English

• Gravity-driven membrane filtration (GDM) has been widely utilized in decentralized water treatment, and industrial water purification due to its negligible energy consumption, and low operation & maintenance [1,2]. However, the high DOC concentration in the effluent and low stable flux limit its wide application [3]. The poor DOC removal was related to the hydrolysis of particulate organic matter by microorganisms within the biocake layer [4]. Pretreatment before the membrane was an effective way to enhance pollutant removal and improve water quality. The earlier study found that integrating the biofilm reactor into the GDM reactor was beneficial in improving DOC removal during long-term filtration, leading to improved permeate quality [5]. Previous literature coupling sand layers with GDM suggested that the removal efficiency of organics was improved by 20%-25% [6]. Besides, vermifiltration (VF) pretreatment was developed for decentralized wastewater treatment and indicated that effective pollutants were removed, such as 95.97% TOC and 45.25% NH₃-N via the VF process [7]. Granular active carbon (GAC) as an environmentally friendly material, has been used in the water treatment field and can be coupled with GDM in the form of pretreatment to contribute to significant pollutant removal (e.g., DOC, and UV₂₅₄) and effluent quality improvement [8-10].

  Besides, the flux level was another factor to impacted the extensive application of GDM technology, which was closely related to the structural characteristics of the biocake layer [11]. The formed biocake layer on the membrane surface without any interventions became dense and flat, resulting in severe membrane fouling and low membrane flux. Many scholars have made efforts hoping to engineer a rough, heterogeneous, and porous biocake layer beneficial to GDM flux improvement [5]. In-situ modification of the biocake layer is an effective measure to alleviate membrane fouling and enhance the membrane flux of the GDM system. A previous report suggested that pre-coating manganese oxides on the membrane surface of GDM system engineered heterogeneous biocake structures and the stable flux improved by 20% [12]. Another study found that depositing larger particles on the membrane surface helped to form a more heterogeneous and permeable biocake layer structure, which was beneficial to alleviate membrane fouling [13]. In recent years, magnetic particles as a fouling reducer were introduced to the membrane bioreactor (MBR) to control the membrane fouling due to their advantages of magnetobiological effects and easy separation [14]. Magnetic particles could affect the MBR microbial community structure, to reduce the pioneer bacteria from bulk sludge and result in less SMP production [14]. Another report even pointed out that magnetic particles selectively enhanced the organic substances' degrading bacteria growth and decreased the ratio of bacteria related to membrane fouling [15]. It was inferred that introducing magnetic particles into the GDM would probably produce an effect on GDM membrane fouling. However, the effects of magnetic particles on the biocake layer properties and GDM performance were still unclear. Furthermore, the previous studies mainly focused on the single pretreatment or biocake layer modification, and there was no report on the combined effects of magnetic particles and GAC, and their respective role in the GDM performance was not clearly expounded.

  Therefore, this study combined GAC and magnetic particles to improve the GDM system performance. Furthermore, the effect of GAC and magnetic particles on the GDM flux variation and removal performance was evaluated. The possible mechanism of combined GAC and magnetic particles was discussed from the biocake layer structure, composition, and microbial community. This study aimed to provide a fundamental for developing sustainable and high-efficiency GDM technology.

  The descriptions of feed water and experimental setup were present in Text S1 (Supporting information) and setup schematic diagram was shown in Fig. S1 (Supporting information). The detailed information and characteristics of magnetic particles were provided in Text S2 and Fig. S2 (Supporting information). The experimental materials were listed in Text S3 (Supporting information). The analytic methods and high-throughput sequencing were offered in Text S4 (Supporting information) and Text S5 (Supporting information), respectively. The details of the economic analysis are shown in Text S6 (Supporting information).

  The flux developments with the filtration in the three GDM groups are displayed in Fig. 1a and were roughly divided into two phases, i.e., flux decline and flux stability. In the phase of flux decline (days 1–12), the flux rapidly dropped with a decrease rate of 85%, 81%, and 78% in GDM1, GDM2, and GDM3, respectively, which was attributed to the fast biocake formation and membrane pore blocking [16]. Compared with GDM1, the dropping rate in the flux of GDM2 was slower since fewer pollutants were rejected by the GDM module due to GAC pretreatment (Fig. S3 in Supporting information). The slowest decrease rate of flux was obtained in GDM3, which was related to that the magnetic particles deposited on the membrane surface served as a dynamic barrier to prevent direct contact between containments and the membrane surface [17]. The results suggested the magnetic particles assisted GDM conferred higher capacity in membrane fouling mitigation. In the phase of flux stability (days 13–42), the flux in GDM1, GDM2, and GDM3 stabilized at 2.3, 4.0, and 4.5 L m⁻² h⁻¹, respectively. Compared to GDM1, the stable flux of GDM2 increased by about 74% due to GAC pretreatment reducing pollutant load into the membrane module. The GDM3 stable flux was improved by 96% and 13% than GDM1 and GDM2, suggesting magnetic particles had a significant advantage in mitigating membrane fouling. Supportively, a previous study reported that magnetic particles could reduce membrane filtration resistance and improve membrane permeability [15]. Other literature also demonstrated that the magnetic particles could decrease biomass secretion and change microbial community, contributing to lower membrane fouling [18,19]. In this study, the mitigating membrane fouling by the magnetic particles was attributed to the modified biocake layer characteristics and their bioeffect on microbial community and biomass. Overall, combining GAC and magnetic particles contributed to the furthest mitigating membrane fouling and improving flux.

  Figure 1

  

  Figure 1.  Flux and resistance developments of GDM systems. (a) Flux variation, (b) total resistance, and (c) stable flux. Capital letters represent the significant difference analyzed by post hoc Tukey's test (P < 0.05).

  DownLoad: Full-Size Img PowerPoint

  The contaminants removals were analyzed to evaluate the effect of magnetic particle assistance on the GDM performances, as depicted in Fig. S3 and detailed in Text S7 (Supporting information).

  The extracellular polymeric substance (EPS) and soluble microbial products (SMP) in the biocake layer were closely related to the membrane fouling [20]. As shown in Figs. 2a and b, the biocake layer in GDM1 had relatively higher EPS and SMP concentrations than the other GDM systems, which was consistent with its denser structure of biocake layer (Fig. 3) and lower stable flux (Fig. 1). Compared with GDM1, coupling GAC to GDM system was beneficial to reducing the content of EPS and SMP, accounting to average reductions of 36.8% for EPS and 22.3% for SMP, which was related to effectively removal of pollutants by GAC filter (Fig. S3) and positive variations in the microbial community of the biocake layer (Fig. 4) [21]. Compared with GDM2, magnetic particle assistance could further benefit the reduction of EPS and SMP, achieving a synchronous removal of 12% LB-EPS polysaccharide, 13% TB-EPS protein, and 21% SMP, which was attributed to the extra bioeffect of magnetic particles. According to previous literature, adding magnetic particles into the MBR system significantly reduced the secretions of SMP polysaccharide (SMPc) and EPS protein (EPSp) due to the enhancement of dehydrogenase activity [22]. Another study pointed out that magnetic particles could induce microbial community evolution and decrease the ratio of microorganisms related to metabolite secretion, resulting in less EPS and SMP secretion [14,15].

  Figure 2

  

  Figure 2.  Biomass analysis: (a, b) Polysaccharide and protein in the biocake layer. (c-e) The fluorescent pollutant of EPS and SMP in GDM1, GDM2, and GDM3, respectively.

  DownLoad: Full-Size Img PowerPoint

  Figure 3

  

  Figure 3.  Morphological characterization. (a-c) SEM of GDM1, GDM2, and GDM3. (d-f) AFM 2D image of GDM1, GDM2, and GDM3. (g-i) AFM 3D image of GDM1, GDM2, and GDM3. (j) Surface EDS of GDM3, (k) cross-section EDS of GDM3.

  DownLoad: Full-Size Img PowerPoint

  Figure 4

  

  Figure 4.  Microbial community composition: (a) bacterial phylum, (b) eukaryotic phylum, (c) bacterial genus, and (d) eukaryotic genus.

  DownLoad: Full-Size Img PowerPoint

  The three-dimensional fluorescence spectrum was used to analyze the fluorescence characteristics of EPS and SMP (Figs. 2c-e). The fluorescence spectrum of EPS and SMP mainly included peak1 (Ex/Em = 225/325 nm) and peak2 (Ex/Em = 275/325 nm), which was assigned to aromatic proteins and tryptophan protein-like substances [23], demonstrating protein was the main component of biomass. The fluorescence intensity followed the order of TB-EPS highest, SMP second, and LB-EPS lowest, which was consistent with their protein content distribution trend. Besides, the fluorescence intensity of TB-EPS, loosely bound EPS (LB-EPS), and SMP in the biocake layer of GDM2 was significantly reduced by GAC. By contrast, with the magnetic particle involvement, the fluorescence intensity in GDM3 had a slight decline relative to GDM2, suggesting that magnetic particle assistance contributed to decreasing further fluorescence pollutants in EPS and SMP, unified with protein content variations. Overall, combining GAC and magnetic particles could furthest diminish the concentrations of both EPS and SMP and fluorescence intensity, which was beneficial to the membrane fouling mitigation and flux improvement.

  The biocake layer morphology was systematically characterized to investigate the effect of GAC and magnetic particles on it. As shown in Fig. 3a, the biocake layer of GDM1 was relatively flat and dense, which was consistent with its high biomass, resulting in a relatively low stable flux level (2.3 L m⁻² h⁻¹, Fig. 1). Concerning GDM2 (Fig. 3b), porous and heterogeneous structures of biocake layer were observed due to GAC prefilter, which helped to improve the membrane permeability (4.0 L m⁻² h⁻¹, Fig. 1). As for GDM3, the magnetic particles involvement made the biocake layer highly heterogeneous and rough and appeared more water transport channels (Fig. 3c) [4].

  The AFM was used to characterize the biocake layer roughness. As described in Fig. 3d, GDM1 had a relatively low roughness (255 ± 32 nm) of biocake layer without any obvious peak and valley fluctuation, consistent with its flat and dense structure. The biocake layer roughness in GDM2 (Fig. 3e) was elevated by about 58% more than in GDM1, and peak and valley fluctuation related to the water holes was observed. Besides, GDM3 noticed more obvious peak and valley structures (Fig. 3f) and showed that there were more water channels, consistent with the results observed by scanning electron microscope (SEM). GDM3 had the highest roughness (663 ± 59 nm), owing to the magnetic particle deposition in the biocake layer, which was demonstrated by Fe mapping (Figs. 3g-i). Overall, magnetic particle assistance was beneficial to engineer more open, connected, and rough biocake layer structures, favorable for the GDM stable flux improvement.

  The bacterial and eukaryotic community constitutions were analyzed to explore the effect of interference by GAC and magnetic particles on them. As illustrated in Table S1 (Supporting information), compared to GDM1, coupling GAC into GDM (GDM2) could significantly improve the eukaryotic abundance and diversity based on the indexes of Simpson, Shannon, Abundance-based Coverage Estimator (ACE), and Chao1. Furthermore, GDM3 had higher Simpson, ACE, and Chao1 than GDM1 and GDM2, indicating that adding magnetic particles into GDM was beneficial to improving the microbial diversity and richness of biocake layer [24]. A comprehensive analysis of Veen graphs was provided in Text S8 (Supporting information).

  In the phylum, Planctomycetota, Proteobacterial, and Cyanobacteria were the dominant bacterial phylum in GDM systems (Fig. 4a). Planctomycetota was the most abundant species in three GDM systems, accounting for 38.2%, 32.7% and 32.3% in GDM1, GDM2, and GDM3. Proteobacterial was relevant to pollutant removal since they included various functional bacteria for the biodegradation of organic pollutants [25]. The more abundant Proteobacterial in GDM2 (38.1%) and GDM3 (33.3%) contributed to the effective removal of organic pollutants than in GDM1 (26.8%). Besides, GDM1 had a higher Cyanobacteria percentage (16.2%) than GDM2 (5.2%) and GDM3 (7.9%). In the GDM system, alga breeding restricted other microorganisms' growth within the biocake layer and promoted more secretions released by microorganisms [26]. Thus, the relatively low Cyanobacteria ratios in GDM2 and GDM3 contributed to reducing microbial secretion, consistent with the analytical results of EPS and SMP, conducive to mitigating the membrane fouling.

  In the eukaryotic phylum level (Fig. 4b), Ciliophora with cilium allowed itself to cross via the biocake layer, which was beneficial for dredging the fouling layer. Cercozoa with cilium was able to wriggle, which could make the biocake layer more porous [27]. Besides, other eucaryons (i.e., Annelida and Rotifera) characterized by strong athletic ability were also detected in GDM systems (classified as the "others"). The movements of eukaryotic microorganisms played an important role in modifying biocake layer structure [28]. In this study, the considerable richness of eucaryon endowed with mobile features in GDM2 and GDM3 was conducive to engineering a more rough and open structure.

  The top 12 bacterial evolution at the genus level was visualized by Circos graphs (Fig. 4c). These Gemmata (11.3%–14.7%), Chloroplast_norank (3.1%–9.2%), Fimbriiglobus (4.5%-5.6%), and Pirellula (1.6%-7.8%) dominant biocake layer. Thereinto, Gemmata, Fimbriiglobus, and Pirellula were divided into Planctomycetota. Chloroplast_norank belonging to Cyanobacteria accounted for a higher ratio (9.2%) in GDM1 than that in GDM2 (3.1%) and GDM3 (4.9%), which was responsible for high biomass in GDM1. Notably, many bacterial genera were classified into others with a proportion of 43.6%, 49.3% and 55.9% in GDM1, GDM2 and GDM3, which further proved that magnetic particles assistance helped to improve the microbial abundance in biocake layer, unified with the result in Fig. S4 (Supporting information).

  In the eukaryotic genus level, Prunus, Desmodesmus, and Selaginella were the dominant species in the biocake layer (Fig. 4d). Prunus related to Streptophyta, occupied large proportions of 13.8% in GDM1, 17.2% in GDM2, and 3.8% in GDM3. Desmodesmus belonging to Chlorophyta, accounted for a relatively high abundance in GDM1 (16.6%), however, the presence of GAC and magnetic particles decreased its ratio, with percentages of 8.1% in GDM2 and 9.1% in GDM3, respectively. Selaginella was included in Streptophyta, with a percentage of 0.8% in GDM1, 11.0% in GDM2, and 20.5% in GDM3. Previous study reported magnetic particles could induce the microbial community evolution, and change microbial richness and diversity due to their bioeffect [29]. Another literature indicated that magnetic particles affected microorganisms associated with biomass secretions and resulted in fewer metabolites, contributing to membrane fouling mitigation [14].

  Overall, magnetic particle assistance enhanced the microbial abundance and diversity, and changed the microbial community structure, contributing to less production of EPS and SMP compared to GDM1 and GDM2, positive for stable flux improvement.

  Economic analysis was performed to evaluate the economic feasibility of magnetic particle-assisted GDM technology. As shown in Fig. S5a (Supporting information), the GDM1 had a relatively higher cost (0.21958 US$/m³) than GDM2 (0.21548 US$/m³) and GDM3 (0.20318 US$/m³), due to its low stable flux (2.3 L m⁻² h⁻¹), resulting in a higher membrane investment (about 84%, Fig. S5b in Supporting information). Although GAC increased the material investment with about a 40% ratio in total cost, the total cost of GDM2 did not increase compared with GDM1. Because introducing GAC improved the stable flux by 74% in GDM2, significantly reducing membrane cost (reducing by 41.7%). Besides, the GDM with magnetic particle assistance (GDM3) had the lowest cost (0.20318 US$/m³), with a decrease of 7.5% and 5.7% relative to GDM1 and GDM2, indicating magnetic particle-assisted GDM technology was economically feasible. The lowest cost in GDM3 was because magnetic particles helped to increase the stable flux by 96%, contributing to an obvious reduction in the membrane investment (reducing by 48.8%). Overall, introducing magnetic particles into the GDM system significantly decreased the membrane investment, leading to the lowest total cost. Therefore, magnetic particle-assisted GDM technology was economically feasible.

  The potential mechanism of combined GAC and magnetic particles in enhancing GDM filtration performance is described in Fig. 5. GAC effectively reduced the polluted loading into the membrane module and improved permeate quality. Besides, the GAC promoted the formation of a rough and heterogeneous biocake layer. For magnetic-particle, the GDM performance improvement by magnetic particle was attributed to the biocake layer characteristics modification and the magnetic particle bioeffect. On the one hand, the magnetic particles (GDM3) presence contributed to engineering a more open and connected structure with higher roughness relative to GDM1 and GDM2, which was favorable for the improvement of stable flux [3]. Previous publications proved that the presence of magnetic particles would induce microbial community evolution, and change microbial richness and diversity, resulting in microbial metabolite variation which was attributed to their bioeffect [14,15,29]. Combined GAC and magnetic particles contributed to GDM performance improvement.

  Figure 5

  

  Figure 5.  Mechanism of performance improvement in the magnetic particles assisted GDM system.

  DownLoad: Full-Size Img PowerPoint

  In this study, magnetic particles were introduced and their effect on GDM performance, biocake layer structure, composition, and microbial community was investigated. The conclusions were obtained as follows: GDM3 achieved highly effective pollutant removals (85% COD_Mn, 95% UV₂₅₄, and 65% DOC) and significant flux improvement (96%) than GDM itself due to the combined GAC and magnetic particle. GAC mainly helped to reduce pollutant load and improve permeated quality while magnetic particles contributed to engineering more open and connected structures with high roughness (663 ± 59 nm). Besides, adding magnetic particles significantly changed the microbial community structure of GDM3, benefiting to producing less EPS and SMP compared to GDM1 and GDM2 due to their bioeffect. Furthermore, GDM3 has the best economic benefit, i.e., low total cost with a decrease of 7.5% and 5.7% than GDM1 and GDM2.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Xishou Guo: Writing – original draft, Visualization, Methodology, Investigation, Conceptualization. Haochun Wang: Writing – review & editing, Visualization, Investigation. Zixin Ma: Writing – review & editing, Visualization. Jinlong Wang: Writing – review & editing, Visualization, Methodology, Formal analysis. Yuchao Tang: Supervision, Resources, Funding acquisition. Guibai Li: Supervision, Resources. Heng Liang: Supervision, Resources, Funding acquisition. Xiaobin Tang: Writing – review & editing, Supervision, Resources, Funding acquisition.

  Acknowledgments

  This work was supported by the National Key Research and Development Program of China (No. 2023YFC3208002), National Natural Science Foundation of China (No. 52370007), Excellent Youth Foundation of Hei Long Jiang Province of China (No. YQ2022E034), and Anhui Provincial Key Laboratory of Environmental Pollution Control and Resource Reuse (No. 2021EPC02).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi: 10.1016/j.cclet.2024.110781.

  
  
• 1. [1]

     M. Peter-Varbanets, F. Hammes, M. Vital, et al., Water Res. 44 (2010) 3607–3616. doi: 10.1016/j.watres.2010.04.020

  2. [2]

     Y.T. Tsai, I.V. Maggay, A. Venault, et al., Sep. Purif. Technol. 279 (2021) 119720. doi: 10.1016/j.seppur.2021.119720

  3. [3]

     W. Pronk, A. Ding, E. Morgenroth, et al., Water Res. 149 (2019) 553–565. doi: 10.1016/j.watres.2018.11.062

  4. [4]

     X. Guo, S. Jiang, Y. Wang, et al., Sep. Purif. Technol. 303 (2022) 122122. doi: 10.1016/j.seppur.2022.122122

  5. [5]

     B. Wu, F. Hochstrasser, E. Akhondi, et al., Water Res. 93 (2016) 133–140. doi: 10.1016/j.watres.2016.02.021

  6. [6]

     A. Ding, J.L. Wang, D.C. Lin, et al., Chemosphere 191 (2018) 253–261. doi: 10.1016/j.chemosphere.2017.10.034

  7. [7]

     J.L. Wang, X.B. Tang, Y.Y. Liu, et al., Resour. Conserv. Recycl. 177 (2022) 106008. doi: 10.1016/j.resconrec.2021.106008

  8. [8]

     J. Zhang, S. Yu, J. Wang, et al., Water Res. 231 (2023) 119606. doi: 10.1016/j.watres.2023.119606

  9. [9]

     X. Tang, W. Pronk, A. Ding, et al., Chem. Eng. J. 333 (2018) 289–299. doi: 10.1016/j.cej.2017.09.111

  10. [10]

      S. Shao, Y. Feng, H. Yu, et al., Water Res. 108 (2017) 240–249. doi: 10.1016/j.watres.2016.10.081

  11. [11]

      J. Ding, H. Xiao, X. Huang, et al., Chin. Chem. Lett. 34 (2023) 108316. doi: 10.1016/j.cclet.2023.108316

  12. [12]

      X.B. Tang, B.H. Xie, R. Chen, et al., Chem. Eng. J. 397 (2020) 125248. doi: 10.1016/j.cej.2020.125248

  13. [13]

      A. Chomiak, B. Sinnet, N. Derlon, et al., Water Res. 64 (2014) 177–186. doi: 10.1016/j.watres.2014.06.045

  14. [14]

      Y. Liu, Q. Liu, J. Li, et al., Bioresour. Technol. 249 (2018) 377–385. doi: 10.1016/j.biortech.2017.10.027

  15. [15]

      H. He, X. Xin, W. Qiu, et al., Water Res. 209 (2022) 117897. doi: 10.1016/j.watres.2021.117897

  16. [16]

      X. Guo, Y. Wang, S. Jiang, et al., Chemosphere 310 (2023) 136692. doi: 10.1016/j.chemosphere.2022.136692

  17. [17]

      J. Wang, X. Tang, X. Cheng, et al., Desalination 516 (2021) 115243. doi: 10.1016/j.desal.2021.115243

  18. [18]

      Q.K. Hong, K.M. Wang, C.H. Wu, et al., J. Environ. Chem. Eng. 11 (2023) 109436. doi: 10.1016/j.jece.2023.109436

  19. [19]

      G.U. Semblante, S.D.R. Tampubolon, S.J. You, et al., J. Membr. Sci. 435 (2013) 62–70. doi: 10.1016/j.memsci.2013.02.003

  20. [20]

      W.X. Zhang, F. Jiang, Water Res. 157 (2019) 445–453. doi: 10.1016/j.watres.2018.07.069

  21. [21]

      X. Tang, A. Ding, W. Pronk, et al., J. Cleaner Prod. 197 (2018) 721–731. doi: 10.1016/j.jclepro.2018.06.155

  22. [22]

      H. Wang, Z. Chen, J. Miao, et al., Bioresour. Technol. 209 (2016) 318–325. doi: 10.1016/j.biortech.2016.03.010

  23. [23]

      R. Chen, H. Zhang, J. Wang, et al., Water Res. 224 (2022) 119111. doi: 10.1016/j.watres.2022.119111

  24. [24]

      R. Chen, H. Zhang, D. Lin, et al., Resour., Conserv. Recycl. 190 (2023) 106839. doi: 10.1016/j.resconrec.2022.106839

  25. [25]

      S.J. Ballinger, I.M. Head, T.P. Curtis, et al., Water Sci. Technol. 46 (2002) 543–550. doi: 10.2166/wst.2002.0532

  26. [26]

      R. Chen, H. Liang, J.L. Wang, et al., Sci. Total Environ. 771 (2021) 145372. doi: 10.1016/j.scitotenv.2021.145372

  27. [27]

      T. Cavalier-Smith, Int. J. Syst. Evol. Microbiol. 52 (2002) 297–354. doi: 10.1099/00207713-52-2-297

  28. [28]

      N. Derlon, N. Koch, B. Eugster, et al., Water Res. 47 (2013) 2085–2095. doi: 10.1016/j.watres.2013.01.033

  29. [29]

      B. Ma, S. Wang, Z. Li, et al., Bioresour. Technol. 225 (2017) 377–385. doi: 10.1016/j.biortech.2016.11.130

• 

  Figure 1  Flux and resistance developments of GDM systems. (a) Flux variation, (b) total resistance, and (c) stable flux. Capital letters represent the significant difference analyzed by post hoc Tukey's test (P < 0.05).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Biomass analysis: (a, b) Polysaccharide and protein in the biocake layer. (c-e) The fluorescent pollutant of EPS and SMP in GDM1, GDM2, and GDM3, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Morphological characterization. (a-c) SEM of GDM1, GDM2, and GDM3. (d-f) AFM 2D image of GDM1, GDM2, and GDM3. (g-i) AFM 3D image of GDM1, GDM2, and GDM3. (j) Surface EDS of GDM3, (k) cross-section EDS of GDM3.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Microbial community composition: (a) bacterial phylum, (b) eukaryotic phylum, (c) bacterial genus, and (d) eukaryotic genus.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Mechanism of performance improvement in the magnetic particles assisted GDM system.

  下载: 全尺寸图片 幻灯片
•