Production of unburned calcium silicon filter material (UCSFM) from oyster shell and its performance investigation in an A/O integrated biological aerated filter reactor (A/O-BAF)

Abstract

A considerable amount of the oyster shells as a waste product of mariculture is produced every year, which leads to a major disposal problem with coastal regions of China. This work researched the feasibility of utilizing oyster shell as raw materials to produce unburned calcium silicon filter material (UCSFM) in an unburned process. The chemical composition, physical property and microstructure of UCSFM were determined. The results indicated that oyster shells could be utilized to produce the UCSFM, which belongs to alkaline filter material. The UCSFM was invoked as a bio-carrier in an A/O integrated biological aerated filter reactor (A/O-BAF) for wastewater treatment, and its processing performance was investigated at different ammonia nitrogen loads. The CODcr and NH₃–N removal rates of UCSFM BAF were slightly higher than UCSFM A/O-BAF, but UCSFM A/O-BAF showed a higher TN removal rate of ammonia nitrogen loads between 0.130 kg (m³ d)⁻¹ and 0.496 kg (m³ d)⁻¹ and the average TN removal rates of UCSFM A/O-BAF and UCSFM BAF were 41.88% and 7.48%, respectively. Pyrosequencing analysis of the 16S rRNA gene in anoxic/oxic zones of A/O-BAF process indicated that the microbial communities were quite different from different ammonia nitrogen loads. Thiobacillus, Thiothrix, Hydrogenophaga, Flavobacterium were the dominant genera shared in the oxic zones, while VadinHA17_norank, Desulfobacter, Aeromonas and SB-1_norank were the dominant genera shared in the anoxic zones. The microbial population and physicochemical properties of the UCSFMs explained the excellent performance of the UCSFM A/O-BAF reactor well. Therefore, the UCSFM produced from the oyster shell was appropriate for serving as a bio-carrier in A/O-BAF reactors.

---

1. Introduction

Oyster shell (OS) is a form of waste product that is produced by taking shellfish as a principal part of the food in China. A large amount of oyster shells are produced every year. As for Shandong Province, shellfish farms cover 63 000 ha of coastal oceans and produce approximately 704 900 kg of OS every year.¹ Oyster shell waste brought about a great deal of social and environmental problems in coastal regions of China. For example, oyster shell waste can give off bad odors and contaminates the surrounding environment when it is dumped into coastal water or reclaimed land. Utilization and development of oyster shell waste is still in its primary stages. Waste oyster shells have special components and contain a large amount of calcium, micro elements and various amino-acids, and therefore they are often used as agriculture fertilizers, construction materials and chicken feed. All these usages are restricted. Oyster shell waste has caused a widespread concern because of its social and environmental problems, particularly in China. Therefore, trying to find a feasible way to balance the environment and economy is an imminent problem of mariculture areas.

Biological aerated filters (BAFs), as novel, flexible and effective bioreactors, are applied widely all over the world as an attached biomass process, which is associated with superior economical upgrading of the technology.^2–4 They could remove chemical oxygen demand (COD), suspended solids (SS) and ammonia nitrogen (NH₃–N) in a single unit,⁵ but as the denitrification of nitrates in BAF did not have anoxic stage, it was not effective in removing the total nitrogen (TN). In order to remove SS, COD, NH₃–N, and TN, modified BAFs were designed by injecting air at the midway of the column in an upflow mode at the bottom of the reactor.⁶ The modified BAFs consisted of the anoxic zone (A zone) and oxic zone (O zone). The effluent from O zone was recycled into A zone to finish nitrogen removal. Two different functional zones were integrated into a single unit that made A/O BAFs relatively compact compared with the activated sludge process. Filter media plays a critical role in the design and operation of the process. The characteristics of filter media not only affect the initial capital outlay, operation mode and process design of BAFs, but are also associated with the daily running cost, such as air influx and backwashing.^7,8 The fact that some waste materials, such as clinoptilolite tailings,⁵ polyethylene plastic,⁹ fly ash¹⁰ and grain-slag,¹¹ have been successfully used as filter media for the biological aerobic filter provides us a new approach to dispose off the oyster shell waste, which could obtain a “win–win” result by treating wastewater with waste.

In this study, unburned calcium silicon filter material (UCSFM) was developed with the principal raw material of waste oyster shell. The objective of this work is to investigate the feasibility of converting oyster shell waste into the UCSFM by an unburned process, and examining the performance of the UCSFM serving as the bio-carrier in the A/O-BAF reactor. The result of this work may lead to an alternative waste oyster shell utilization process.

2. Materials and methods

2.1 Materials

UCSFM was comprised of oyster shells, zeolite tailing, pore forming agent-sodium salt and cement. OS was collected from an oyster farm near Qingdao City. Zeolite tailings were taken from Weifang Trading Company Ltd. Shandong Province of China and cement was obtained from Shandong Shanshui Cement Group Ltd. The chemical constituents of the mixtures of oyster shells, zeolite tailing and cement are displayed in Table 2. Zeolite tailing and OS were to serve as the major raw materials. Pore forming material – sodium salt and cement was invoked as addition agents. The four raw materials produce a new type of filter material-UCSFM. The UCSFM were composed of oyster shells, zeolite tailings, sodium salt and cement with a mass ratio of 8 : 3 : 2 : 2. Haydites were only composed of clay.

The preparation process of the UCSFM was shown in Fig. 1. UCSFM was produced as follows: first, zeolite tailings and OS were warmed to 115 °C for 1.5 h to eliminate water and volatile organic compounds, and then were crushed into powder (Wiley Mills, USA) using a grinder and sieved with 100 meshes. Next, these mixtures were blended in a blender mixer, and tap water was then injected into a rotational disk to pelletize these mixtures in 3 mm sizes. Finally, semi-finished products were diverted to cement curing box with a constant temperature and humidity for 7 days. The UCSFM was obtained after washing.

Fig. 1 Illustration for the UCSFM preparation process.

2.2 Reactor description and operation

2.2.1 Reactor description. Two lab-scale reactors were composed of organic glass (Fig. 2). The reactors were designed to be 2.50 m in height and 0.08 m in inner diameter in an upflow mode. The height of the filter material was 1.20 m with an empty bed volume of 6.03 L. To create an anoxic zone (A zone), air was introduced midway of the A/O-BAF reactor using an air diffuser, while air was introduced into the bottom of BAF reactor. BAF was run as control reactors and A/O-BAF as a test reactor. The air flow rate and wastewater rate was pumped into reactors with an air flow meter, and a peristaltic pump, respectively. To remove the accumulated suspended substance and shedding of biofilm, backwashing was carried out every 48 h.

Fig. 2 Experimental schematic view of the water treatment process.

2.2.2 Wastewater characteristics. The influent water of A/O-BAF was the effluent from the primary settling tanks of the sewage water station of University of Jinan in Shandong, China. Table 1 shows the influent water characteristics.

Table 1 Characteristics of the test wastewater

Parameter	Range	Parameter	Range
CODcr (mg L^−1)	100.23–600.87	pH value	6.81–7.07
NH3–N (mg L^−1)	20.12–120.24	TP (mg L^−1)	12.12–18.32
Temperature	23−25 °C

---

2.2.3 Start-up. The two reactors were operated with activated sludge from Everbright Water 〈Jinan〉 Limited. The biofilm culturing process is as follows: first, the activated sludge was introduced and aerated without inflow water for 7 days. Then, the two reactors began continuous operation with water flow of 3.0 L h⁻¹, air flow of 9.0 L h⁻¹ and hydraulic retention time (HRT) of 2 h. Finally, the two reactors reached a steady state after approximately 6 weeks with the same operating conditions.

2.2.4 Operation of the experimental BAFs. The whole test mainly consists of two steps. The operation parameters of each step are as follows:

• Stage one (15 days). Ammonia nitrogen loading rates 0.130 kg (m³ d)⁻¹, HRT 3 h, pH 6.8–7.0, and temperature controlled at 23–25 °C. The average concentration of dissolved oxygen (DO) was 3.65, 0.0 and 3.52 mg L⁻¹ in BAF, anoxic and aerobic zones of the A/O-BAF process, respectively.

• Stage two (15 days). Ammonia nitrogen loading rates 0.496 kg (m³ d)⁻¹, HRT 3 h, pH 6.8–7.0 and temperature controlled at 23–25 °C. The average concentration of dissolved oxygen (DO) was 3.65, 0.0 and 3.52 mg L⁻¹ in BAF, anoxic and aerobic zones of the A/O-BAF process, respectively.

2.3 Analytical methods

The chemical composition of raw materials and UCSFM was examined using an X-ray fluorescence spectrometer (XRF, AXIOS-PW4400). The bulk density, particle diameter and hydrochloric acid soluble rate of UCSFM and haydite were checked according to the method reported by Kent et al.⁷ The cylinder pressure strength and crushing strength were examined according to the method reported by Qiu et al.¹² The crystal structures of raw material powder and UCSFM were examined by X-ray diffraction (XRD, D8 FOCUS). The surface and aperture distribution of UCSFM were examined by scanning electron microscopy (SEM, QUANTA 250 FEG), mercury intrusion porosimetry (MIP, ASAP2020), respectively. The concentrations of chemical oxygen demand (CODcr), NH₃–N, TN, and TP were examined according to standard methods.¹³ The pH, temperature and dissolved oxygen were detected by probes.

2.4 Analysis of the microbial community in the biomass from the A/O-BAF process

The biofilm samples were taken out of the anoxic/oxic zones of A/O-BAF process of different ammonia nitrogen loads. The UCSFMs were soaked in sterile water, and the biofilm was then removed from the UCSFMs using a sterile brush. Finally approximately 1 g of wet biofilm was collected into a 10 mL sterile plastic test tube for DNA extraction. The biofilm was collected using clean cotton swabs and placed into different sterile containers. All biofilm specimens were initially kept at 4 °C, and the DNA was isolated within an hour. Before DNA extraction, the cotton swabs were rinsed gently with sterile water to remove the residual biofilm. The DNA was extracted using a DNeasy Tissue kit (OMEGA E.Z.N.A, USA). The DNA concentration of the purified amplicons was checked by 1% agarose gel electrophoresis. The amount of DNA was estimated using a Nano Drop spectrophotometer (Nano Drop Technologies, Wilmington, DE, USA). The dehydrated samples were sent to Shanghai Major Biological Medicine Science and Technology Co Ltd, where the pyrosequencing analyses were performed using an Illumina Miseq PE300 system. Barcodes with 10 bp were used for each sample. The 16S rRNA genes were amplified with the primers for the region: 338F (5′-ACTCCTACGGGAGGCAGCA-3′), in which the tag was included, and 806R (5′-GGACTACHVGGGTWTCTAAT-3′).

3. Results and discussion

3.1 The characteristics of filter materials

3.1.1 Chemical composition. The chemical analyses of raw materials and UCSFM are presented in Table 2. The UCSFM consisted mainly of CaO (47.23%) and SiO₂ (20.41%), followed by Al₂O₃ (4.71%), SO₃ (3.81%), Fe₂O₃ (2.40%) and MgO (1.85%). Moreover, a few transition metals, such as TiO₂, MnO, V₂O₅ and Cr₂O₃, a some alkali metal oxides involving Na₂O, K₂O and SrO (1.15% in total) were also found to co-exist. As shown in Table 2, the concentrations of Al₂O₃, CaO, SiO₂ in UCSFM compared with raw materials were found to decrease by 1.32%, 1.34%, and 3.57%, respectively. The concentration of SiO₂ has the largest change because the fineness of SiO₂ powder is less than that of the other raw material and it is more difficult to stick to the UCSFM. In addition, the largest relative change rate of all components was displayed by Na₂O. Sodium salt is the pore forming material of the preparation of filter material, and therefore Na₂O is lost after the completion of the pore forming stages.

Table 2 Chemical composition of the raw materials and UCSFM (wt%)

Component	Units	Raw materials	UCSFM	Component	Units	Raw materials	UCSFM
Na2O	%	0.69	0.22	TiO2	%	0.26	0.25
MgO	%	2.15	1.85	V2O5	%	0.03	0.06
Al2O3	%	6.03	4.71	MnO	%	0.19	0.23
SiO2	%	23.98	20.41	Cr2O3	%	0.02	0.03
P2O5	%	0.16	0.16	Fe2O3	%	2.15	2.40
SO3	%	4.49	3.81	SrO	%	0.09	0.10
K2O	%	0.83	0.57	Cl	%	0.10	0.12
CaO	%	48.57	47.23	Others	%	10.2	17.8

---

As showed in Fig. 3. UCSFM had complex crystal structures that mainly ascribed to the raw materials' composition. There is a large amount of CaCO₃ and Ca₃SiO₅ (3CaO·SiO₂) and a small amount of SiO₂ in the raw material powder, and the diffraction peaks are higher than that of UCSFM. Moreover, there are also some steamed bread peaks. The diffraction peak of UCSFM is relatively small, and the main crystal structure is of CaCO₃ and SiO₂. In addition, the steamed bread peak that appears in raw material powder was not observed in the UCSFM powder. In view of the abovementioned phenomenon, the reasons were as follows: the steamed bread peaks of raw material powder should be the pore forming crystal structure and the pore forming material had been cleared, therefore it does not appear that those were the same steamed bread peaks. The concentration of SiO₂ in UCSFM is increased. The main reason is that Ca₃SiO₅ (3CaO·SiO₂) was activated in filter material curing process and some of these active SiO₂ were freed, and therefore the SiO₂ diffraction peaks of XRD pattern are increased in the UCSFM powder, while the diffraction peak of Ca₃SiO₅ (3CaO·SiO₂) disappeared.

Fig. 3 X-ray diffraction patterns of the raw material powder (A) and UCSFM (B).

3.1.2 Physical properties. According to light aggregate and its test method (GB/T1743-1998) and water treatment filter material (CJ/T43-2005) measurement method, the volume density, cylinder pressure strength, crushing strength, dissolution rate of hydrochloric acid, and other apparent performances of the material filter were measured. The results are presented in Table 3.

Table 3 Comparison of the apparent property between UCSFM and haydite

Filter material	Particle diameter (mm)	Bulk density (kg m^−3)	Cylinder pressure strength (MPa)	Crushing strength (N)	Hydrochloric acid soluble rate
UCSFM	3–5	680–710	3.82	165–190	42.56%
Haydite	3–5	610–700	3.0 (qualified product)	110–220	2.82%
			4.0 (superior product)

---

Table 3 shows that the density of UCSFM is similar to haydite. When cylinder pressure strength is approximately 3.82 MPa, UCSFM belongs to qualify products. The maximum crushing strength of UCSFM is slightly smaller than that of haydite, but the quality is more uniform than that of haydite. UCSFM is an alkaline filter material and the hydrochloric acid solubility is larger, therefore it cannot be applied to treat acidic wastewater. The experimental results and its chemical composition are closely related.

3.1.3 Microstructure properties of UCSFM. Fig. 4(A) and (B) indicate the microstructure properties of UCSFM. It turned out that UCSFM had rougher surfaces distributed among large pores, and there were plenty of arch structures. In addition, there are several crystal columns in the cross-section of UCSFM. The rough surface distributed over large pores and internal microstructure was propitious for microorganisms to grow onto the UCSFM.

Fig. 4 SEM photographs and aperture distribution of UCSFM: (A) surface; (B) cross-section and (C) pore diameter.

Fig. 4(C) shows the aperture distribution of UCSFM. The pole of UCSFM is concentrated mainly on 0.01 nm and 0.1 nm, but there are some larger apertures between 10 nm and 100 nm or over 100 nm, the largest are 228 nm. The holes of the pore capacity below 0.1 nm are 3.135 cm³ g⁻¹ (contributing 85.48% of the total pore volume). The holes of the pore capacity between 0.1 nm and 10 nm are 0.547 cm³ g⁻¹ (contributing 11.70% of the total pore volume). The holes of the pore capacity between 10 nm and 100 nm are 0.074 cm³ g⁻¹ (contributing 1.63% of the total pore volume). The holes of the pore capacity over 100 nm are 0.029 cm³ g⁻¹ (contributing 1.19% of the total pore volume). According to the abovementioned analysis, there are rich pores and many small pores could rely on large pores, therefore there are good biofilm adhesion and adsorption properties. These characteristics show that UCSFM is fitting for support media in A/O BAF.

3.2 Removal characteristics of UCSFM A/O-BAF at different ammonia nitrogen loads

UCSFM A/O-BAF and BAF were operated for 6 months after the start-up biofilters. During steady-state conditions, the temperature of the wastewater was controlled from 23 °C to 25 °C, pH values between 6.8 and 7.0, and HRT approximately 3 h. In order to meet the demand of microbial growth, the carbon nitrogen ratio was controlled to 5 : 1; therefore, CODcr of the influent also showed a greater change with the improvement of ammonia nitrogen. The CODcr, NH₃–N and TN of influent and effluent in two reactors during stage one and two are shown in Fig. 5–7.

3.2.1 CODcr removal in two reactors. The influent and effluent CODcr at different ammonia nitrogen loads into two reactors are illustrated in Fig. 5. From Fig. 5, it is observed that both the reactors have excellent CODcr removal efficiency. UCSFM BAF had a slightly higher removal efficiency of CODcr than UCSFM A/O-BAF at ammonia nitrogen loads between 0.130 kg (m³ d)⁻¹ and 0.496 kg (m³ d)⁻¹. The average CODcr removals of UCSFM BAF and UCSFM A/O-BAF were 84.64% and 73.09%, respectively.

Fig. 5 Influent and effluent CODcr at different ammonia nitrogen loads in the two reactors.

When the ammonia nitrogen load was 0.130 kg (m³ d)⁻¹, the average effluent concentrations of CODcr in UCSFM A/O-BAF and UCSFM BAF were about 43.18 mg L⁻¹ and 27.03 mg L⁻¹, respectively, and the average CODcr removals were approximately 69.38% and 80.66%, respectively. In the case of 0.496 kg (m³ d)⁻¹, the average removal rate of CODcr in UCSFM A/O-BAF and UCSFM BAF were 76.79% and 88.62%, respectively, and the average effluent concentrations of CODcr were approximately 108.83 mg L⁻¹ and 53.39 mg L⁻¹, respectively. These results suggest that the ammonia nitrogen load could influence the CODcr removal rate.

3.2.2 NH₃–N removal in two reactors. The influent and effluent NH₃–N in the two reactors at different ammonia nitrogen loads are shown in Fig. 6. As shown in Fig. 6, when ammonia nitrogen load was 0.130 kg (m³ d)⁻¹, both reactors processed an excellent removal rate toward NH₃–N, the average concentrations of NH₃–N in the effluent of UCSFM A/O-BAF and UCSFM BAF were approximately 8.26 mg L⁻¹ and 0.25 mg L⁻¹, respectively, and the average NH₃–N removals of these two reactors were approximately 70.15% and 99.10%, respectively. However, the NH₃–N removal rates of both reactors decreased in the case of 0.496 kg (m³ d)⁻¹, and the average effluent NH₃–N in the effluent of UCSFM A/O-BAF and UCSFM BAF were 55.80 mg L⁻¹ and 7.31 mg L⁻¹, respectively, and average removal rates were 48.59% and 93.21%, respectively.

Fig. 6 Influent and effluent NH₃–N at different ammonia nitrogen loads in the two reactors.

The results indicated that both the reactors show a significant reduction for ammonia nitrogen removal rates when the ammonia nitrogen loads range from 0.130 to 0.496 kg (m³ d)⁻¹. The ammonia nitrogen removal rate of UCSFM A/O-BAF decreased by 21.56%. The following reasons can account for these results. First, when the ammonia nitrogen load increases, the CODcr load also increases. Second, a large number of organic matter in the anoxic section of A/O-BAF failed to remove CODcr, so NH₃–N was not nitrified by nitrobacteria because of insufficient biomass. Third, heterotrophic bacteria and autotrophic bacteria competed for nutrients, such as substrates, DO and inhabitation area of the medium in the oxic section of A/O-BAF. This was helpful for heterotrophic bacteria when a higher organic load was induced by increasing the ammonia nitrogen load.¹⁴ As a consequence, autotrophic bacteria were deprived and ammonia nitrogen removal rate decreased rapidly.

3.2.3 TN removal in two reactors. The influent and effluent TN of the two reactors at different ammonia nitrogen loads are presented in Fig. 7. As can be seen from Fig. 7, A/O-BAF had a higher TN removal efficiency than BAF at ammonia nitrogen loads between 0.130 kg (m³ d)⁻¹ and 0.496 kg (m³ d)⁻¹. The average TN removal of A/O-BAF and BAF were 41.88% and 7.48%, respectively. The effluent TN concentrations of A/O-BAF and BAF were 23.78–86.59 mg L⁻¹ (on average, 55.27 mg L⁻¹) and 44.51–117.53 mg L⁻¹ (on average, 81.55 mg L⁻¹), respectively.

Fig. 7 Influent and effluent TN at different ammonia nitrogen loads in the two reactors.

In the experiment performed with the ammonia nitrogen load at 0.130 kg (m³ d)⁻¹, the average effluent TN concentrations of A/O-BAF and BAF were approximately 25.34 mg L⁻¹ and 49.18 mg L⁻¹, respectively, and the average TN removal rates were approximately 50.63% and 4.34%, respectively. In the case of 0.496 kg (m³ d)⁻¹, the average TN were reduced to 85.21 mg L⁻¹ and 113.93 mg L⁻¹, respectively, and the average removal rates were 33.14% and 10.62%, respectively. The results indicated that the ammonia nitrogen load into two reactors has different effects. When the ammonia nitrogen loads ranged from 0.130 to 0.496 kg (m³ d)⁻¹, the TN removal rate of A/O-BAF was reduced by 17.49%, while the TN removal rate of BAF increased by 6.28%.

Overall, UCSFM consisted mainly of CaO (47.23%). It can strip CaCO₃ into the wastewater to buffer the pH and maintain acceptable pH values, and therefore enhanced the ability of TN removal rate and resistance to the ammonia nitrogen load effect.

3.3 The microbial species diversity and community structure of A/O-BAF

Pyrosequencing analysis was performed to reveal the microbial diversity and distribution of anoxic/oxic zones of A/O-BAF process of different ammonia nitrogen loads. After trimming, sorting and quality control, 30289 (AOBFON1), 25009 (AOBFAN1), 25896 (AOBFON2) and 25275 (AOBFON2) high quality sequence tags were clustered into 551 (AOBFON1), 493 (AOBFAN1), 370 (AOBFON2) and 461 (AOBFON2) operational taxonomic units (OTUs) at 3% distance thresholds (Table 4). The abundance-ACE (based coverage estimator), Chao, Simpson and Shannon are also presented in Table 4. Higher numbers of OTUs were estimated at the anoxic zone samples (626 and 626 in AOBFON1, 485 and 476 in AOBFON2, respectively) on the basis of ACE and Chao. This revealed that the richness of the bacterial communities in the anoxic zone of A/O-BAF operated on the ammonia nitrogen load of 0.130 kg (m³ d)⁻¹ was higher than the ammonia nitrogen load of 0.496 kg (m³ d)⁻¹. However, for the oxic zone samples, the richness of the bacterial communities in AOBFAN2 was greater than in AOBFAN1 (Table 4). The Shannon index values of AOBFAN1 at a 3% distance were higher than AOBFON1, respectively, indicating higher diversity of AOBFAN1 compared to AOBFON1, while similar higher diversity was also presented in AOBFAN2 than AOBFON2. The differences in the microbial richness and diversity of the anoxic and oxic zones of A/O-BAF process could be attributed to dissolved oxygen (DO). The Shannon diversity index of AOBFON1 and AOBFAN1 at 3% distance were higher than that of AOBFON2 and AOBFAN2, respectively, indicating higher diversity of AOBFON1 compared with AOBFAN1. The differences in the microbial richness and diversity of A/O-BAF process could be attributed to the different ammonia nitrogen load.

Table 4 Similarity-based OTUs and species richness estimates of the bacterial phylotypes in the four samples

Sample ID	Reads	0.97
		OTU	ACE	Chao	Shannon	Simpson
AOBFON1	30 289	551	626 (601 662)	626 (597 676)	4.37 (4.35, 4.39)	0.0359 (0.0350, 0.0369)
AOBFAN1	25 009	493	561 (537 596)	586 (548 651)	4.52 (4.50, 4.54)	0.0263 (0.0255, 0.0270)
AOBFON2	25 896	370	476 (440 530)	485 (439 563)	3.72 (3.70, 3.74)	0.0572 (0.0558, 0.0586)
AOBFAN2	25 275	461	569 (534 620)	593 (541 679)	4.31 (4.29, 4.33)	0.0300 (0.0293, 0.0306)

---

To further illustrate the distinctions of the microbial communities between the anoxic and oxic zones of A/O-BAF process of different ammonia nitrogen loads, all of the four samples were compared at the genus level. The relative abundance of the most dominant taxa in the A/O-BAF process is shown in Fig. 8.

Fig. 8 Relative abundance of the genera identified in the samples obtained from the anoxic and oxic zones of A/O-BAF process at different ammonia nitrogen loads.

Microbial groups were observed in the oxic zones in A/O-BAF process of different ammonia nitrogen loads in Thiobacillus (13.70%), Thiothrix (8.64%), Saprospiraceae_uncultured (5.70%), Hydrogenophaga (4.95%), Pseudoxanthomonas (4.95%), Flavobacterium (4.25%) in AOBFON1 and Hydrogenophaga (24.91%), Flavobacterium (12.53%), Arcobacter (11.52%), Thiobacillus (8.52%), Acinetobacter species (4.50%), Thiothrix (3.39%) in AOBFON2.

Although both samples had the same dominant genus (Thiobacillus, Thiothrix, Hydrogenophaga, and Flavobacterium), the microbial communities between AOBFON1 and AOBFON2 were significantly distinct on the genus level. Thiobacillus were reported to be the main degraders for thiocyanate and widespread in coking wastewater treatment systems. Meanwhile, they have been extensively applied for denitrification processes of the superior denitrification capacity.^15,16 Thiothrix can reduce nitrite and/or nitrous oxides (denitrification) under anaerobic growth. Hydrogenophaga, an autotrophic H₂-oxidizing bacterium that utilizes hydrogen as an energy source,¹⁷ was found to be accumulated in the AOBFON2 sample (24.91% – Hydrogenophaga). However, it only accounted for 4.95% in AOBFON1. It should be noted that Hydrogenophaga could degrade 4-aminobenzene sulfonate,¹⁸ xenobiotic compounds, such as polychlorinated biphenyl and methyl tert-butyl ether.^19,20 Microbes belonging to the Flavobacteriaceae family, such as Flavobacterium, were also distinguished when utilizing acetate to stimulate activated sludge.²¹ However, it should be established that organic carbon and active denitrifying population of microbes were also associated.

Interestingly, the microbial community structures between AOBFAN1 and AOBFAN2 samples were also different from the anoxic zone of A/O-BAF process of the different ammonia nitrogen loads (Fig. 7). Anaerolineaceae_uncultured was the most dominant bacteria in the AOBFAN1 sample of a relative abundance of 11.35%, followed by vadinHA17_norank (9.91%), Desulfobacter (9.76%), Desulfomicrobium (4.26%), Aeromonas (4.26%), SB-1_norank (4.16%). VadinHA17_norank was the most dominant bacteria in AOBFAN2 samples of a relative abundance of 10.44%, followed by ST-12K33_norank (7.55%), Desulfobacter (7.43%), Bacteroides (6.51%), SB-1_norank (5.43%), and Aeromonas (5.41%). VadinHA17_norank, Desulfobacter, Aeromonas and SB-1_norank were the dominant genera shared by both AOBFAN1 and AOBFAN2 samples, and these genera have been reported to be frequently responsible for the degradation of compounds in biofilters.^22–25 Desulfobacter and Aeromonas are other anaerobic bacterias in the anoxic zone.^26–28 The microbial population and physicochemical properties of the UCSFMs well explained the excellent performance of the UCSFM A/O-BAF reactor.

4. Conclusions

A considerable amount of oyster shells is produced as waste product of mariculture every year, which leads to a major disposal problem with coastal regions of China. This work investigated the possibility of converting the oyster shell waste into the UCSFM by an unburned process, and examined the performance of the UCSFM serving as the bio-carrier in the A/O-BAF reactor. The results were as follows:

(1) The feasibility of UCSFM as a biofilm supported by A/O-BAF was verified quite satisfactorily according to the results of characteristics of UCSFM.

(2) UCSFM BAF had a slightly higher COD and NH₃–N removal efficiency compared to UCSFM A/O-BAF, but UCSFM A/O-BAF showed higher TN removal efficiency of ammonia nitrogen loads between 0.130 kg (m³ d)⁻¹ and 0.496 kg (m³ d)⁻¹. UCSFM A/O-BAF and UCSFM BAF had average TN removals of 41.88% and 7.48%, respectively. The richness of the bacterial communities in the anoxic zone of A/O-BAF operated on the ammonia nitrogen load of 0.130 kg (m³ d)⁻¹ was higher than that under the ammonia nitrogen load of 0.1496 kg (m³ d)⁻¹.

(3) Pyrosequencing analysis of the 16S rRNA gene in anoxic/oxic zones of A/O-BAF process indicated that the microbial communities between the two samples were quite different to different ammonia nitrogen loads. Thiobacillus, Thiothrix, Hydrogenophaga, and Flavobacterium were the dominant genera shared in the oxic zones, while VadinHA17_norank, Desulfobacter, Aeromonas and SB-1_norank were the dominant genera shared in the anoxic zones. The microbial population and physicochemical properties of the UCSFMs well explained the excellent performance of the UCSFM A/O-BAF reactor.

Therefore, oyster shell waste can be converted to the UCSFM, which has good biofilm adhesion and adsorption properties and can be used as an excellent bio-carrier for microorganisms in A/O-BAF reactor.

Acknowledgements

This work was funded by the National Natural Science Foundation of China (NSFC51508227, NSFC51278225, and NSFC51178207), A Project of Shandong Province Higher Educational Science and Technology Program (No. J14LG02), Shandong Province Science and Technology Development Plan-Policy Guidance Project (No. 2013YD17003).

References

1. C. L. Li, A. H. Song, W. Hu, Q. C. Li, B. Zhao, R. J. Zhu, Y. Zhang and D. P. Ma, Status analyzing and developing counter-measure of cultured scallop industry in Shandong province, Mar. Sci., 2011, 35(3), 93–98.
2. Y. W. Ding, L. Wang, B. Z. Wang and Z. Wang, Removal of nitrogen and phosphorus in a combined A(2)/O-BAF system with a short aerobic SRT, J. Environ. Sci., 2006, 18, 1082–1087.
3. G. Farabegoli, A. Chiavola and E. Rolle, The biological aerated filter (BAF) as alternative treatment for domestic sewage. Optimization of plant performance, J. Hazard. Mater., 2009, 171, 1126–1132.
4. D. Gao, Z. Li, J. Guan, Y. Li and N. Ren, Removal of surfactants nonylphenol ethoxylates from municipal sewage-comparison of an A/O process and biological aerated filters, Chemosphere, 2014, 97, 130–134.
5. Y. Feng, J. Qi, L. Chi, D. Wang, Z. Wang, K. Li and X. Li, Production of sorption functional media (SFM) from clinoptilolite tailings and its performance investigation in a biological aerated filter (BAF) reactor, J. Hazard. Mater., 2013, 246–247, 61–69.
6. J. H. Ha and S. K. Ong, Nitrification and denitrification in partially aerated biological aerated filter (BAF) with dual size sand media, Water Sci. Technol., 2007, 55(1–2), 9–17.
7. T. D. Kent, C. S. B. Fitzpatrick and S. C. Williams, Testing of Biological Aerated Filter (BAF) Media, Water Sci. Technol., 1996, 34(3–4), 363–370.
8. M. F. Rosa and A. A. L. Furtado, Biofilm development and ammonia removal in the nitrification of a saline wastewater, Bioresour. Technol., 2008, 65, 135–138.
9. O. Francisco and H. Ernesto, Wastewater treatment with a double-layer submerged biological aerated filter, using waste materials as biofilm support, Environ. Manage., 2002, 65, 79–84.
10. S. Han, Q. Yue, M. Yue, B. Gao, Y. Zhao and W. Cheng, Effect of sludge-fly ash ceramic particles (SFCP) on synthetic wastewater treatment in an A/O combined biological aerated filter, Bioresour. Technol., 2009, 100, 1149–1155.
11. Y. Feng, Y. Yu, Q. Duan, J. Tan and C. Zhao, The characteristic research of ammonium removal in grain-slag biological aerated, Desalination, 2010, 263, 146–150.
12. L. P. Qiu and J. Ma, Characters of biofilm and its microorganism population in biological aerobic filter, China Environ. Sci., 2005, 25, 214–217.
13. State Environmental Protection Administration of China, Monitoring and Analysis Methods of Water and Wastewater, China Environmental Science Press, Beijing, 4th edn, 2002.
14. A. Albuquerque, A. Gonzalez-Martinez and F. Osorio, Effect of aeration on steady-state conditions in non- and partially aerated low-loaded biofilter, J. Environ. Sci. Technol., 2012, 9, 395–408.
15. H. R. Beller, P. S. Chain, T. E. Letain, A. Chakicherla, F. W. Larimer and P. M. Richardson, et al., The genome sequence of the obligately chemolithoautotrophic, facultatively anaerobic bacterium Thiobacillus denitrificans, J. Bacteriol., 2006, 188, 1473–1488.
16. J. Y. Park and Y. J. Yoo, Biological nitrate removal in industrial wastewater treatment: which electron donor we can choose, Appl. Microbiol. Biotechnol., 2009, 82, 415–429.
17. K. Yoon, Y. Sakai, N. Tsukada, K. Fujisawa and H. Nishihara, Purification and biochemical characterization of a membrane-bound [NiFe]-hydrogenase from a hydrogen-oxidizing, lithotrophic bacterium, Hydrogenophaga sp. AH-24, FEMS Microbiol. Lett., 2009, 290, 114–120.
18. H. M. Gan, S. Shahir, Z. Ibrahim and A. Yahya, Biodegradation of 4-aminobenzenesulfonate by Ralstonia sp PBA and Hydrogenophaga sp PBC isolated from textile wastewater treatment plant, Chemosphere, 2011, 82, 507–513.
19. P. B. Hatzinger, K. McClay, S. Vainberg, M. Tugusheva, C. W. Condee and R. J. Steffan, Biodegradation of methyl tert-butyl ether by a pure bacterial culture, Appl. Environ. Microbiol., 2001, 67, 5601–5607.
20. A. J. Lambo and T. R. Patel, Cometabolic degradation of polychlorinated biphenyls at low temperature by psychrotolerant bacterium Hydrogenophaga sp. IA3-A, Curr. Microbiol., 2006, 53, 48–52.
21. T. Osaka, S. Yoshie, S. Tsuneda, A. Hirata, N. Iwami and Y. Inamori, Identification of acetate- or methanol-assimilating bacteria under nitrate reducing conditions by stable-isotope probing, Microb. Ecol., 2006, 52, 253–266.
22. I. Krustok, J. Truu, M. Odlare, M. Truu, T. Ligi, K. Tiirik and E. Nehrenheim, Effect of lake water on algal biomass and microbial community structure in municipal wastewater-based lab-scale photobioreactors, Appl. Microbiol. Biotechnol., 2015, 99(15), 6537–6549.
23. I. V. Trubitsyn, E. V. Belousova, M. N. Tutukina, A. Y. Merkel, G. A. Dubinina and M. Y. Grabovich, Expansion of ability of denitrification within the filamentous colorless sulfur bacteria of the genus Thiothrix, FEMS Microbiol. Lett., 2014, 358(1), 72–80.
24. P. Jordan, E. A. Marchand, J. Silverstein and M. Hernandez, Development and Application of Small-Subunit rRNA Probes for Assessment of Selected Thiobacillus Species and Members of the Genus Acidiphilium, Appl. Environ. Microbiol., 2000, 66(7), 3065–3072.
25. D. C. Yang, W. T. Im, M. K. Kim and S. T. Lee, Pseudoxanthomonas koreensis sp. nov. and Pseudoxanthomonas daejeonensis sp. nov, Int. J. Syst. Evol. Microbiol., 2005, 55, 787–791.
26. R. M. Abed, N. Musat, F. Musat and M. Mussmann, Structure of microbial communities and hydrocarbon-dependent sulfate reduction in the anoxic layer of a polluted microbial mat., Mar. Pollut. Bull., 2011, 62(3), 539–546.
27. A. M. Vieira, R. Bergamasco, M. L. Gimenes, C. V. Nakamura and B. P. Dias Filho, Microbial populations of an upflow anaerobic sludge blanket reactor treating wastewater from a gelatin industry, Environ. Technol., 2001, 22(12), 1477–1485.
28. T. Hao, L. Wei, H. Lu, H. Chui, H. R. Mackey, M. C. van Loosdrecht and G. Chen, Characterization of sulfate-reducing granular sludge in the SANI(R) process, Water Res., 2013, 47(19), 7042–7052.

---