Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor

Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor

Accepted Manuscript Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor Cilai Tang, Pengfe...

1MB Sizes 0 Downloads 17 Views

Accepted Manuscript Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor Cilai Tang, Pengfei Sun, Jiali Yang, Yingping Huang, Yonghong Wu PII: DOI: Reference:

S0960-8524(19)30001-X https://doi.org/10.1016/j.biortech.2019.01.001 BITE 20875

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

28 November 2018 30 December 2018 1 January 2019

Please cite this article as: Tang, C., Sun, P., Yang, J., Huang, Y., Wu, Y., Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor, Bioresource Technology (2019), doi: https:// doi.org/10.1016/j.biortech.2019.01.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Kinetics simulation of Cu and Cd removal and the microbial community adaptation in a periphytic biofilm reactor

Cilai Tang1, 2, Pengfei Sun2, Jiali Yang2, Yingping Huang1, Yonghong Wu1, 2*

1

College of Hydraulic & Environmental Engineering, China Three Gorges University, Yichang, 443002, China

2

Zigui Ecological Station for Three Gorges Dam Project, Institute of Soil Science, Chinese Academy of Sciences, 71 East Beijing Road, Nanjing 210008, China

Corresponding author: Dr. Yonghong Wu 8, Daxue Road, Yichang 443002 China Tel: (+86)-0717 6392160 Email: [email protected]

1

Abstract Periphytic biofilm reactor (PBfR) shows great potential in pollutants removal. However, few studies were focused on mathematical model of pollutants removal in PBfR. A three-step PBfR was designed and a new model was developed to simulate the kinetics of Cu and Cd removal from simulated wastewater. The results show that the PBfR could remove 99.0% Cu and 99.7% Cd from liquid wastewater. The experiment data could be well fitted with a high correlation coefficients both for Cu and Cd. The microbial community in the PBfR could be self-adjusted to tolerate the toxicities of Cu and Cd, resulting in sustainable and high decontamination efficiencies. The eukaryote in the PBfR played a vital role in Cu and Cd removal. The prokaryote showed negative effect on Cu and Cd removal, though it had more diversity than eukaryote. This study provides a new approach for Cu and Cd removal and their kinetics simulation in photoautotrophic bioreactor.

Key words: Periphytic biofilm reactor; Kinetics; Copper; Cadmium; Monod-model

2

1. Introduction Bioreactor is a powerful tool to remove various pollutants from industrial (Bakar et al., 2018), domestic (Smith et al., 2012), urban (Parladé et al. 2018), and municipal wastewaters (Li et al. 2018). Therefore, it is always a research hotspot (Mannina et al. 2017, Zhang et al. 2017). Recently, photoautotrophic bioreactors (PBRs) such as algae-based PBRs have attracted increasingly attention due to their potential practical application (Luo et al. 2018). Compared to algae-based PBRs, however, periphytic biofilm reactor (PBfR) is an emerging PBRs which also shows great potential in wastewater treatment (Shangguan et al. 2015). Unlike traditional PBRs, PBfR is more abundant in microbial species which contains algae, bacteria, fungi, protozoa and, etc. (Shangguan et al., 2015). As it is well known, the decontamination efficiency of PBRs depends on the concentration of biomass. And the concentration of biomass are positive with the growth of microbes (Luo et al. 2018, Wang et al. 2018). Therefore, the more abundant microbial species in PBfR would be more efficient in decontamination (Luo et al. 2018). Moreover, the mechanism will be more complex and maybe different from traditional PBRs (Wang et al. 2018). Although lots of studies have been conducted to probe the effects of operational conditions on PBfR, few studies were focused on mathematical model of pollutants removal in PBfR. Generally, in order to better control and optimize the operational conditions, kinetic model is an effective approach, which will relate many key parameters and simplify the complex problem. Mathematical model is an effective tool to reveal the complicated processes of 3

contaminant removal in bioreactor. It is helpful for the design, prediction, and control of the bioreactor system (Liu et al. 2003, Xu et al. 2016). For PBRs, the mathematical relationship between nutrients (N, P) removal and environmental factors (e.g., pH, illumination intensity, temperature, etc.) and the growth of alga in bioreactor are always the key point (Wang et al. 2018, del Rio-Chanona et al. 2018). In fact, the growth mode of alga in PBRs is relatively simple, and Monod equation is widely used to simulate it (Ghaffar et al. 2017; Lo´pez-Meza et al. 2016). But it is quite intricate in PBfR via self-regulating. The interaction among different microorganisms, such as predation, competition, and symbiosis, are the main driving forces to renew the biomass of periphytic biofilm (Scherlach and Hertweck 2018, Liu et al. 2018). The various microbes in PBfR exhibit totally different growth behaviors (mechanisms) from in traditional PBRs. The conventional Monod model can not exactly simulate such complicated processes. Therefore, new model should be developed to better describe the relationship between the growth of biomass and pollutant removal in PBfR. But few studies have been focused on this topic. Cd and Cu are two major heavy metal pollutants in the cropland of China. 7% and 2.1% of croplands in China are contaminated by Cd and Cu, respectively, indicating the urgency and importance of developing strategies for Cd and Cu decontamination (EPA 2014, Bian et al. 2014). Many technologies have been used to remove heavy metals from wastewater, such as adsorption, ion exchange, chemical precipitation, membrane filtration, coagulation and flocculation, and biological process (Liu et al. 2018; Sajid et al. 2018). Among these techniques, adsorption including physical and biologic process is 4

an effective approach (Velkova et al. 2018). Periphytic biofilm could effectively entrap Cu even at high concentration of Cu exposure (Liu et al. 2018). But it is not sure whether PBfR is effective to remove Cu and Cd simultaneously. Consequently, a novel three-step PBfR was designed to treat Cu and Cd contaminated wastewater simultaneously. A modified saturation-type (Monod) kinetics model for Cd and Cu removal by the PBfR was developed based on the growth of periphytic biofilm. 2. Methods and materials 2.1 Experiment setup 2.1.1 Design of the PBfR A three-step PBfR consisting of three main parts was designed: the filter tank (tank 1), the core periphytic biofilm tank (tank 2), and the sedimentation tank (tank 3). All the tanks are connected by silicone tubes (length: 20 cm; inner diameters: 1cm) (Figure 1). For the tank 2, it is composed of four parts (Figure 1): 1) flat rack, a polyvinyl chloride (PVC) board (length: 45 cm, width: 35 cm, and height: 25 cm) which was used to load all the systems; 2) 14 PVC pipes with an inner diameter of 6 cm in two rows to load the intertwining of Poly Ethylene (PE) pipes; 3) 14 transparent spiral PE pipes (inner diameters: 2 mm; the total volume: 4.4 L; and the total surface area: 4.472 cm2) for periphytic biofilm loading were intertwined onto the 14 transparent PVC pipes to constitute the reaction section with the purpose of enhancing the hydraulic retention time; 4) lighting system, 2 energy saving lamps (15 W, 20 cm) were fixed at the bottom of flat rack (Figure 1A) as the light source. During the operation, wastewater was pumped into tank 1 to filter large size particle solid, and then decontamination by biofilm in tank 2, 5

followed by tank 3 to further remove the contaminant and suspended solid by sedimentation before discharge. 2.1.2Periphytic biofilm incubation Eutrophic lake water was collected from Xuanwu Lake in Nanjing of China, and used for periphytic biofilm culture. Periphytic biofilm was enriched using Woods Hole culture medium (WC medium) under the conditions of 25 ± 1 oC, 50 mol photons m-2 s-1, and a light/dark cycle of 12 h:12 h (Sun et al. 2017). The enriched periphytic biofilm was then pumped into the PE pipes pretreated using 0.1 M HCl for 12 h and flushing with distilled water for 3 times. After that, periphytic biofilm-loaded PE pipes were placed into WC medium for several days under the aforementioned conditions to assist the stable adhesion of periphytic biofilm onto the PE pipes before being installed into the PBfR for experiments. 2.1.3 Experiment operation During the 24-day wastewater treatment process, the PBfR was placed in a greenhouse under the same conditions with that of periphytic biofilm incubation. The simulated wastewater was prepared using CuSO4·5H2O,CdCl2, NH4Cl, KH2PO4 with the final concentrations of Cu 2.0 mg/L, Cd 0.1 mg/L, 2.0 mg/L nitrogen and 0.4 mg/L P. NaHCO3 was used to adjust the pH to 7.0. During the operation, the simulated wastewater was pumped using a peristaltic pump at a constant velocity of 5 mL/min (hydraulic retention time: 1.0 L/d). Dissolved oxygen (DO) in the periphytic tank was maintained in the range of 6.41±0.32 and 8.11±0.41 mg/L via aeration. To minimize periphytic biofilm shedding and stimulate periphytic biofilm growing, the bioreactor 6

was periodically backwashed based on our previous method (Shangguan et al. 2015). 2.1.4 Sampling and analyses During the operation period, each of 50 mL effluents (outflow1 and 2) were sampled every 2 days. DO in the outflows was determined with a DO meter (HQ40d, HACH, USA). The concentrations of Cu and Cd in the outflows were determined using a inductively coupled plasma mass spectrometry (ICP-MS, Optima 7700DV, Thermo Fisher) (Yang et al. 2015). The dynamics of Cu and Cd removal by the PBfR were fitted using modified Monod models (Ghaffar et al. 2017, López-Meza et al. 2016). Statistical analysis of experimental data was performed using SPSS 16.0. The removal efficiency (RE) of pollutants by the PBfR was calculated by the following equation: RE(%) =

𝐶0 −𝐶𝑡 𝐶0

× 100

Eq.1

Where, C0 and Ct are the pollutants concentration (mg/L) in feed and effluent at the time of t, respectively. In addition, periphytic biofilms in tank 2 were collected at the day of 0, 5th, and 15th to analyze the variation of microbial communities in periphytic biofilms. The genomic DNA of each periphytic biofilm sample was extracted according to the manufacturer's instruction (MOBIO PowerSoil® DNA Isolation Kit, Germany). The V4-V5 region of 16S rRNA gene and V4 region of 18S rRNA gene were amplified and sequenced using Hiseq2500 to obtain the components of prokaryote and eukaryote in each periphytic biofilm samples, respectively. For PCR amplification, a forward primer 907R (5’-CCCCGTCAATTCMTTTRAGTTT-3’)

and

a

reverse

primer

515F

(5’-GTGCCAGCMGCCGCGGTAA-3’) were used for V4-V5 region of 16S rRNA gene, 7

and a forward primer 528F (5’-GCGGTAATTCCAGCTCCAA-3’) and a reverse primer 706R (5’-AATCCRAGAATTTCACCTCCAA-3’). 2.2 Model construction 2.2.1 The development of Monod-X model Periphytic biofilm in tank 2 is the key component of PBfR. The pollutants removal by the PBfR mainly depends on the growth of periphytic biofilm (Wu 2016). Mathematically, the growth rate of microorganisms is calculated according to Monod equation as following: û𝑆

𝜇 = 𝐾𝑠+𝑆

Eq. 2

In which, μ is the growth rate; S is the concentration of limiting substrates; Ks is the monod saturation constant; û is the maximum specific growth rate. For the removal of Cu and Cd by the PBfR, it can be expressed using the modified Monod model as following (Xu et al. 2016): ds 𝑑𝑡

k𝑆

= 𝐾 +𝑆 𝑋 𝑠

Eq. 3

In which, Ks is the monod saturation constant, and it is the substrate concentration (mg/L) when µ=µmax/2; S is substrate concentration (mg/L); k is the maximum specific substrate removal rate (d-1); X is the biomass concentration (mg/L). In tank 2, the biomass would be increased due to the growth of periphytic biofilm under the appropriate condition. Then, the endogenous respiratory decay rate was considered at a low level, and the biomass of periphytic biofilm in the bioreactor at any time can be expressed as: 𝑦 = 𝑋0 + 𝑦0 (𝑆0 − 𝑆)

Eq. 4 8

Where y0 is periphytic biofilm yield coefficient (mg periphytic biofilm/mg substrate); X0 is initial biomass of periphytic biofilm (mg/L); S0 is initial concentration of substrate (mg/L). Calculation according to SCu/SCd, then S0

=

Cu/Cd which was

plugged into Monod equation, and then a new equation was generated as following: ds 𝑑𝑡

k𝑆

= 𝐾 +𝑆 (𝑋0 + 𝑦0 (𝑆0 − 𝑆))

Eq. 5

𝑠

S is the Cu and Cd concentrations (mg/L) at any time in outflows and they decrease with the increase of treatment time. Thus, it was speculated that the above equations for Cu and Cd will be plotted as decreasing curves, which are consistent with the changing trends of concentrations of Cu and Cd in artificial wastewater treated by the PBfR. 2.2.2 Parameters estimation and evaluation Based on the experiment data, X0 varied within the range of 10-70 mg/L. It not only showed little effect on model construction, but also reduced the difficulty in data fitting. The parameter of y0 is stable, and thus the value of 0.67 mg periphytic biofilm/mg substrate was recommended in this study. Therefore, only k and Ks were remain unknown to build the corresponding Monod-Cu/Cd models (Xu et al. 2016). We set y=ds/dt, X0=50, and y0=0.67 to simplify the fitting process. Consequently, two models of Monod-Cu and Monod-Cd for Cu and Cd removal by the PBfR could be obtained. 3. Results and discussion 3.1Performance of the PBfR The results of Cu and Cd removal by PBfR were presented in Figure 2. The residual Cu in both outflow 1 and 2 changed in wave-modes which was probably due to the enrichment and release of Cu by periphytic biofilm in the reactor (Yang et al. 2015). 9

After 24d operation, the Cu removal efficiency by tank 2 and tank 3 were 64.0% and 99.0%, respectively (Figure 2A), indicating that both tank 2 and tank 3 could efficiently remove Cu. Finally, the average removal rate of Cu by the PBfR was 1.54 mg/d. For Cd, the removal efficiency reached equilibrium at 2ndd. After 24d, the average removal efficiency of Cd by the tank 2 was 80.0% and the total Cd removal efficiency by the PBfR was 99.7%, indicating the importance of tank 3 in Cd removal (Figure 2B). The average removal rate of Cd by the PBfR was 0.07 mg/d. The results showed that the PBfR exhibited synchronous and high-efficiency for Cu and Cd removal. The high removal efficiencies of Cu and Cd by PBfR were mainly due to both the abundant extracellular polymeric substances in periphytic biofilm (Liu et al. 2018), and microbial activity in tank 2 (Alam et al. 2015). On the other hand, Cu and Cd would pose toxicity to periphytic biofilm via inhibiting its photosynthesis and carbon utilization functions (Yang et al. 2015), and then induce the death of some periphytic biofilm, which led to the release of Cu and Cd to liquid phase, especially for Cu. It was consistent with the obtained results. 3.2 Models establishment for Cu and Cd removal by PBfR As shown in Table 1, the k for Cu and Cd removal were 0.0023 d-1and 0.0007 d-1, respectively. Generally, the parameter k indicates the reaction velocity. For easier enrichment of heavy metals, the value of k is bigger, vice versa, its value is lower (Mora et al. 2015). Additionally, the Ks for Cu and Cd removal were 0.0667 mg/L and 0.0065 mg/L, respectively. Ks shows relatively smaller and opposite effect on heavy metals enrichment compared to k, namely, the smaller of Ks, the easier enrichment of heavy 10

metals (Mora et al. 2015). Generally, the correlation coefficient reflects the conformity of fitting. The high correlation coefficients (R2=0.8309 for Cu and R2=0.8812 for Cd) imply good fitting of Cu and Cd in artificial wastewater by PBfR using Monod-Cu and Monod-Cd models. Finally, two models for Cu and Cd removal by the PBfR are obtained as following: For Cu removal: 0.0023𝑆

𝑦 = (0.0667+𝑆) (50 + 0.67 × (2 − 𝑆))

Eq. 6

For Cd removal: 0.0007𝑆

𝑦 = (0.0065+𝑆) (50 + 0.67 × (0.1 − 𝑆))

Eq. 7

In which, S is the concentrations of Cu and Cd (mg/L) in outflows at any time. 3.3 Models validation Model validation was based on the comparison between the model prediction and the experiment data for Cu and Cd removal by the PBfR (Xu et al. 2016). As shown in Figure 4, the experimental data for Cu and Cd removal by the PBfR could well match the model except for individual data. It is consistent with the correlation coefficients in Table 1. During the validation, the abnormal fluctuation of experimental data was probable due to the inhibition effect of periphytic biofilm by released Cu and Cd (Yang et al. 2015). Generally, Monod model is appropriate to simulate the growth rules of microorganisms (López-Meza et al. 2016, Lobry et al. 1992). The results show that this model derived from Monod-X is also appropriate for Cu and Cd removal by PBfR. It provides a new approach to model pollutants removal by PBfR and similar phototrophic bioreactors. 11

3.4 Microbial community in PBfR 3.4.1Taxonomic composition of periphytic biofilm in tank 2 Microorganisms play a key role for Cu and Cd removal in tank 2. Thus, the microbial communities were further studied in tank 2 at different periods of treatment. For the prokaryote, Proteobacteria was the most abundant phylum in the beginning of test (Day 0), comprising 33.1% of all phyla. Its relative abundance increased to 40.4% and 70.8% at the 5th and 15th d, respectively (Figure 4A). It is consistent with previous study that Cu and Cd contamination increased the abundance of Proteobacteria in bioreactor (Liu et al. 2017) and soil (Feng et al. 2018). However, a remarkable decrease of Cyanobacteria was observed from initial 33.6% to 18.74 and 7.4% at 5thd and 15thd, respectively. It implied that Cyanobacteria is sensitive to Cu and Cd. The abundance of Bacteroidetes increased firstly from 12.2% to nearly 20%, and then decreased to below 4%. Other functional bacterias such as Firmicutes, Planetomyeetes, Acidobacteria, Chlorobi, Actinobacteria, Verrucomicrobia, Spirochaetes, Thermi, Armatimonadetes, TM6, OD1, Nitrospirae and NKB19 were very small and they almost kept stable in the whole period of operation. For Eukaryote, the abundance of Opisthokonta decreased remarkably from 77.6% (Day 0) to 34.0% (Day 5) and 22.3% (Day 15) (Figure 4B). It indicated that Opisthokonta was sensitive to Cu and Cd exposures. While the abundance of Archaeplastida greatly increased from 0.9% (Day 0) to 6.6% (Day 5) and 68.3% (Day 15), implying that it had strong toleration to Cu and Cd toxicities. For SAR, its change was dynamic. Its zenith was observed at 5th day, but similar proportion at beginning and 15th day. The other functional bacterias decreased dramatically from 10% 12

at 5th day to below 2% after 15 days. The result indicated that the microorganisms in the PBfR could be self-adjusted to adapt to new environment and keep a sustainable and high removal efficiencies of Cd and Cu. Different dominant bacteria existed at different periods of operation and contributed to fast removal of Cu and Cd. 3.4.2 Contribution of microbial communities for Cu and Cd removal Partial Least Squares Path Modelling was employed to synthetically illustrate the contribution of prokaryote and Eukaryote to Cu and Cd removal in PBfR (Figure 5). Prokaryote showed negative effect to both Cu (path coefficient: -0.87) and Cd (path coefficient: -0.80) removal, while eukaryote showed positive effect to Cu (path coefficient: 0.58) and Cd (path coefficient: 0.68) removal. Although the diversity of eukaryote was less than prokaryote, the eukaryote played the primary role in Cu and Cd removal. 4. Conclusion A mathematical model was developed to simulate the kinetics of Cu and Cd removal in the PBfR. The developed model could well reflect the experiment data of Cu and Cd removal by PBfR. The PBfR could self-adjust the microbial communities to tolerate Cu and Cd exposures. Eukaryote played a vital role in Cu and Cd removal in PBfR. However, prokaryote showed negative effect to Cu and Cd removal, though it had more diversity than eukaryote. These results not only provide a new approach for treating Cu and Cd contaminated wastewater, but also offer a reference for modeling pollutants removal dynamics by biofilm-based bioreactors.

13

Acknowledgements This work was supported by National Natural Science Foundation of China (31772396, 21876097) and the Natural Science Foundation of Jiangsu Province, China (BK20150066). This work was also supported by Chinese Academy of Sciences Interdisciplinary Innovation Team.

Reference Alam, M.A., Wan, C., Zhao, X.Q., Chen, L.J., Chang, J.S. and Bai, F.W. (2015) Enhanced removal of Zn(2+) or Cd(2+) by the flocculating Chlorella vulgaris JSC-7. Journal of Hazardous Materials 289, 38-45. Bian, R., Joseph, S., Cui, L., Pan, G., Li, L., Liu, X., Zhang, A., Rutlidge, H., Wong, S. and Chia, C. (2014) A three-year experiment confirms continuous immobilization of cadmium and lead in contaminated paddy field with biochar amendment. Journal of Hazardous Materials 272, 121-128. Del Rio-Chanona, E.A., Liu, J., Wagner, J.L., Zhang, D., Meng, Y., Xue, S. and Shah, N. (2018) Dynamic modeling of green algae cultivation in a photobioreactor for sustainable biodiesel production. Biotechnology and Bioengineering 115(2), 359-370. EPA, http://www.sdpc.gov.cn/fzgggz/ncjj/zhdt/201404/t20140418_607888.html Feng, G., Xie, T., Wang, X., Bai, J., Tang, L., Zhao, H., Wei, W., Wang, M. and Zhao, Y. (2018) Metagenomic analysis of microbial community and function involved in cd-contaminated soil. BMC microbiology 18(1), 11. Ghaffar, S., Stevenson, R.J. and Khan, Z. (2017) Effect of phosphorus stress on 14

Microcystis aeruginosa growth and phosphorus uptake. PLOS One 12(3), e0174349. Li, R.H., Wang, X.M. and Li, X.Y. (2018) A membrane bioreactor with iron dosing and acidogenic co-fermentation for enhanced phosphorus removal and recovery in wastewater treatment. Water Research 129, 402-412. Liu, Y., Xu, H.L., Yang, S.F. and Tay, J.H. (2003) Mechanisms and models for anaerobic granulation in upflow anaerobic sludge blanket reactor. Water Research 37(3), 661-673. Liu, J., Wang, F., Wu, W., Wan, J., Yang, J., Xiang, S. and Wu, Y. (2018) Biosorption of high-concentration Cu (II) by periphytic biofilms and the development of a fiber periphyton bioreactor (FPBR). Bioresource Technology 248, 127-134. Liu, X., Hua, Y., Tang, S., Mi, F., Hui, P., Lu, G., Liu, Z. and Zhi, D. (2017) Effects of single and combined copper/perfluorooctane sulfonate on sequencing batch reactor process and microbial community in activated sludge. Bioresource Technology 238, 407-415. Lobry, J.R., Flandrois, J.P., Carret, G. and Pave, A. (1992) Monod's bacterial growth model revisited. Bulletin of Mathematical Biology 54(1), 117-122. Luo, Y., Leclech, P. and Henderson, R.K. (2018) Assessment of membrane photobioreactor (MPBR) performance parameters and operating conditions. Water Research 138, 169-180. López-Meza, J., Araíz-Hernández, D., Carrillo-Cocom, L.M., López-Pacheco, F., Rocha-Pizaña, M.d.R. and Alvarez, M.M. (2016) Using simple models to describe 15

the kinetics of growth, glucose consumption, and monoclonal antibody formation in naive and infliximab producer CHO cells. Cytotechnology 68(4), 1287-1300. Mannina, G., Capodici, M., Cosenza, A. and Di, T.D. (2018) Nitrous oxide from integrated fixed-film activated sludge membrane bioreactor: Assessing the influence of operational variables. Bioresour Technology 247, 1221-1227. Mora, M., Dorado, A.D., Gamisans, X. and Gabriel, D. (2015) Investigating the kinetics of autotrophic denitrification with thiosulfate: Modeling the denitritation mechanisms and the effect of the acclimation of SO-NR cultures to nitrite. Chemical Engineering Journal 262, 235-241. Parladé, E., Homdiaz, A., Blánquez, P., Martínezalonso, M., Vicent, T. and Gaju, N. (2018) Effect of cultivation conditions on β-estradiol removal in laboratory and pilot-plant photobioreactors by an algal-bacterial consortium treating urban wastewater. Water Research 137(11), 86-96. Sajid, M., Nazal, M.K., Ihsanullah, Baig, N., Osman, A.M. (2018) Removal of heavy metals and organic pollutants from water using dendritic polymers based adsorbents: A critical review. Separation and Purification Technology 191, 400-423. Scherlach, K. and Hertweck, C. (2018) Mediators of mutualistic microbe-microbe interactions. Natural Product Reports 35(4), 303-308. Shangguan, H., Liu, J., Zhu, Y., Tong, Z. and Wu, Y. (2015) Start-up of a spiral periphyton bioreactor (SPR) for removal of COD and the characteristics of the associated microbial community. Bioresource Technology 193(10), 456-462. Sun, P., Zhang, J., Esquivel-Elizondo, S., Ma, L. and Wu, Y. (2018) Uncovering the 16

flocculating potential of extracellular polymeric substances produced by periphytic biofilms. Bioresource Technology 248, 56-60. Velkova, Z., Kirova, G., Stoytcheva, M., Kostadinova, S., Todorova, K., Gochev, V. (2018) Immobilized microbial biosorbents for heavy metals removal. Engineering in Life Sciences 18(12), 871-881. Wang, M., Payne, K.A., Tong, S. and Ergas, S.J. (2018) Hybrid algal photosynthesis and ion exchange (HAPIX) process for high ammonium strength wastewater treatment. Water Research 142, 65-74. Wu, Y. (2016) Periphyton: Functions and Application in Environmental Remediation, Elsevier. Xu, G., Yin, F., Chen, S., Xu, Y. and Yu, H.-Q. (2016) Mathematical modeling of autotrophic denitrification (AD) process with sulphide as electron donor. Water Research 91, 225-234. Zhang, W., Tang, B. and Bin, L. (2017) Research Progress in Biofilm-Membrane Bioreactor: A Critical Review. Industrial & Engineering Chemistry Research 56(24), 6900-6909.

17

Figure and table captions Figure 1. The flow diagram of the PBfR (Tank 1: filter tank; Tank 2: periphytic biofilm tank; Tank 3: sedimentation tank). A is the periphytic biofilm reactor in Tank 2, Tank 2, and B is the inner structure of A (1: flat rack; 2: PVC pipe; 3: PE pipe; 4: clapboard; 5: lighting system; 6: wastewater flow) Figure 2. The concentration profile of Cu (A) and Cd (B) in the outflow. Figure 3. The model validation using experimental data (A: Monod-Cu; B: Monod-Cd). Figure 4. Taxonomic composition of periphytic biofilm in tank 2. (A) prokaryote, (B) eukaryote. Figure 5. Contribution of prokaryote and eukaryote to Cu and Cd removal analyzed by Partial Least Squares Path Modelling.

Table 1. Fitting parameters of modified Monod model for Cu and Cd

18

Figure 1. The flow diagram of the PBfR (Tank 1: filter tank; Tank 2: periphytic biofilm tank; Tank 3: sedimentation tank). A is the periphytic biofilm reactor in Tank 2, Tank 2, and B is the inner structure of A (1: flat rack; 2: PVC pipe; 3: PE pipe; 4: clapboard; 5: lighting system; 6: wastewater flow)

19

Figure 2. The concentration profile of Cu (A) and Cd (B) in the outflow.

20

Figure 3.The model validation using experimental data (A: Monod-Cu; B: Monod-Cd).

21

Figure 4.Taxonomic composition of periphytic biofilm in tank 2. (A) prokaryote, (B) eukaryote.

22

Figure 5. Contribution of prokaryote and eukaryote to Cu and Cd removal analyzed by Partial Least Squares Path Modelling.

23

Table 1. Fitting parameters of modified Monod model for Cu and Cd Model Equation Reduced Chi-Sqr Adj. R-Square

B

k Ks

Model Equation Reduced Chi-Sqr Adj. R-Square

B

k Ks

Monod-Cu y=((K×x)/(Ks+x)) ×(50+0.67×(2-x)) 0.0118 0.8309 Value Standard Error 0.0023 5.4394E-4 0.0667 0.0026 Monod-Cd y=((K×x)/(Ks+x))×(50+0.67×(0.1-x)) 1.81E-05 0.8812 Value Standard Error 6.73E-05 1.19E-05 0.0065 1.12E-04

24

Highlights



A three-step periphytic biofilm reactor (PBfR) was designed for Cu and Cd removal.



The PBfR could sustainably and effectively remove Cu and Cd from wastewater.



The experiment data could be well fitted with a modified Monod model.



The microbial community could be self-adjusted to tolerate the toxicities of Cu or Cd.



The eukaryote in PBfR played a vital role in Cu and Cd removal.

25