Introduction

The nitrification–denitrification process is fundamental to traditional nitrogen removal in biological wastewater treatment. Nitrification involves two consecutive reactions: the oxidation of ammonia to nitrite by ammonia-oxidizing bacteria (AOB), followed by the oxidation of nitrite to nitrate by nitrite-oxidizing bacteria (NOB). Partial nitrification conveniently produces nitrite as an intermediate compound for denitrification, which reduces the oxygen supply and organic carbon for denitrification. Because AOB and NOB coexist in a nitrification system, it is difficult to achieve partial nitrification. The control of experimental parameters is necessary to inactivate or remove NOB and to favor the growth of AOB. Several factors were reported to be important for obtaining partial nitrification, such as the solids retention time (SRT), hydraulic retention time (HRT), dissolved oxygen (DO), pH, temperature, free ammonia, etc. (Zhang et al. 2014; Choi and Ahn 2014).

AOB are slow-growing and low-yield organisms; thus, the microorganisms are easy to wash out from an activated sludge system. By immobilizing the microorganisms on surfaces of carriers, the process can prevent loss of biomass, hence increasing the treatment efficiency (Li et al. 2009; Ras et al. 2011). Biofilms are structurally complex polymeric network with pores and channels which multicellular communities of microorganisms embedded in a matrix of extracellular polymeric substances (EPS) (Wagner et al. 2009; Wang et al. 2014). The EPS are produced by the secretions of the microbial population, cell lysis, and hydrolysis products, including proteins (PN), polysaccharides (PS), nucleic acids, phospholipids, and various heteropolymers. The composition, distribution, and properties of the EPS (solubility and bindability) determine many of the important properties of the biofilms such as their strength, elasticity, sorption capacity for adsorbates, and stratification of the biofilms (Li et al. 2008; Moryl et al. 2014).

The properties and architecture of biofilms can vary with the different strains of microbial species, growth stages, and changes in environmental conditions, such as nutrient concentration, DO, ionic strength, pH, SRT, and HRT (Chen et al. 2013; Wang et al. 2014). It was reported that the physical and chemical states of the EPS were affected by wastewater compositions, and the amount, composition of EPS, and distribution of EPS were influenced by the carbon sources (Wang et al. 2014; Moryl et al. 2014).

As autotrophic bacteria, nitrifiers (AOB and NOB) could obtain energy from the oxidation of ammonium/nitrite for growth without organic carbon involved during biological nitrification process (Ma et al. 2013). In the presence of organic compounds, nitrifiers are usually outcompeted by much faster-growing heterotrophs for oxygen and nutrients, and the competition between heterotrophic and nitrifying bacteria for oxygen and space in biofilm systems is enhanced compare to sludge nitrification system (Oyanedel-Craver et al. 2009). Consequently, the presence of organic carbon may lead to a decrease in nitrification efficiency or even a failure of the process. The forming and development of nitrifying biofilm on the surface of carrier materials can be a very slow process. However, faster-growing heterotrophic bacteria can produce higher production of EPS which enhance adhesion and retain nitrifying bacteria in the nitrification system (Furumai & Rittmann 1994; Tsuneda et al. 2001). Therefore, it is suggested that the presence of moderate amount of organic carbon source is advantageous for nitrification process and development of nitrifying biofilms.

At present, there have been very limited studies on the characteristics of partial nitrification biofilms and the effect of organic carbon on the partial nitrification and structure of biofilms. In this work, the differences in the characteristics of partial nitrification under autotrophic and heterotrophic conditions were compared, the effects of organic carbon on the production of EPS in biofilms and their composition were investigated by chemical analysis and three-dimensional excitation–emission matrix fluorescence spectroscopy (EEM), and fluorescence staining and confocal laser scanning microscope (CLSM) were used to determine differences in the architecture and EPS distribution of the autotrophic and heterotrophic biofilms.

Materials and methods

Biofilm reactors

The study was carried out in two identical laboratory-scale fixed-bed biofilm reactors (FBBR), which were made of Perspex with a working volume of 5.5 L in each reactor. The carrier material consisted of hollow polyhedral polyethylene balls. Each polyethylene ball had an average diameter of 25 mm, a specific surface area of 460 m2 m−3, and a density of 820 g L−1. Air was supplied from the bottom of the reactors.

Feeding media and operational strategy

Each reactor was inoculated with activated sludge from the Sibao Municipal Wastewater Plant in Hangzhou, China. The seed sludge mixed liquid suspended solid (MLSS) concentration was 15.4 g L−1, and the SVI was 60.7 mL g−1. The two bioreactors were provided with synthetic wastewater containing NH4Cl as the main substrate along with minerals. NaHCO3 was used as the inorganic carbon resource for the autotrophic reactor (R1), while glucose was added to the heterotrophic reactor (R2) as an organic carbon resource at a ratio of chemical oxygen demand (COD)/N = 2.6:1. The composition of the synthetic wastewater is as follows (g L−1): NaCl 0.585, KCl 0.074, MgSO4·7H2O 0.049, CaC12 0.128, KH2PO4 0.054, NH4Cl 0.19–1.2, NaHCO3 0.6–3.7, and glucose 0.12–0.78. A solution of trace elements was added at 1 ml L−1. The trace element concentration in this solution was as follows (g L−1): ZnSO4·7H2O 0.43, H3BO3 0.0111, CuSO4·5H2O 0.25, NiCl2·6H2O 0.19, MnCl2·4H2O 0.99, NaMoO4·2H2O 0.22, CoCl2·6H2O 0.24, and Na2EDTA·2H2O 1.5. The reactors were operated at pH 8.0 ± 0.2 using NaHCO3, and the temperature was maintained at 30 ± 1 °C. The HRT values were controlled at 8–10 h by adjusting the influent flow. The DO concentration was maintained at 0.5–1.0 mg L−1. A stepwise increase in the NH4 +-N concentration was performed from 50 to 310 mg L−1 in increments of 20 mg L−1.

Analysis of water quality

NH4 +-N, NO2 −-N, NO3 −-N, COD, and alkalinity were regularly measured using standard methods (APHA 1998). The total COD removal, ammonia removal, and partial nitrification rate were evaluated by comparing the inlet and outlet values of COD, ammonium, nitrate, and nitrite concentrations. The DO, pH, ORP, and temperature of the two reactors were continuously measured online during the performance using a pH/ORP analyzer (model P33, HACH, USA), and DO was measured using a dissolved oxygen sensor (model 5790000, HACH, USA).

EPS characterization

EPS extraction protocol

EPS were extracted according to the previously described protocol that was validated on activated sludge or biofilm samples (Liang et al. 2010; Abzac et al. 2010; Lee et al. 2013). Two biofilm samples were obtained from the hollow polyhedral polyethylene balls of the two reactors using a razor blade. Approximately 1.0 g (dry weight) of each biofilm was processed into a 10-mL suspension with the addition of Milli-Q water, and each biofilm sample was processed in triplicate for EPS fractionation. The suspension was pretreated using ultrasound (20 kHz, 40 W, 30 s, 4 °C) and centrifugation (2,000×g, 15 min, 4 °C). The supernatant liquor was collected as soluble EPS. The sediments were re-suspended with Milli-Q water to 10 mL, and 0.06 mL formamide (37 %) was added to the suspension followed by stirring with an orbital shaker (20–30 rpm, 1 h). The suspension was centrifuged (5,000×g, 15 min, 4 °C), and the supernatant was collected as loosely bound EPS (LB-EPS). The collected sediments were re-suspended again with PBS (pH 7) to obtain another 10-mL suspension. Tightly bound EPS (TB-EPS) were then extracted by chemical protocols: formaldehyde plus NaOH (1 N, 3 h, 4 °C). Biofilm samples were centrifuged (20,000×g; 15 min, 4 °C), and the pellets were washed twice in phosphate-buffered saline (PBS), pH 7. Each biofilm sample was subdivided into three 10-mL aliquots containing approximately 5 g VSS L−1 for triplicate extractions. All EPS fraction samples were filtered before measurement through polyvinylidene fluoride membranes with a pore size of 0.22 μm.

EEM analysis of EPS

EEM spectra of the EPS samples from the different methods were obtained using 1 mL sample volumes. All three-dimensional EEM spectra were measured using luminescence spectrometry (F-4500 FL Spectrophotometer, HITACHI, Japan). The EEM spectra were collected by scanning the samples with scanning emission spectra from 250 to 600 nm at 5-nm increments by varying the excitation wavelength from 250 to 500 nm at 5-nm sampling intervals. Excitation and emission slits were maintained at 5.0 nm, and the scanning speed was maintained at 1,200 nm min−1. The spectrum of deionized Milli-Q water (Millipore, USA) was used as the blank. The software Origin8.0 (OriginLab Inc., USA) was used to process the EEM data, which illustrates the elliptical shape of contours where the X-axis represents the emission spectra and the Y-axis represents the excitation spectra.

Biofilm fluorescence staining and confocal laser scanning microscopy analysis

To visualize the distribution of the cells and EPS of the biofilms, the biofilm samples were stained with different fluorescently labeled probes, and confocal laser scanning microscopy (CLSM; LSM780, Zeiss, Germany) was employed to probe the internal structure of the biofilms.

The biofilm samples were taken by a razor blade and placed on two glass slides and then stained using the scheme proposed by Chen et al. (2007). Proteins were stained with fluorescein isothiocyanate (FITC). Concanavalin tetramethylrhodamine conjugate (ConA) was used to bind to α-mannopyranosyl and α-glucopyranosyl sugar residues. Whole cells were probed with the cell wall-permeable nucleic acid stain SYTO 63. The dead cells were stained by the cell wall-impermeable stain SYTOX blue. Nile red was applied to stain lipids, and calcofluor white was utilized to stain β-polysaccharides. FITC, calcofluor white, and Concanavalin A conjugated with tetramethylrhodamine were purchased from Sigma (St. Louis, USA). SYTO 63, SYTOX blue, and Nile red were used from Molecular Probes (Carlsbad, CA, USA). The probes were detected on six channels with corresponding excitations and emissions according to the report of Chen et al. (2007): FITC (Ex 488 nm, Em 499–552 nm), ConA (Ex 561 nm, Em 561–613 nm), SYTO 63 (Ex 633 nm, Em 650–704 nm), SYTOX blue (Ex 458 nm, Em 459–505 nm), Nile red (Ex 514 nm, Em 622–692 nm), and calcofluor white (Ex 405 nm, Em 411–482 nm).

After staining, the biofilms were analyzed immediately using CLSM. Six fields of view were chosen randomly on the biofilms and viewed using a ×20 objective. The staining and viewing processes were performed twice to reduce deviation. Three-dimensional fluorescent images of the biofilms were obtained, and the thicknesses of the biofilms were measured using the ZEN 2010 Viewer software (Zeiss).

Results

Start-up of the autotrophic and heterotrophic partial nitrification

The two biofilm reactors were operated for 6 months under autotrophic (R1) and heterotrophic (R2) conditions. The partial nitrification operating results of R1 and R2 are illustrated in Fig. 1a, b, respectively. The influent ammonia concentrations were increased stepwise from 50 to 310 mg N L−1. For R1, the ammonia removal efficiency increased from 33 to 65 % after 10 days of operation when the influent ammonia concentration was 50 mg N L−1. Subsequently, the ammonia removal efficiency gradually increased along with the increase in the ammonia load, and an ammonium removal efficiency of over 85 % was obtained at the end of this experiment. The nitrosation rate (NO2 −-N/(NO2-N+NO3-N)) remained at approximately 80 %, and the effluent nitrate concentration was under 40 mg L−1. The ammonia removal efficiency and nitrosation rate reached their highest values at the highest influent ammonia concentration. The reason is speculated as follows: (1) a high concentration of ammonia induces a high concentration of free ammonia, which inhibits the growth of NOB, subsequently resulting in an improved nitrosation rate (Bassin et al. 2012) and (2) a low air flow rate was required to maintain an approximately 1 mg L−1 DO concentration, which caused uneven stirring of the solution by aeration, thus leading to an uneven DO distribution in the reactor. The DO was higher than 1 mg L−1 in some zones, which favored the growth of NOB (Wei et al. 2014).

Fig. 1
figure 1

Nitrogen conversation in the autotrophic (a) and heterotrophic (b) biofilm reactors during partial nitrification domestication

Full size image

R2 started up with a substrate COD loading between 127 and 826 mg L−1 during the entire stage, maintaining the influent C/N ratio of 1.0 (COD/N = 2.6). Figure 1b shows the partial nitrification domestication of R2. The ammonia removal efficiency varied from 79 to 95 % when the ammonia loading was 90 mg L−1 and remained in a stable phase till the end of the experiment. Despite some fluctuations caused by the change in the COD and ammonia concentrations, the nitrosation rate remained stable, varying between 80 and 99 % during the subsequent phases. Both the ammonia removal efficiency and the nitrosation rate in R2 were higher than those in R1, and it took a shorter time for R2 to reach the highest ammonium removal efficiency and nitrosation rate. Moreover, a higher biomass concentration was observed in R2 compared to R1, which is in accordance with the thicker biofilms on the carriers. When the biomass was accumulated in a relatively short period of time, more nitrifying bacteria were colonized on the biofilms, which led to full ammonia removal (Bassin et al. 2012).

During the experiment process, the total nitrogen loss in R2 ranged between 5 and 37 %, which was higher than it was in R1 (0.2–16 %), and the effluent nitrate ranged between 0.4 and 37.6 mg L−1, which was a lower level compared with R1 (7.1–48.3 mg L−1), indicating the presence of denitrification in R2. However, the existence of nitrate implied that heterotrophs in the R2 system remained at a relatively low level, which prevented the system from a full denitrification (Ge et al. 2012). The higher nitrosation in R2 also could be attributed to a lower homogeneous DO (<1.0 mg L−1) distribution that came from the high amount of aeration disturbing the bulk solution well. On the other hand, a variety of heterotrophic bacteria have been reported to be capable of nitrification, and most of heterotrophic nitrifiers also perform denitrification aerobically. The existence of heterotrophic nitrifiers might also improve the ammonia removal efficiency and the nitrosation rate under heterotrophic condition. However, the nitrification rates of heterotrophic nitrifiers are suggested lower than autotrophic nitrifiers under the conditions suitable for autotrophic nitrifiers (Otte et al. 1999). Besides, considering the low nitrogen loss in R2, the heterotrophic nitrification-aerobic denitrification process in R2 system is speculated only a small percentage compare with autotrophic nitrification.

EEM fluorescence spectra analysis of the EPS fractions

Three-dimensional EEM fluorescence spectroscopy has been utilized to identify the chemical composition of EPS. Each EEM fluorescence spectrum gives information about the chemical composition and structure of EPS samples (Domínguez et al. 2010). The EEM fluorescence spectra of the soluble EPS, LB-EPS, and TB-EPS fractions of R1 and R2 biofilms are illustrated in Fig. 2. Four peaks were identified for the soluble EPS extracted from both types of biofilms (Fig. 2a, b). Peak A is related to fulvic acid-like substances, and peak C is related to humic acid-like substances according to Chen et al. (2003). Peak B is assigned to the reported tyrosine & protein-like substances. Peak T is related to tryptophan & protein-like substances. Four peaks appeared in the following locations for LB-EPS of the R1 biofilms (Fig. 2c). Peak A and peak C are located in the region of humic acid-like, peak A is associate to fulvic acid-like substance, peak C is related to humic acid-like substance. Peak B is located in region IV (soluble microbial by-product (SMP-like)) and corresponding to tyrosine and protein-like. Peak T is also located in region IV and corresponding to protein-like contains tryptophan. Five peaks were observed in the following locations for LB-EPS of the R2 biofilm (Fig. 2d). Peak A is assigned to fulvic acid-like substance, and peak C is related to humic acid-like polymers. Peak T1 is located in region IV (soluble microbial by-product (SMP)-like substances) and is related to tryptophan & protein-like substances. Peak T2 is located in the region of II (aromatic protein) and corresponds to tryptophan. Peak B is located in region IV and corresponds to tyrosine & protein-like substances (Chen et al. 2003).

Fig. 2
figure 2

EEM fluorescence spectral identifications of EPS. a Soluble EPS of R1. b Soluble EPS of R2. c LB-EPS of R1. d LB-EPS of R2. e TB-EPS of R1. f TB-EPS of R2

Full size image

Six peaks were identified for the TB-EPS of both R1 and R2 fraction extracted by the formaldehyde plus NaOH (F + N) method (Fig. 2e, f). Peaks B, T1, and T2 are attributed to the protein fluorescence in which peak B is reported as tyrosine protein-like substances, peak T1 is associated with the tryptophan protein-like substances, and peak T2 is located in region II (aromatic protein) and is associated with compounds derived from proteins. Peaks C1, C2, and C3 are related to humic acid-like substances. As observed, part of peak T1 extends to the region of V (humic acid-like) for both TB-EPS; it can be inferred that the existence of protein encapsulation by humic substances in EPS results in overlap of the two peaks.

The fluorescence intensity and peak intensity ratios of the EEM spectra could be employed for quantitative analysis (Liu et al. 2011; Domínguez et al. 2010). The peak intensity ratios are used to analyze the differences in the composition of the EPS extracted from the two biofilms in this research as the concentrations of the two types of EPS are different. The peak intensity ratios are shown as ratios to Peak C, as this component is present in all EPS samples. For the soluble EPS, the intensity ratios (R1, R2) are as follows: A/C 1.3 and 0.5, T/C 3.8 and 4.1, and B/C 2.5 and 1.5. For LB-EPS, the intensity ratios (R1, R2) are as follows: A/C 1.0 and 1.0, B/C 15.0 and 2.3, T/C 3.3 (R1), T1/C 5.2 (R2), and T2/C (R2) 1.0. For TB-EPS extracted by F + N, T1/C1 4.0, 1.9; T2/C1 0.5, 0.3; T1/C2 8.1, 12.3; T2/C2 1.0, 1.7; T1/C3 8.1, 12.3; and T2/C3 1.0, 1.7. The ratios of T/C are greater than the A/C and B/C ratios for soluble EPS of both R1-EPS and R2-EPS, suggesting that tryptophan-like protein presents the highest concentration among the compositions of soluble EPS and that tyrosine-like protein is the second-highest concentration. The concentration of fulvic acid-like substances is higher than that of humic acid-like substances for R1-EPS. However, for R2-EPS, fulvic acid-like substances are only half the quantity of humic acid-like substances according to the intensity ratios. For LB-EPS of the R1 extraction, tyrosine shows the highest concentration, and tryptophan shows the second highest. Fulvic acid-like substances show the same quantity as humic acid-like substances. For LB-EPS of R2 extraction, tryptophan and protein-like substances are the highest and tyrosine and protein-like substances are the second highest. Tryptophan, fulvic acid-like substances, and model humic acid polymers have the same quantities. For TB-EPS extracted by the F + N method, the composition characteristics are similar between the R1 and R2 extractions, i.e., the tryptophan protein-like substances are the highest composition, and aromatic protein–proteins and humic acid-like substances related to peaks C2 and C3 have the same concentration. The intensity ratios also indicated that, in spite of the differences in the concentration of corresponding compositions, the concentration ratios between each composition are similar in the two EPS samples.

CLSM image analysis: the structural characteristics and EPS distribution of biofilms

Fluorescence staining images of R1-biofilm

Figure 3 shows the component distribution of R1-biofilm cells by two-dimensional CLSM image, suggesting that protein (Fig. 3b, green), lipid (Fig. 3c, yellow), α-d-glucopyranose polysaccharides (Fig. 3d, cyan blue), and nucleic acids (Fig. 3e, red) were the main components of the partial nitrification cells, and β-polysaccharides (Fig. 3g) were mainly a part of EPS. Protein composes the outer wall of cells, and α-d-glucopyranose polysaccharides, lipids, and nucleic acids were entrapped within the cells; thus, the colors representing lipid, α-d-glucopyranose polysaccharides, and intracellular nucleic acids were not distinct in the combined image of Fig. 3a–g (Fig. 3h).

Fig. 3
figure 3

Two-dimensional confocal laser scanning microscopy (CLSM) image showing the components distribution of R1-biofilm cells (scale bar = 20 μm). a Optical microscopy photograph. b proteins (FITC). c lipids (Nile red). d α-d-Glucopyranose polysaccharides (ConA). e Total cells (SYTO 63). f Dead cells (SYTOX blue). g β-Polysaccharides (calcofluor white). h Combined image of (a)–(g)

Full size image

The views for the three-dimensional CLSM image were determined to be mature biofilms based on the EPS having a higher proportion of DNA and a lower PN/PS ratio compared to the other samples (Mahendran et al. 2012). The R1-biofilm CLSM images (Fig. 4) revealed that β-polysaccharides covered the outer surface of the biofilm and were distributed throughout the whole biofilm as the predominant EPS content. According to the relevant fluorescent intensity data, β-polysaccharides present the highest proportion among the EPS contents. The microbial cells of the R1-biofilm are mainly located on the interface of the biofilm and the bulk phase (opposite the arrows near the sides of Fig. 4a). The protein was mainly distributed in the upper-middle portion of the R1-biofilm, which played the role of a core. The dead cells and α-d-glucopyranose polysaccharides have the same distributed features as protein. The R1-biofilms contained a low amount of lipids compared to the other components, and these were mainly distributed in the upper portion of the R1-biofilm. The nucleic acids (total cells) were distributed throughout the R1-biofilm but were concentrated in the upper portion near the bulk phase.

Fig. 4
figure 4

Three-dimensional CLSM image of the structural characteristics and EPS distribution of R1-biofilm. a CLSM projection and orthogonal images of biofilm (the plane showing the interface of biofilms and carriers) (scale bar = 50 μm). b Interface of biofilm and bulk phase. c Intact of three-dimensional CLSM image (interface of biofilm and bulk phase is upward). d Intact of three-dimensional CLSM image (interface of biofilm and carrier is upward)

Full size image

Images derived by CLSM for the biofilm samples showed the thickness of the R1-biofilm that was presented (Fig. 4c, d). The thickness of R1-biofilm samples ranged from 40 to 120 μm. Voids and uneven thickness are illustrated in Fig. 4a, and in the xy and xz projections, voids were open spaces within the biofilms matrix that transport nutrients and oxygen into the depths of the biofilms. The heterogeneous biofilm layer revealed the complex forming process of the biofilms in that the microbial film tends to be more concentrated in certain regions of the support material (Bassin et al. 2012). In addition, biofilm detachment caused by abrasion, erosion, and sloughing is considered to be a mechanism of heterogeneous biofilm. The relatively low microorganism count can be observed in Fig. 4, which revealed a limited rate of microorganism growth for autotrophic biofilms (Han et al. 2005).

Fluorescence staining images of R2-biofilms

Figure 5a–h shows the component distribution of R2-biofilm cells by 2-dimensional CLSM image. Abundant filamentous bacteria were present in the R2-biofilm and stained with FITC (Fig. 5b). Extracellular nucleic acid in the R2-biofilm was less than that of the R1-biofilm. Similar to the R1-biofilm, β-polysaccharides (Fig. 5g) were the main component of the EPS rather than a composition of cells (Fig. 5e, f), and protein (Fig. 5b), lipid (Fig. 5c), and α-d-glucopyranose polysaccharides (Fig. 5d) have the same distribution characteristics on the cells.

Fig. 5
figure 5

Two-dimensional CLSM image showing the components distribution of R2-biofilm cells (scale bar = 20 μm). a Optical microscopy photograph. b Proteins (FITC). c Lipids (Nile red). d α-d-Glucopyranose polysaccharides (ConA). e Total cells (SYTO 63). f Dead cells (SYTOX blue). g β-Polysaccharides (calcofluor white). h Combined image of (a)–(g)

Full size image

The distribution profile of the EPS with biofilm depth was recorded and is shown in Fig. 6. β-polysaccharides are the predominant EPS component on the surface of the R2-biofilm (Fig. 6); on the other hand, α-d-glucopyranose polysaccharides were more evident on the interface of the biofilm and bulk phase than they are on the R1-biofilm. Large quantities of cells crowded in the upper layer of the biofilms and are embedded in the β-polysaccharides matrix (Fig. 6a, b). Based on the fluorescence intensity data, β-polysaccharides were evenly distributed within the biofilms with the higher concentration. Lipids, proteins, α-d-glucopyranose polysaccharides, and nucleic acids were present both within cells and the EPS matrix, for which the concentration increased along with increasing depth from the upper surface to the base layer, but they were barely seen in the layer of the interface between the biofilm and carrier. Apart from the similar distributive characteristics of the four analyzed components, α-d-glucopyranose polysaccharides have the highest concentration on the upper surface of the R2-biofilm compared with the others. Dead cells were barely seen in the upper surface and accumulated in the base portion near the carrier (as shown in the xy and xz projections of Fig. 6a, carrier sides opposite the arrows), suggesting that the R2-biofilm layers were formed on the carrier in chronological order. Therefore, all EPS compositions were distributed more evenly throughout the R2-biofilm with denser concentrations compared with the R1-biofilm.

Fig. 6
figure 6

Three-dimensional CLSM image showing the structural characteristics and EPS distribution of R2-biofilm. a CLSM projection and orthogonal images of biofilm (the plane showing the interface of biofilm and bulk phase) (scale bar = 50 μm). b Interface of biofilm and carrier. c Intact of three-dimensional CLSM image (interface of biofilm and bulk phase is upward). d Intact of three-dimensional CLSM image (interface of biofilm and carrier is upward)

Full size image

The R2-biofilm thickness was recorded for several fields of view, and one of them is shown in Fig. 6c, d. The average thickness ranged from 40 to 250 μm. In contrast to the CLSM results of the R1-biofilm, the R2-biofilms showed more homogeneous and more compact structure, which is consistent with the research of (Bassin et al. 2012). The biomass was evenly distributed throughout the biofilms when the highest organic load was applied to the nitrification systems.

Discussion

Performance of partial nitrification

The characteristic of autotrophic partial nitrification was different than that of heterotrophic partial nitrification in this research, and partial nitrification with the COD/N ratio of 2.6:1 in R2 appeared to obtain better ammonia removal efficiency and nitrosation rate. Under autotrophic conditions, an ammonia removal efficiency of 80–89 % and nitrosation rate of 75–86 % were achieved when the ammonia load was at a high range of 250 to 310 mgN L−1. Both the ammonia removal efficiency and nitrosation rate increased with an increasing load of ammonia. Compare with autotrophic biofilm system, the ammonia removal efficiency and nitrosation rate reached a high level in shorter period of time (<20 days) in the heterotrophic biofilm system. Ammonia removal rate and nitrosation rate increased from 21 to 72 % and 39 to 71 %, respectively, with the ammonia load increased from 50 to 70 mg L−1. When the NH4 +-N load increased to 90 mg L−1, both ammonia removal rate and nitrosation rate were over 80 % and subsequently continued to increase as the treatment process went on. With the increase in NH4 +-N load, the final ammonia removal rate and nitrosation rate were as high as 97 and 99 %, respectively. Most of the ammonia removal rate and ideal nitrosation rate were achieved in a shorter time compared to the autotrophic bioreactor.

It was demonstrated from this study that an appropriate amount of organic carbon resource present in the partial nitrification system is advantageous for obtaining better results even though nitrification is an autotrophic process. Zielińska et al. (2012) proposed that a lower COD/N ratio caused lower COD removal rates and a higher diversity and abundance of AOB. The low COD removal rate was related to the limiting supply of nutrients for the growth of heterotrophs that resulted in low consumption of COD. Bassin et al. (2012) suggested that feeding some organic compounds to the nitrification system is a promising strategy to lower the time necessary for biofilm development without causing a detrimental effect on the nitrification performance and favoring the accumulation of nitrifying bacteria in the biofilms during the start-up phase. However, the high COD/N ratio was reported to be detrimental to the nitrification system. Ma et al. (2013) found that an increase in the COD/N ratio from 3.0 to 10.0 resulted in the proliferation of the overall bacterial communities. However, with a further increase in the COD/N ratio from 10.0 to 60.5, the diversity of the overall bacterial communities was decreased and nitrification declined due to the competition between heterotrophs and autotrophs leading to the washout of the autotrophs. It was suggested that a range of 2 to 3 for the COD/N ratio was beneficial for partial nitrification.

Biochemical properties of partial nitrification biofilms

The fluorescence intensity and peak intensity ratios of EEM spectra indicate that the proportions of protein-like substances and humic substances of the EPS derived from the two biofilms are different. In addition, the locations of fluorescence peaks show the differences in the components in the two different EPS. The compositional analysis of extracted EPS shows that higher quantities of EPS were generated by the heterotrophic biofilm in comparison with the autotrophic biofilm, and the concentrations of each component of EPS derived from the heterotrophic biofilm were all higher compared with the autotrophic biofilm-derived EPS. It was indicated that the presence of organic carbon induced a greater production of EPS, which is in agreement with the report of Bassin et al. (2012). It was suggested in the literature that environmental conditions such as the C/N ratio of the influent, hydrodynamic shear and oxygen supply, etc., have an influence on the quantity and composition of EPS (Li et al. 2008; Bridier et al. 2010). For the two partial nitrification systems, the organic carbon was the only different condition, however, to maintain DO of R2 at 0.5–1.0 mg L−1, aeration intensity in R2 was adjusted higher for demand of higher quantity of biomass, which induced higher hydrodynamic shear and subsequently increased the production of polysaccharides.

The PS and PN were dominant in both TB-EPS derived from different biofilms, but PN concentrations were higher than PS in soluble EPS and LB-EPS. Bound EPS play more important roles in the biofilms than the soluble EPS (Li et al. 2008), and the proportions of PN and PS in TB-EPS were reported in some of the literatures (Adav et al. 2007; Wang et al. 2014). In this study, the PN/PS ratios of the autotrophic and heterotrophic biofilms were 0.96 and 0.69, respectively. The PN and PS of TB-EPS are present in approximately equal proportions in R1, but PS concentration shows a nearly twofold higher than PN in R2. The PN/PS ratio in R1 biofilm-derived EPS is in agreement with the result that PN/PS was present in proportions ranging between 1 and 1.7 in biofilm-derived EPS (Mahendran et al. 2012). The concentration of PS in R2 is higher than that of PN, which may be attributed to conversion of the influent glucose (Ras et al. 2011). The higher concentration of DNA in the R2 EPS may came from cell death and decay due to the higher concentration of biomass and rapid biofilm development, which lead to higher levels of cell death (Fagerlind et al. 2012).

Morphologic properties of partial nitrification biofilms

CLSM images show the different structural distributions of EPS and cells between R1 and R2 biofilms. Biofilm structure plays an important role in its functional properties and overall wastewater treatment performance because it determines external mass transfer in biofilm system, and the external mass transfer significantly affects the rates of aerobic carbon oxidation and nitrification (Mahendran et al. 2012). R2 biofilms were thicker compared with R1 biofilms because both organic loading and hydraulic shear could increase biofilm thickness. However, modeling studies have found that complete oxygen penetration occurs when the biofilm is 20 μm or less (Picioreanu et al. 2004), which implies that both R1 and R2 biofilms are partially oxygen penetrated, subsequently resulting in aerobic, anoxic, and anaerobic growth zones present in both biofilms. It is inferred that there are more anoxic and anaerobic growth zones existing in R2 biofilms, and the presence of a small proportion of denitrification in R2 supports this hypothesis. However, the heterogeneous and porous structure of the biofilm induced a complex pattern of oxygen penetration, correspondingly, a complex microbial community structure, and the high partial nitrification efficiency in the R2 treatment system verified the result. Meanwhile, the dead cells mainly distributed on the layer next to the carrier in R2 were likely due to substrate diffusion limitation in thick biofilms (Fagerlind et al. 2012), and dead cells distributed evenly throughout the whole R1 biofilm may be attributed to the homogeneous substrate diffusion through the relatively thin biofilms.

The location of the EPS in biofilms is important, as it has been found to form a complex with extracellular enzymes (Frølund et al. 1995). More EPS (expressed as β-polysaccharides) were generated in the region near the bulk liquid (wastewater) in both biofilms as it was the highly aerobic region, which was consistent with the report of Li et al. (2008). β-Polysaccharides presented a network over the entire biofilms and play a role in connecting cells and stabilizing the structure of biofilms (Mcswain et al. 2005).

The microbial cells of the R1 biofilm mainly distributed in the upper-middle portion near the bulk phase can be explained by the biofilm’s age. Wagner et al. (2009) proposed that the microorganisms located near the bulk phase can be attributed to the better supply of oxygen, glucose, nutrients, and trace elements with increasing age.

Extracellular DNA (eDNA, the main nucleic acid component in the biofilms) is derived from lysed cells, and some are produced by bacteria through a mechanism that involves the release of small vesicles from the outer membrane. eDNA acts as a nutrient store and strengthens the biofilm matrix, as well as provides adhesive support and protection to microbial cells (Grande et al. 2011; Jakubovics et al. 2013). For R2-biofilm, eDNA trends to increase along with depth, which may be attributed to the increase in dead cells or nutrient requirements with depth of thick biofilms.

The results illustrate the similar aspects of the structures and EPS distributions of R1-biofilm and R2-biofilm: (1) polysaccharides and proteins are the major EPS constituents of biofilms, and polysaccharide contents are higher than protein contents from both R1-EPS and R2-EPS; (2) β-polysaccharides and cells (live) concentrate in the outer layers of biofilms, and β-polysaccharides not only are the significant components of EPS in the outer layers but also distribute throughout the biofilms; and (3) proteins, lipids, α-d-glucopyranose polysaccharides, and dead cells are mostly present within the layers of biofilms. There are some delicate distinctions between the two types of biofilms: (1) The total amount of R2-EPS is significantly higher than that of R1-EPS; (2) R2-biofilms are more even and compact in structure compared with R1-biofilm; (3) proteins, lipids, α-d-glucopyranose polysaccharides, and nucleic acids mainly distribute in the upper-middle portion of the R1-biofilm, but for R2-biofilms, the components trend to increase along with depth from the upper surface to the base layer; and (4) for R1-biofilm, dead cells distribute evenly throughout the whole biofilms, whereas they trend to concentrate on the carrier side for R2-biofilm.