by unmodified biochar are 55.5% (total As), 44% (III) and 67% (V). In Phase 2, BM and ZBMC presented better removal effect on As, indicating that the combination of biochar and compound microbial agent could enhance the purification effect of CWs on As. Microorganisms are involved in many aspects of As cycling in the environment. Various polar functional groups (e.g. –COOH, –NH2, –PO3 4 , -SH on the cell wall) on the surface of microorganisms can react quantitatively with As ions (e.g. ion exchange, cooperative binding or complexation) to immobilize As (Unz and Shuttleworth, 1996). Takeuchi et al. (2007) have found that Marinomonas communis growing in the culture medium containing 5 mg/L As (Ⅴ) can absorb 2290 mg/kg As (dry weight). Microorganisms do have a certain adsorption effect on As in biogas slurry. Additionally, biochar can be used as a carrier of microbial agents. Therefore, their combination can improve the removal ability of As in the CWs. 3.3. Absorption of heavy metals by water spinach In Phase 1, the content of heavy metals in the aboveground part of water spinach was: Cu > Zn > As (Fig. 5), which was consistent with the concentration of heavy metals in biogas slurry: Cu > Zn > As (Fig. 4). The effects of different treatments on the contents of Cu and Zn in the aboveground part of water spinach were basically the same. The highest and lowest contents of Cu and Zn in the aboveground part of water spinach were found under ZC and ZBMC, respectively. The corresponding contents of Cu were 58.28 ± 0.53 mgkg1 and 46.55 ± 0.19 mgkg1, respectively. The corresponding contents of Zn were 6.32 ± 0.08 mgkg1 and 4.97 ± 0.02 mgkg1, respectively (Fig. 5). Except for ZC, the contents of Cu and Zn in each treatment group were lower than those in CK. Cu content showed no significant differences between CK and B, ZC and BC (P > 0.05), but significant differences were observed under other treatments (P < 0.05). For Zn, its content in aboveground water spinach of each treatment group was lower than that in CK, showing no significant differences (P > 0.05). The content of Cu was CK > B > Z, and that of Zn was CK > Z > B, suggesting that zeolite was better than biochar in reducing Cu accumulation in water spinach. For Zn, the result was opposite, which was consistent with the effect of the two fillers alone on the removal rate of Cu and Zn in biogas slurry mentioned above. It indicates that fillers alone can reduce the enrichment of Cu and Zn by wetland plants through adsorption and filtration. There were no significant differences in the contents of Cu and Zn between Z and ZM (P > 0.05), but significant differences between Z and ZC (P < 0.05) were detected. The contents of Cu and Zn were ZC > Z > ZM. No significant differences were found in the contents of Cu and Zn among B, BM and BC (P > 0.05). The contents of Cu and Zn were BC > B > BM. The lowest Cu and Zn contents were found under ZBMC. The results demonstrated that the cumulative amount of Cu and Zn in the aboveground part of water spinach using microbial agent combined with zeolite and biochar was slightly lower than that applied with zeolite and biochar alone. The cumulative amount of Cu and Zn in the aboveground part of water spinach using chlorella combined with zeolite and biochar was slightly higher than that applied with zeolite and biochar

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

8

Zn concentration (mg/kg)

80

Cu concentration (mg/kg)

9

60

40

20

0

6

4

2

0 CK

Z

B

ZM ZC BM BC ZBMC

CK

Experimental treatment

Z

B

ZM ZC BM BC ZMBC

Experimental treatment

As concentration (mg/kg)

3.0 2.5 2.0 1.5 1.0 0.5 0.0 CK

Z

B

ZM ZC BM BC ZMBC

Experimental treatment Fig. 5. Cu, Zn and As contents in the aboveground part of water spinach treated with different combinations of fillers and biosorbents in Phase 1.

alone. However, the increased or decreased amount was small, which may be caused by the fact that Cu and Zn can promote the growth of microalgae at low concentrations but inhibit the growth of microalgae at high concentrations. This phenomenon is considered as self-balancing effect (Calabrese, 1999). Cu and Zn at low concentrations in this experiment could also promote the proliferation of chlorella to a certain extent. Additionally, the root system of water spinach has a huge surface area, which provides a good environment for increasing microbial attachment, secreting various organic compounds as carbon source of chlorella and microorganisms. With combined application of chlorella and fillers, due to its own gravity, ion adsorption with sediment matrix at the bottom of the pond, and utilization of root exudates of water spinach for self-proliferation, chlorella and its adsorbed heavy metals accumulated massively in the rhizosphere at the bottom of the CWs, resulting in partial uneven distribution of heavy metals in the CWs. Compared with single application of fillers, the enrichment in the aboveground part of water spinach could be improved to a certain extent. In this experiment, the microbial agent contained various microorganisms. Heavy metals in wastewater could be removed by biosorption, enrichment and flocculation. Microorganisms could be used as carriers, and their combined application with microbial agent could play a synergistic role in the removal of Cu and Zn in wastewater. In addition, ZBMC caused the lowest accumulation of Cu and Zn in the aboveground part of water spinach, suggesting that, like nitrogen (3.1.2), the two biosorbents were applied simultaneously to form algal-bacterial synergistic coexistence system, which could enhance the removal of heavy metals from wastewater (Loutseti et al., 2009). The highest and lowest As contents were found under BM and B, being 2.75 ± 0.14 mgkg1 and 0.94 ± 0.15 mgkg1, respectively (Fig. 5). As accumulation in the aboveground part of water spinach under BM and BC was higher than that under CK but As accumula-

tion under other treatments was lower than that under CK. Significant differences were not found in As content between CK and Z (P > 0.05), but they were observed between CK and B (P < 0.05). The content of As was CK > Z > B, suggesting that zeolite and biochar could reduce As accumulation in water spinach, especially biochar. As content showed no significant differences among Z, ZM and ZC (P > 0.05), and the As content was ZC > ZM > Z. Significant differences were detected in As content among B, BM and BC (P < 0.05), and the content of As was BM > BC > B. Therefore, compared with zeolite and biochar alone, As accumulation in the aboveground part of water spinach increased under the combination of the two biosorbents with zeolite and biochar, which was different from the enrichment of Cu and Zn in water spinach. The cause is that microorganism-mediated heavy metal removal in CWs is diverse and complex (Sheoran and Sheoran, 2006). Moreover, the removal processes and environments for conventional heavy metals such as Cu and Zn are diverse. The two elements can be removed in different environments. Cu and Zn are cations which can precipitate under normal pH conditions, while As is a neutral or anionic ion which can be removed under specific conditions (with matters tolerant to Fe, S and Ca, etc.) (Lizama et al., 2011). It has been considered that the main removal pathway of As is bioprecipitation of arsenic sulphide formed under sulfatereducing bacteria (Duncan et al., 2004). Lizama et al. (2011) have found that Fe can co-precipitate with As, and As and Fe can coprecipitate with organic sulfides in anaerobic environment. In this study, the oxygen content in CWs decreased in the later phase (see Section 3.1.1), which was beneficial to the formation of As to insoluble sulfide compounds as well as the accumulation of chlorella at the bottom of CWs, increasing As concentration in the rhizosphere and promoting the enrichment of As in water spinach. When biochar and zeolite were applied alone, no such sedimentation would occur, which reduced the concentration of heavy metals in biogas

10

X. Guo et al. / Science of the Total Environment 703 (2020) 134788

As concentration mg/L

12

CK Z B ZM ZC BM BC ZBMC

9 6 3 0

Roots

Shoots

Fig. 6. As content in aboveground and underground parts treated with different combinations of fillers and biosorbents in Phase 2.

slurry and water spinach. B presented the best effect on reducing As accumulation in water spinach, followed by ZBMC. Therefore, in this experiment, both fillers and biosorbents in wetlands could reduce the concentration of heavy metals in water to a certain extent, but different combinations could also affect the distribution of total heavy metals in different space of wetlands, thereby affecting the absorption of heavy metals by water spinach. In Phase 2, the accumulative As content of the aboveground and underground parts of water spinach (Fig. 6) was as follows: underground part > aboveground part. To be consistent with Phase 1, the highest and lowest As content in the aboveground part of water spinach in Phase 2 were found under BM and B, being 2.05 ± 0.27 mgkg1 and 0.62 ± 0.04 mgkg1, respectively (Fig. 6). As content presented no significant differences between CK and Z and B (P > 0.05), but significant differences were displayed between CK and other treatments (P < 0.05). ZBMC and B caused the highest and lowest As contents in the underground part, being 9.77 ± 2.26 mgkg1 and 5.98 ± 0.43 mgkg1, respectively (Fig. 6). There were no significant differences in As content under different treatments (P > 0.05). The results demonstrated that As enrichment in the roots of water spinach could be increased by the combination of the two fillers and the two biosorbents in CWs, the cause of which is consistent with Phase 1. Only B (biochar alone) could reduce As accumulation in water spinach. As content in water spinach was the lowest in both aboveground and underground parts under this treatment. To sum up, adding zeolite and biochar as adsorbents to the CWs of water spinach could reduce the accumulation of heavy metals such as Cu, Zn and As in water spinach. The cumulative Cu content of all treated water spinach met the standards of formulated feed for piglets (Cu  200 mgkg1) and the early formulated feed for growing-finishing pigs (20 kg ~ 60 kg) (Cu  150 mgkg1) in the Allowable Amount of Cu in Feeds (GB 26419–2010). The accumulative Zn content of all treated water spinach conformed to the Allowable Amount of Zn in Feeds (NY 929–2005). The accumulative As content of all treated water spinach met the Hygienical Standard for Feeds (GB 13078–2017) (As  10 mg.kg1). Therefore, in the CWs of water spinach, the wetland plant water spinach can be used as green fodder to return to agricultural production.

4. Conclusion In this study, the strengthening effects of different combinations of fillers and biosorbents on the removal of pollutants from biogas slurry were investigated, which is valuable for the largescale application of wetlands in sewage purification. However, previous studies have not fully explored this issue. In this experiment, fillers and their different combinations with biosorbents could

enhance the purifying ability of CWs of water spinach to nutrients and heavy metals in biogas slurry to varying degrees. The concentration of nutrients in biogas slurry of CWs decreased gradually with the treatment time, and the removal rate of nutrients under each treatment was >60%. The removal efficiency of COD under each treatment increased significantly from Phase 1 to Phase 2, but differences in the removal effect were insignificant between the treatments. The combination of zeolite, biochar, compound microbial agent and chlorella showed the optimal removal effect on TN and NH+4-N among all treatments. The combination of biochar and chlorella maintained a good removal effect on TP. The removal effect of heavy metals from biogas slurry by CWs of water spinach was As > Zn > Cu. The removal rate of Cu under each treatment was<1%. The removal rates of Zn and As were 8.15%–23.69% and 35.38%–83.89%, respectively. The combination of biochar and microorganisms presented significant advantages in the removal of As from biogas slurry. The combination of zeolite, biochar, compound microbial agent and chlorella had the best effect on reduction of the accumulation of Cu and Zn in the aboveground part of water spinach under all treatments. Biochar showed the best effect on reduction of As enrichment in the aboveground and underground parts of water spinach. The combination of fillers and biosorbents can be selected according to the actual demand. However, chlorella is not suitable for growing in biogas slurry at a high concentration and has demand for light and temperature. Compound microbial agent contains various microorganisms, and its efficient removal of pollutants also needs appropriate conditions. Therefore, in the further study, we will explore the optimal condition for bioremediation of wastewater and the interaction mechanism between fillers and organisms as well as between different organisms. Conflict of interest Authors declare that there is no conflict of interest to declare. Acknowledgments This study was sponsored by the National Key Research and Development Program of China (2017YFD0800900) and the Doctoral Startup Research Project of China West Normal University (17E052). References Álvarez-Rogel, J., Gómez, M.D.C.T., Conesa, H.M., Párraga-Aguado, I., GonzálezAlcaraz, M.N., 2018. Biochar from sewage sludge and pruning trees reduced porewater Cd, Pb and Zn concentrations in acidic, but not basic, mine soils under hydric conditions. J. Environ. Manage. 223, 554–565. APHA, 1995. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC. Ahmad, M., Rajapaksha, A.U., Lim, J.E., Zhang, M., Ok, Y.S., 2013. Biochar as a sorbent for contaminant management in soil and water: a review. Chemosphere 99 (3), 19–33. Bruch, I., Fritsche, J., Bänninger, D., Alewell, U., Sendelov, M., Hürlimann, H., Hasselbach, R., Alewell, C., 2011. Improving the treatment efficiency of constructed wetlands with zeolite-containing filter sands. Bioresour. Technol. 102 (2), 937–941. Bi, R., Zhou, C., Jia, Y., Wang, S., Li, P., Reichwaldt, E.S., Liu, W., 2019. Giving waterbodies the treatment they need: A critical review of the application of constructed floating wetlands. J. Environ. Manage. 238, 484–498. Calabrese, E.J., 1999. Evidence That hormesis represents an ‘‘Overcompensation” response to a disruption in homeostasis. Ecotoxicol. Environ. Saf. 42, 135–137. Duncan, W.F.A., Mattes, A.G., Gould, W.D., Goodazi, F., 2004. Multi-stage biological treatment system for removal of heavy metal contaminants. In: Rao, S.R., Harrison, F.W., Kosinski, J.A., Amaratunga, L.M., Cheng, T.C., Richards, G.G. (Eds.), The Fifth International Symposium on Waste Processing and Recycling in Mineral and Metallurgical Industries. Canadian Institute of Mining, Metallurgy and Petroleum, Hamilton, Ontario, Canada, pp. 469–483. De Godos, I., Blanco, S., García-Encina, P.A., Becares, E., Muñoz, R., 2009. Long-term operation of high rate algal ponds for the bioremediation of piggery wastewaters at high loading rates. Bioresour. Technol. 100 (19), 4332–4339.

X. Guo et al. / Science of the Total Environment 703 (2020) 134788 Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F., Burr, M.D., Brisson, J., Campera, A.K., Stein, O.R., 2009. Microbial processes influencing performance of treatment wetlands: a review. Ecol. Eng. 35 (6), 987–1004. Galletti, A., Verlicchi, P., Ranieri, E., 2010. Removal and accumulation of Cu, Ni and Zn in horizontal subsurface flow constructed wetlands: contribution of vegetation and filling medium. Sci. Total Environ. 408 (21), 5097–5105. Gai, X., Wang, H., Liu, J., Zhai, L., Liu, S., Ren, T., Liu, H., 2014. Effects of feedstock and pyrolysis temperature on biochar adsorption of ammonium and nitrate. PLoS One 9 (12) e113888. Haarstad, K., Bavor, H.J., Mæhlum, T., 2012. Organic and metallic pollutants in water treatment and natural wetlands: a review. Water Sci. Technol. 65 (1), 76–99. Hina, K., Hedley, M., Camps-Arbestain, M., Hanly, J., 2015. Comparison of Pine Bark, Biochar and Zeolite as Sorbents for NH+4-N Removal from. Water. CLEAN–Soil, Air, Water 43 (1), 86–91. Islam, M.S., Ueno, Y., Sikder, M.T., Kurasaki, M., 2013. Phytofiltration of arsenic and cadmium from the water environment using Micranthemum umbrosum (JF Gmel) SF Blake as a hyperaccumulator. Int. J. Phytorem. 15 (10), 1010–1021. Li, J.H., Lv, G.H., Bai, W.B., Liu, Q., Zhang, Y.C., Song, J.Q., 2016. Modification and use of biochar from wheat straw (Triticum aestivum L.) for nitrate and phosphate removal from water. Desalination Water Treatment 57 (10), 4681–4693. Liu, J.Z., Wu, Y.H., Wu, C.X., Muylaert, K., Vyverman, W., Yu, H.Q., Munoz, R., Rittmann, B., 2017. Advanced nutrient removal from surface water by a consortium of attached microalgae and bacteria: a review. Bioresour. Technol. 241, 1127–1137. Loutseti, S., Danielidis, D.B., Economou-Amilli, A., Katsaros, C., Santas, R., Santas, P., 2009. The application of a micro-algal/bacterial biofilter for the detoxification of copper and cadmium metal wastes. Bioresour. Technol. 100 (7), 2099–2105. Lizama, K., Fletcher, T.D., Sun, G., 2011. Removal processes for arsenic in constructed wetlands. Chemosphere 84 (8), 1032–1043. Liu, R.B., Zhao, Y.Q., Doherty, L., Hu, Y.S., Hao, X.D., 2015. A review of incorporation of constructed wetland with other treatment processes. Chem. Eng. J. 279, 220– 230. Lesage, E., Rousseau, D.P.L., Meers, E., Van de Moortel, A.M.K., Du Laing, G., Tack, F.M. G., De Pauw, N., Verloo, M.G., 2007. Accumulation of metals in the sediment and reed biomass of a combined constructed wetland treating domestic wastewater. Water Air Soil Pollut. 183 (1–4), 253–264. Ladu, J.L.C., Lu, X., Zheng, M., Wei, T., 2014. Application of A2/O Bio-reactor Constructed Wetlands for Removing Organic and Nutrient Concentrations from Rural Domestic Sewage. J. Environ. Sci. 4, 709–718. Luederitz, V., Eckert, E., Lange-Weber, M., Lange, A., Gersberg, R.M., 2001. Nutrient removal efficiency and resource economics of vertical flow and horizontal flow constructed wetlands. Ecol. Eng. 18 (2), 157–171. Maucieri, C., Florio, G., Borin, M., 2018. Ligneous-cellulosic, nitrophilous and wetland plants for biomass production and watertable protection against nutrient leaching. Eur. J. Agron. 96, 77–86. Maznah, W.W., Al-Fawwaz, A.T., Surif, M., 2012. Biosorption of copper and zinc by immobilised and free algal biomass, and the effects of metal biosorption on the growth and cellular structure of Chlorella sp. and Chlamydomonas sp. isolated from rivers in Penang, Malaysia. J. Environ. Sci. 24 (8), 1386–1393. Mondal, P., Majumder, C.B., Mohanty, B., 2008. Effects of adsorbent dose, its particle size and initial arsenic concentration on the removal of arsenic, iron and manganese from simulated ground water by Fe3+ impregnated activated carbon. J. Hazard. Mater. 150 (3), 695–702. Mustapha, H.I., van Bruggen, J.J.A., Lens, P.N.L., 2018. Fate of heavy metals in vertical subsurface flow constructed wetlands treating secondary treated petroleum refinery wastewater in Kaduna, Nigeria. Int. J. Phytoremediat. 20 (1), 44–53. Miller, R.L., Jensen, B.J., Munns, B.T., Cardon, G.E., 2011. Use of steel slag to remove dissolved phosphorus from lagoon supernatant. Trans. ASABE 54 (1), 191–196. Merrikhpour, H., Jalali, M., 2013. Comparative and competitive adsorption of cadmium, copper, nickel, and lead ions by Iranian natural zeolite. Clean Technol. Environ. Policy 15 (2), 303–316.

11

Mays, P.A., Edwards, G.S., 2001. Comparison of heavy metal accumulation in a natural wetland and constructed wetlands receiving acid mine drainage. Ecol. Eng. 16 (4), 487–500. Martinez, M.E., Jimenez, J.M., El Yousfi, F., 1999. Influence of phosphorus concentration and temperature on growth and phosphorus uptake by the microalga Scenedesmus obliquus. Bioresour. Technol. 67 (3), 233–240. Neeraj, G., Krishnan, S., Kumar, P.S., Shriaishvarya, K.R., Kumar, V.V., 2016. Performance study on sequestration of copper ions from contaminated water using newly synthesized high effective chitosan coated magnetic nanoparticles. J. Mol. Liq. 214, 335–346. Nilanjana, D., Vimala, R., Karthika, P., 2008. Biosorption of heavy metals-An overview. Indian J. Biotechnol. 7 (2), 159–169. Pilon-Smits, E., 2005. Phytoremediation. Annu. Rev. Plant Biol., 15–39 Peng, Z., Wen, J., Liu, Y., Zeng, G., Yi, Y., Fang, Y., Deng, J., Cai, X., 2018. Heavy metal leachability in soil amended with zeolite-or biochar-modified contaminated sediment. Environ. Monit. Assess. 190 (12), 751. Reddy, K.R., OConnor, G.A., Gale, P.M., 1998. Phosphorus sorption capacities of wetland soils and stream sediments impacted by dairy effluent. J. Environ. Quality 27 (2), 438–447. Seo, D.C., Cho, J.S., Lee, H.J., Heo, J.S., 2005. Phosphorus retention capacity of filter media for estimating the longevity of constructed wetland. Water Res. 39 (11), 2445–2457. Sheikhhosseini, A., Shirvani, M., Shariatmadari, H., 2013. Competitive sorption of nickel, cadmium, zinc and copper on palygorskite and sepiolite silicate clay minerals. Geoderma 192, 249–253. Sheoran, A.S., Sheoran, V., 2006. Heavy metal removal mechanism of acid mine drainage in wetlands: a critical review. Miner. Eng. 19, 105–116. Tang, J., Zhu, W., Kookana, R., Katayama, A., 2013. Characteristics of biochar and its application in remediation of contaminated soil. J. Biosci. Bioeng. 116 (6), 653– 659. Takeuchi, M., Kawahata, H., Gupta, L.P., Kita, N., Morishita, Y., Ono, Y., Komai, T., 2007. Arsenic resistance and removal by marine and non-marine bacteria. J. Biotechnol. 127 (3), 434–442. Unz, R.F., Shuttleworth, K.L., 1996. Microbial mobilization and immobilization of heavy metals. Curr. Opin. Biotechnol. 7 (3), 307–310. Vogel, M., Günther, A., Rossberg, A., Li, B., Bernhard, G., Raff, J., 2010. Biosorption of U (VI) by the green algae Chlorella vulgaris in dependence of pH value and cell activity. Sci. Total Environ. 409 (2), 384–395. Wang, Z.H., Guo, H.Y., Shen, F., Yang, G., Zhang, Y.Z., Zeng, Y.M., Wang, L.L., Xiao, H., Deng, S., 2015. Biochar produced from oak sawdust by Lanthanum (La)-involved pyrolysis for adsorption of ammonium (NH+4), nitrate (NO 3 ), and phosphate (PO3 4 ). Chemosphere 119, 646–653. Wang, H., Ren, Z.J., 2014. Bioelectrochemical metal recovery from wastewater: a review. Water Res. 66, 219–232. Xu, R., Zhang, Y., Liu, R., Cao, Y., Wang, G., Ji, L., Xu, Y., 2019. Effects of different substrates on nitrogen and phosphorus removal in horizontal subsurface flow constructed wetlands. Environ. Sci. Pollut. Res. 26 (16), 16229–16238. Yadav, A.K., Kumar, N., Sreekrishnan, T.R., Satya, S., Bishnoi, N.R., 2010. Removal of chromium and nickel from aqueous solution in constructed wetland: mass balance, adsorption-desorption and FTIR study. Chem Eng J. 160 (1), 122–128. Yalcuk, A., Ugurlu, A., 2009. Comparison of horizontal and vertical constructed wetland systems for landfill leachate treatment. Bioresour. Technol. 100 (9), 2521–2526. Zhu, T., Jenssen, P.D., Maehlum, T., Krogstad, T., 1997. Phosphorus sorption and chemical characteristics of lightweight aggregates (LWA)-potential filter media in treatment wetlands. Water Sci. Technol. 35 (5), 103–108. Zhou, Y., Gu, T., Yi, W., Zhang, T., Zhang, Y., 2019. The release mechanism of heavy metals from lab-scale vertical flow constructed wetlands treating road runoff. Environ. Sci. Pollut. Res. 26 (16), 16588–16595.