Bioresource Technology 172 (2014) 22–31

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Startup pattern and performance enhancement of pilot-scale biofilm process for raw water pretreatment Guang-feng Yang a, Li-juan Feng a,c, Qi Yang a, Liang Zhu a,b,⇑, Jian Xu a, Xiang-yang Xu a,b a

Department of Environmental Engineering, Zhejiang University, Hangzhou 310058, China Zhejiang Province Key Laboratory for Water Pollution Control and Environmental Safety, Hangzhou 310058, China c Department of Environmental Engineering, Zhejiang Ocean University, No. 1 Haida South Road, Zhoushan 316022, China b

h i g h l i g h t s  Two pilot-scale biofilm reactors were established with different startup strategies.  Precoated biofilm carriers favor biomass enrichment and reduce nitrite accumulation. 1

 The optimum DO level for adequate nitrification was 1.0–2.6 mg L

.

 The suitable temperature range for adequate nitrification was 21–22 °C.  The presentence of algae increased the risk of disinfection by-products production.

a r t i c l e

i n f o

Article history: Received 27 June 2014 Received in revised form 24 August 2014 Accepted 26 August 2014 Available online 3 September 2014 Keywords: Raw water Biofilm Pretreatment process Nitrogen removal Trihalomethane prediction model

a b s t r a c t The quality of raw water is getting worse in developing countries because of the inadequate treatment of municipal sewage, industrial wastewater and agricultural runoff. Aiming at the biofilm enrichment and pollutant removal, two pilot-scale biofilm reactors were built with different biological carriers. Results showed that compared with the blank carrier, the biofilm was easily enriched on the biofilm precoated carrier and less nitrite accumulation occurred. The removal efficiencies of NH+4-N, DOC and UV254 increased under the aeration condition, and a optimum DO level for the adequate nitrification was 1.0–2.6 mg L1 with the suitable temperature range of 21–22 °C. Study on the trihalomethane prediction model indicated that the presentence of algae increased the risk of disinfection by-products production, which could be effectively controlled via manual algae removing and light shading. In this study, the performance of biofilm pretreatment process could be enhanced under the optimized condition of DO level and biofilm carrier. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In developing countries, abundant nutrients are discharged from the municipal sewage, industrial wastewater and agricultural runoff because of their inadequate treatment and disposal, which cause the serious pollution problems of natural waters. Thereinto, the pollution of drinking water source directly threatens the drinking water safety (Chu et al., 2010; Zhang et al., 2013). For example, the major pollutants such as ammonia and organics existing in water could complicate the chlorination process and form disinfection byproducts (Chowdhury et al., 2009; Benner et al., 2013; Han et al., 2013), which would cause health problems to people. However, ⇑ Corresponding author at: Department of Environmental Engineering, Zhejiang University, No. 866 Yuhangtang Road, Hangzhou 310058, China. Tel.: +86 571 88982343; fax: +86 571 28865333. E-mail address: [email protected] (L. Zhu). http://dx.doi.org/10.1016/j.biortech.2014.08.116 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

traditional raw water treatment processes including coagulation, sedimentation, sand-filtration and disinfection, could not remove nitrogen and dissolved organic carbon (DOC) effectively. The biofilm process is a promising alternative for the pretreatment of raw water because of its potential economic advantages, lower secondary pollution, and less disinfection by-products (DBPs) production (Bruce and Douglas, 2002; Yu et al., 2007; Chu et al., 2011; Qian et al., 2011; Yu et al., 2012; Feng et al., 2013; Han et al., 2013; Zhang et al., 2013). But it is difficult to start up because of the limited growth and enrichment of functional microorganism in the oligotrophic niche (Egli, 2010). In addition, the performance of biofilm process is affected by the influent quality, dissolved oxygen (DO) level, temperature and presence of algae (Mallick, 2002; Feng et al., 2013; Han et al., 2013; Zhang et al., 2013). However, few literatures reported the effect of startup pattern on the performance and potential risk of biofilm process for the raw water pretreatment.

23

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

5

In this study, two pilot-scale biofilm reactors with different biological carriers are established for raw water pretreatment. The purposes are to investigate the effect of startup pattern on the performance of pilot-scale biofilm process, and also to reveal its potential risk and control strategy in the presentence of algae.

8

2

7 3

1

2. Methods

4

8

9

10

6

2.1. Characteristics of polluted raw water The polluted raw water was collected from a river located in Hangzhou, China. The major quality are summarized in Table 1. This kind of water is similar to polluted raw water in Eastern China according to the on-the-spot investigation and related literature (Chu et al., 2010; Feng et al., 2013; Zhang et al., 2013). 2.2. Biofilm pretreatment process 2.2.1. Reactor setup Two continuous flow reactors (R1 and R2) with an effective volume of 63.9 L (147 cm  14.5 cm  30 cm) were used in this study (shown in Fig. 1). The reactors filled with the same volumetric carriers (TA-II elastic filler, purchased from Tianyu Environmental Protection Engineering Co., Ltd.; 1.52% of filling ratio), which had a diameter and surface area of 200 mm and 18 m2 m3, respectively. In the whole operation period, the environmental temperature was 10–41 °C, corresponding to an influent temperature of 11–35 °C. The carriers in R1 with precoated biofilm were elastic filler obtained from a continuous reactor feeding synthetic polluted raw water, and the biomass on the carrier is approximately 0.95 ± 0.07 mg TOC g1 carrier. In reactor R2, blank carriers were added as the control.

Fig. 1. Schematic diagram of biofilm pretreatment process: (1) influent tank; (2) peristaltic pump; (3) influent area; (4) water diffuser; (5) aerator; (6) air diffuser; (7) reactor; (8) elastic carrier; (9) effluent area; (10) effluent tank.

total organic carbon (TOC) analyzer (TOC-V CPH, Shimadzu). UV254 was read from a spectrophotometer (UV-2401PC, Shimadzu) at 254 nm wavelength. Chlorophyll a, b and c were used as indicators to quantify the algae in the water, and the ratio change of chlorophyll a, b and c could be used to illustrate the algae species changing. Spectrophotometry was employed to describe the absorbance of chlorophyll extracting solution at the wavelength of 630, 645, 663 and 750 nm (Chinese SEPA, 2002). The levels of chlorophyll a, b and c could be calculated according to equations (1–3) presented below. Chla ¼

Chlb ¼

Chlc ¼

2.2.2. Reactor operation The reactors are operated for more than 6 months, and the whole operation period is divided into four stages (P1, P2, P3 and P4). The objectives and operating parameters are shown in Table 2. At stage P1, the reactors were fed with the polluted raw water, and natural biofilm formation method was used. At stage P2, the effects and the growing characteristics of algae were studied under high temperature. For reducing harmful algae in raw water, the algae were controlled and the operation performance evolution after algae removal was estimated in P3. In P4, the performance was investigated under different DO levels.

ð11:64ðD663  D750 Þ  2:16ðD645  D750 Þ þ 0:10ðD630  D750 ÞÞ  V 1 V2  L ð1Þ ð11:64ðD663  D750 Þ  2:16ðD645  D750 Þ þ 0:10ðD630  D750 ÞÞ  V 1 V2  L ð2Þ ð11:64ðD663  D750 Þ  2:16ðD645  D750 Þ þ 0:10ðD630  D750 ÞÞ  V 1 V2  L ð3Þ

2.3. Analysis methods

2.3.2. Biomass and biofilm analysis The elastic carrier was sampled from reactors and then exposed to ultra-sound wave (300 W) for 4 h to remove the biofilm attached on the carrier. Total solids (TS), total carbon (TC), inorganic carbon (IC), TOC and DOC were used for biomass change (expressed as weight of biomass/weight of carrier, mg g1). TS was detected according to the Chinese State Environ-mental Protection Agency (SEPA) Standard Methods (ChineseSEPA, 2002). TOC analyzer (TOC-V CPH, Shimadzu) was used for analysis of TOC, IC and DOC.

2.3.1. Chemical index analysis The water samples were routinely collected and analyzed using Standard Methods issued by Chinese SEPA (2002). Unfiltered water samples were used for analysis of turbidity, total nitrogen (TN) and total phosphorus (TP). The water samples were pre-filtered using a 0.45 lm (pore size) glass fiber filter before ammonium, nitrite, nitrate, DOC and UV254-detectable compounds (UV254) were analyzed. The pH was determined using pH meter with a selective electrode (METTLER TOLEDO 320, Switzerland). DO meter (YSI Model52, USA) was employed to measure the DO level. DOC was analyzed using a catalyzed combustion

2.3.3. DNA extraction and PCR-DGGE Total DNA of different biofilm samples were extracted using a soil DNA kit (OMEGA) (Feng et al., 2013). Polymerase chain reaction (PCR) was employed to amplify the V3 region of 16S rRNA, and the universal bacterial primers P357GC and P518 were used for DGGE analysis. The PCR was conducted out using a thermal cycler under the conditions same as the description of Han et al. (2013). 5 lL PCR products were detected by electrophoresis on a 0.8% agarose gel stained with goldview. DGGE was performed using an AD-code mutation detection system (Bio-Rad, Hercules, CA, USA) as reported by Feng et al. (2012). Samples containing

Table 1 The major quality of polluted raw water. Parameters

T (°C)

Turbidity (NTU)

TOC (mg L1)

UV254

NH+4-N (mg L1)

1 NO ) 2 -N (mg L

1 NO ) 3 -N (mg L

TP (mg L1)

Range Average (±SD)

11–35 26 ± 11

0–61.4 12.4 ± 12.2

0.5–13.9 4.4 ± 2.9

0.0721–0.1810 0.0992 ± 0.0183

0.21–3.90 1.43 ± 0.90

0.00–0.80 0.15 ± 0.13

0.36–2.33 1.09 ± 0.39

0.04–0.50 0.15 ± 0.09

24

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31 Table 2 Operating parameters in biofilm reactors. Stages

Period (days)

Water temperature ( °C)

NH+4-N loading rate (g m3 d1)

Experimental content

P1 P2 P3

1–25 26–61 62–80(P31) 81–107(P32) 108–184

22.8 ± 5.1 26.2 ± 3.3 27.5 ± 2.5 31.5 ± 2.8 25.4 ± 5.3

0.76 ± 0.39 1.43 ± 0.65 1.39 ± 0.65 1.26 ± 0.85 3.72 ± 0.71

Startup of biofilm reactors Performance at coexistence of bacteria and algae Performance evolution after algae removal Performance recovery at low influent NH+4-N level Operation at various DO levels

P4

35 lL PCR amplicons were loaded onto 8% (w/v) polyacrylamide gels with an acrylamide/bisacrylamide ratio of 37.5. The denaturing gradient ranged from 30% to 60% denaturant (denaturant (100%) contains 7 M urea and 40% (v/v) formamide in 1  TAE). Electrophoresis was first performed for 30 min at 30 V with a temperature of 60 °C, and then 5.5 h at 155 V without changing temperature. Staining gels for 30 min using SYBR Green to obtain the DGGE profile. Gel Documentation System (BIO-Rad Laboratories, Segrate, Italy) was used to capture the DGGE photos and Quantity One software (Version 4.62) was used to analyze the DGGE profile. 2.3.4. Mathematical model assessment There are many studies reported the models for predicting disinfection byproduct (DBP) formation in drinking waters (Chowdhury et al., 2009; Feng et al., 2013), which could be used to assess the performance of raw water pretreatment systems and the risk to the subsequent process. When the formation reaction would not be limited by lacking chlorine in chlorination (residue chlorine-Cl > 1 mg L1), empiric models based on DOC and UV254 (Eqs. (4) and (5)) could be used to predict the DBPs’ formation (Hong et al., 2008; Feng et al., 2013).

TTHM ¼ 0:125½DOC0:852 pH1:801 t0:246

ð4Þ

TTHM ¼ 2:697½UV254 0:654 pH1:718 t 0:227

ð5Þ

1

where TTHM is total THMs (lg L ); [DOC] is DOC concentration (mg L1); [UV254] is UV254 level (cm1); t is reaction time (h). Generally, the chlorination time is more than 2 h. In this study, we used the reaction time of 3 h to predict the THMs. 3. Results 3.1. Start-up of biofilm reactors (P1) During the start-up period (P1), the nitrification of R1 was more complete and stable than that of R2 (p > 0.01) (Fig. 2a, b and g) due to the carrier with precoated biofilm. As shown in Fig. 2(a and b), the NH+4-N removal efficiencies (ARE) of R1 and R2 are 65.2 ± 17.8% and 63.6 ± 14.3%, respectively. T-test showed that ARE in both reactors were not significantly different (p > 0.01). The result showed that the two kinds of startup strategies had no significant difference during the start-up periods. However, the effluent NH+4-N concentration in R1 was less than 0.5 mg L1 (threshold NH+4-N concentration of Standards for Drinking Water Quality in China, GB5749-2006) with an average level of 0.30 ± 0.08 mg L1, which was lower and more stable than that of R2 (effluent NH+4-N level of 0.12–0.67 mg L1). Because the nitrite concentration in the effluent of R2 was greater than that of R1 (p > 0.01) (Fig. 2g), the nitrification of R1 was more complete. This result showed that using carrier precoated biofilm to startup the process could effectively reduce the accumulation of nitrite. Influent organic matter levels varied from 0.5–13.9 mg L1 (DOC) during the whole experimental period, with an average

value of 4.4 ± 2.9 mg L1 DOC. Compared with R2, the effluent DOC concentration of R1 was low. The denification may be contribute to reduce of organic matter, and the average removal efficiencies of nitrate in R1 and R2 were 42.4 ± 21.7% and 14.7 ± 32.3%, respectively. 3.2. Performance at coexistence of bacteria and algae (P2) The presentence of algae could not affect ammonia oxidation but affect the NO 2 -N removal performance. Although the maximum ARE of R1 and R2 reached to 90.1% and 89.0% respectively, the average removal efficiency of R1 was not significantly increased comparing the values in P1 (p > 0.01), and even decreased in R2 (Fig. 2 a and b). During stage P2, the effluent NH+4-N concentrations in R1 and R2 were 0.37 ± 0.27 and 0.48 ± 0.44 mg L1 respectively, which were higher than those at stage P1. Expect for the first 12 d and the last 7 d of P2, the effluent NO 2 -N levels were no more than 0.08 mg L1, and not significantly different (p > 0.01) in both reactors. However, the NO 3 -N removal was significantly affected due to the oxygen production by photosynthetic reaction. As shown in Fig. 3(a and b), the chlorophyll contents are accumulated due to the retention and the growth of algae in both reactors at stage P2. The algae concentrations were closely linked to the levels of chlorophyll, and the quantitative relationship between chlorophyll and algae, i.e., chlorophyll level in unit algae (mg g1), was expressed in Table 3. The chlorophyll (a, b or c) contents per algae weight were similar to each other in reactors R1 and R2. For example, the levels of chlorophyll a in reactor R1 and R2 were 0.724 and 0.672 mg per gram algae. 3.3. Algae growth and controlling (P3) 3.3.1. Short-term instability after algae removal (P31) At the beginning of stage P3, the methods of manual removing algae and light-shading were used to remove the existed algae and inhibit the growth of algae in the reactors. However, the performance of two reactors was impaired due to the algal cells destruction in short-term operation. The effluent ammonia concentrations were obvious higher than the influent levels in both reactors R1 and R2. As shown in Fig. 2(a and b), the effluent ammonia levels of R1 and R2 in days 62–80 (P31) are 0.78–3.24 and 0.84–4.33 times of those in influent. The maximum TOC release was observed in both reactors, and the maximum effluent TOC levels of R1 and R2 were 3.2 and 7.2 times of those in influent, respectively. At the same time, the effluent UV254 were 0.1234 ± 0.0267 and 0.1379 ± 0.0589 cm1, respectively, both greater than the influent level of 0.0978 ± 0.0181 cm1 (Fig. 2e and f). After the successful algae removal, the chlorophyll levels in effluent and in both reactors were obviously decreased (Fig. 3a–c). 3.3.2. Recovery after algae removal (P32) Although no mechanical aeration was used, in the later operation period of P3 (P32, days 81–107), the DO levels in R1 and R2 were 0.70 ± 0.21 and 0.83 ± 0.33 mg L1, respectively. The average ammonia removal efficiencies in R1 and R2 were 62.0% and 64.5%

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

Fig. 2. The removal performance of nitrogen, turbidity and UV254 in the two reactors: (a), (c) and (e): R1; (b), (d) and (f): R2; (g): R1 and R2.

25

26

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

Fig. 3. The variation of chlorophyll contents in the different parts of biofilm reactors: (a) day 31; (b) day 61; (c) day 101. (1) influent; (2) effluent of R1; (3) effluent of R2; (4) in the front of R1; (5) in the middle of R1; (6) in the end of R1; (7) in the front of R2; (8) in the middle of R2; (9) in the end of R2.

Table 3 Chlorophyll level of biofilm reactors. Reactors

Chlorophyll level per algae wet weight (mg g1) Chl a in algae

Chl b in algae

Chl c in algae

Chl a in algae

Chl b in algae

Chl c in algae

R1 R2

0.724 0.672

0.299 0.299

0.203 0.198

2.292 2.786

1.558 1.848

0.724 0.672

with average effluent ammonia concentrations of 0.32 ± 0.19 and 0.21 ± 0.06 mg L1, respectively. No NO 2 -N accumulation was observed in effluent of both reactors (Fig. 2g), and the effluent NO 3 -N levels were obviously increased. The average TN removal rate of R1 was low to 0.06 g m3 d1, and the average TN removal rate of R2 was higher (0.24 g m3 d1). The results showed that natural reoxygenation provide enough oxygen for nitrification at a low influent NH+4-N level (no more than 1 mg L1), however the TN removal ability was restricted due to a weak denitrification reaction.

Chlorophyll level per algae dry weight (mg g1)

3.4. Optimization of biofilm process At stage P4, two reactors were operated at different DO levels with a high influent NH+4-N level of 2.56 ± 0.49 mg L1. As shown in Table 4 and Fig. 2, the whole operation period (P4) are divided into four sub-periods (P41, P42, P43 and P44) according to the DO levels in reactors. During P41, the influent ammonia concentration reached 2.62 ± 0.25 mg L1, the effluent ammonia levels of R1 and R2 were high to 1.55 ± 0.53 and 1.29 ± 0.46 mg L1 respectively due to oxygen lacking.

Table 4 Nitrogen removal of biofilm reactors at various DO levels. Reactors

Operation period (days)

R1

P4-1 P4-2 P4-3 P4-4 P4-1 P4-2 P4-3 P4-4

R2

108–128 129–143 144–165 166–184 108–128 129–143 144–165 166–184

DO (mg L1)

NH+4-N (mg L1)

1 NO ) 2 -N (mg L

1 NO ) 3 -N (mg L

Influent

Reactor

Influent

Effluent

Influent

Effluent

Influent

Effluent

3.07 ± 1.17 1.94 ± 1.32 4.43 ± 0.37 5.50 ± 2.17 3.07 ± 1.17 1.94 ± 1.32 4.43 ± 0.37 5.50 ± 2.17

0.52 ± 0.14 1.09 ± 0.81 2.32 ± 0.37 3.68 ± 0.77 0.38 ± 0.13 1.03 ± 0.73 2.16 ± 0.75 1.79 ± 0.87

2.62 ± 0.25 2.38 ± 0.36 2.57 ± 0.50 2.37 ± 0.78 2.62 ± 0.25 2.38 ± 0.36 2.57 ± 0.50 2.37 ± 0.78

1.55 ± 0.53 0.82 ± 0.54 0.69 ± 0.35 0.95 ± 0.62 1.29 ± 0.46 0.29 ± 0.09 0.50 ± 0.16 0.66 ± 0.52

0.09 ± 0.03 0.13 ± 0.08 0.25 ± 0.16 0.15 ± 0.12 0.09 ± 0.03 0.13 ± 0.08 0.25 ± 0.16 0.15 ± 0.12

0.02 ± 0.02 0.02 ± 0.01 0.02 ± 0.03 0.03 ± 0.03 0.03 ± 0.03 0.01 ± 0.01 0.03 ± 0.04 0.03 ± 0.03

0.71 ± 0.15 0.92 ± 0.19 1.19 ± 0.40 1.00 ± 0.51 0.71 ± 0.15 0.92 ± 0.19 1.19 ± 0.40 1.00 ± 0.51

1.37 ± 0.18 2.54 ± 0.58 3.68 ± 0.27 2.33 ± 1.16 1.58 ± 0.17 3.04 ± 0.27 3.67 ± 0.41 0.32 ± 0.67

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

For enhancing the ammonia removal, the mechanical aeration was employed in P42, P43 and P44. The NH+4-N removal performance was obviously increased in both reactors R1 and R2 (Shown in Table 4). In reactor R1, the maximum NH+4-N removal efficiency reached to 91.1% (P42) from 70.0% (P41) in R1, the average NH+4-N removal efficiency increased from 40.9% ± 18.9% (P41) to 65.6 ± 22.6% (P42) and 73.4 ± 14.2% (P43). Similar phenomenon was observed in R2, the maximum NH+4-N removal efficiency reached 90.6% during stageP42, which was similar to that of R1.

27

However, average NH+4-N removal efficiency of 87.7 ± 2.5% (P42) in R2 was significant greater than that in R1 (p > 0.01). 3.5. Periodic and spatial variation of pollutant removal performance 3.5.1. Periodic variation of pollutant removal in biofilm reactors At different periods, the daily periodic variation of pollutant removal performance of raw water biofilm pretreatment systems  was studied. The temporal variation of NH+4-N, NO 2 -N, NO3 -N of

Fig. 4. The temporal and spatial variation of pollutant removal performance of biofilm reactors: (a–c) temporal variation, (d–f) spatial variation: (d) day 31; (e) day 61; (f) day 101. (1) influent; (2) effluent of R1; (3) effluent of R2; (4) in the front of R1; (5) in the middle of R1; (6) in the end of R1; (7) in the front of R2; (8) in the middle of R2; (9) in the end of R2.

28

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

both reactors is presented in Fig. 4(a–c). The influent NH+4-N level was 4.47 ± 0.27 mg L1, and the effluent NH+4-N concentration in both reactors were at low levels at night time (0, 6 h). This may be caused by the variation of temperature and DO level in reactors. There are NO 2 -N accumulation in both reactors due to DO consumption. Because of a higher DO concentration in R1 (1.28 mg L1) the NO 2 -N accumulation in R2 was lower than that in R1. The algae in R1 were more than that in R2 (Fig. 3). Although more oxygen was produced in day time, the available DO may be less without mechanical or enough hydraulic diffusion. As shown in Fig. 4(a–c), the effluent NO 3 -N level of R2 was higher than that of R1. However, the TN (=sum of NH+4-N, NO 2 -N and NO 3 -N) removal of R2 was better than R1. This may be attributing to the effects of aerobic denitrifying bacteria (Zhu et al., 2012), which could utilize aerobic denitrifying enzyme to remove the nitrogen. 3.5.2. Spatial variation of reactor performance To investigate the spatial shift of biofilm pretreatment performance, the samples obtained different sampling points was detected. Three sampling points distributed along the length of the reactor (25, 74 and 123 cm from the front to the end of the reactor, respectively) were analyzed. As shown in Fig. 4(d–f), the change in NH+4-N and NO 2 -N concentrations from the front to the end of reactors has a similar trend of gradually decreased. The NO 2 -N accumulation was obvious on day 31 and 61. The removal

efficiency of R2 was inferior to that of R1. However, low NO 2 -N concentration was observed in both reactors. Different from the  change of NH+4-N and NO 2 -N, the shift of NO3 -N on day 101 was  opposite to that on day 31 and 61and the NO3 -N change trend sim ilar to that of NH+4-N and NO 2 -N. On day 101, the NO3 -N concentrations are gradually increased. As shown in Fig. 4(e and f), the NH+4-N of influent was mostly removed at the front of the reactors after the mature of biofilm. Only few NH+4-N removal occurred at the middle and the end of both reactors.

3.6. Biomass evolution of biofilm process At the startup of R1, DOC concentration was 0.76 ± 0.04 mg g1 with a TOC level of 0.95 ± 0.07 mg g1, while there was no biomass on the carrier used in R2. The performance of R2 was inferior to that of R1. On day 61, the biofilm biomass (DOC) of R1 and R2 were 0.78 ± 0.02 and 0.38 ± 0.05 mg g1, respectively, along with TOC concentration of 1.11 ± 0.11 and 0.65 ± 0.12 mg g1, respectively. Without competition of algae for substrates, the biomass in R1 and R2 was further increased. At the end of this experiment, SS increased 3.1 and 2.1 times for R1 and R2, respectively (Fig. 5). The DOC and TOC in the front of R1 were 2.43 and 2.60 times to that of R2. However, the biomass attached on the carrier could not complete desquamated (Zhang et al., 2013), because the inherent defect of biomass extraction method used in the study.

Fig. 5. The variation of biofilm in biofilm reactors: (a) DOC; (b) IC; (c) TC; (d) SS (Sample obtained from: (1, 2) bottom and top in the front of R1 (day 61); (3) in the front of R1 (day 184); (4, 5) bottom and top in the end of R1 (day 61); (6) in the end of R1 (day 184); (7, 8) bottom and top in the front of R2 (day 61); (9) in the front of R2 (day 184); (10, 11) from bottom and top in the end of R2 (day 61); (12) in the end of R2 (day 184); (13) R1 (day 0)).

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

The ratio of DOC/TOC in biofilm were varied from 0.47 to 0.80; the result showed that the microbe has been changed in the biofilm. The PCR-DGGE results (Fig 6 and Table 5) showed that the Shannon diversity index (H) of sample A, B, C, D, E and F were 2.82, 2.88, 2.82, 2.39, 3.11 and 2.54, respectively (biofilm sample obtained from A, top in the front of R1, day 61; B, bottom in the front of R1, day 61; C, top in the front of R2, day 61; D, in the front of R1, day 184; E, in the front of R2, day 184; F, carrier procoted biofilm of R1, day 0). At the same time, the similarity analysis indicated that the change of microbial community is obvious (Table 5), and it’s in accord with the variation of biomass and DOC. 4. Discussion 4.1. Effect of precoated biofilm on the pollutant removal It was reported that after the raw water pretreatment reactor stopped operation for about 1 month, the re-startup period was shortened (2 d) even at a temperature of 6–10 °C with an ARE of 70% because the biofilm was preserved at a low temperature of 2–5 °C, and the effective ammonia oxidation bacteria (AOB) and nitrite oxidation bacteria (NOB) were still alive under this condition (Han et al., 2013). In this study, the startup of R1 could be considered as the re-startup with the precoated biofilm. After the initial two day’s operation, ARE reached to 82.8%. However, the ammonia removal performance was unstable during this period. This may be caused by the facts that the precoated biofilm was formed at a condition of synthetic polluted raw water (using biodegradable CH3OH and KNO3 as carbon source and nitrogen source, respectively as described by Feng et al. (2013). When it is used

29

the complicated real polluted raw water, the activity of biofilm could be further suppressed because organics in the raw water was difficult to be assimilated. As shown in Fig. 6, the biomass characteristics of R1 and R2 were significantly different. The similarity analysis showed that the microorganism similarity in the biofilm was varied from 48.4% (bands D and E) to 67.3% (bands B and C). On day 184, the bacteria species in the biofilm located in the front of R1 (band D) and R2 (band E) had the minimum microorganism similarity, which showed that even operated at the same operation conditions, the biomass formed in the carrier was clearly different (Table 5). The biomass indicated by DOC, TOC, DC and TC of per unit carrier (mg g1) on the carrier in the front of reactor was obvious greater than those in the end in R1. While in R2, the difference of biomass in the front and end of reactor was not obvious. The carrier precoated with biofilm maybe cause a majority of nutrient (nitrogen and organic matter) consumed in the front of R1. The spatial variation of operation performance of both reactors showed that a higher NH+4-N level was consumed in the front of reactor, corresponding to a higher biomass (DOC and TOC) level (Figs. 5 and 6). Zhang et al. (2013) reported that both temperature and NH+4-N rate in raw water pretreatment system significantly affected the biofilm characteristics. In this study, both reactors operated at the same NH+4-N rate and went through the same temperature condition, the difference of biofilm characteristics was caused by different used carrier . The change of SS in two reactors was not always identical to that of the shift of DOC and TOC, which could be used as the indicator of biotic components. A higher SS was always observed in the front of reactor (Fig. 6). This may be caused by the abiotic sediment or suspend solid attached to the carrier (Fig. 2f). 4.2. Process optimization in oligotrophic niche

Fig. 6. The PCR-DGGE profile of biofilm obtained from the pretreatment process: (A) top in the front of R1 (day 61); (B) bottom in the front of R1 (day 61); (C) top in the front of R2 (day 61); (D) in the front of R1 (day 184); (E) in the front of R2 (day 184); (F) carrier procoted biofilm of R1, (day 0).

Table 5 Similarity matrix of DGGE profile analysis. Lane

A

B

C

D

E

F

A B C D E F

100 55.4 61.4 53.8 55.9 60.9

55.4 100 67.3 54.5 61.7 65.8

61.4 67.3 100 59 65.4 60.3

53.8 54.5 59 100 48.4 50.6

55.9 61.7 65.4 48.4 100 61.8

60.9 65.8 60.3 50.6 61.8 100

The standard deviation (SD) values of effluent quality or removal efficiency could be used to characterize the operation stability. A higher SD value means a higher fluctuation of performance. The SD values of NH+4-N in influent and effluent presented in Table 4 showed that the NH+4-N removal performance of R2 was more stable than that of R1. At the same time, the DOC removal was better in whole period of P4 than previous periods (from P1 to P3). The average DOC removal efficiencies of R1 and R2 in P4 were 25.4 ± 19.9% and 17.6 ± 14.4% respectively, with the maximum values of 70.3% and 66.4%, respectively. The UV254 removal performance was also enhanced in this period (Fig. 2e and f). DO level in bioreactors affected the performance of biological pretreatment (Qin et al., 2008; de Vet et al., 2009; Han et al., 2013; Feng et al., 2013; Zhang et al., 2013). The literatures showed that the DO concentration in biofilm raw water pretreatment systems were various from 0.01 mg L1 to saturation level according to different treatment objectives (Qin et al., 2008; de Vet et al., 2009; Feng et al., 2013; Han et al., 2013; Zhang et al., 2013). At a low DO concentration, the performance of NH+4-N removal would be inhibited (Han et al., 2013). During stage P4 of this study, the DO levels in influent and reactor were 3.24 ± 1.0 and 0.70 ± 0.21 mg L1, respectively. The NH+4-N removal was satisfactory when the influent loading rate was low. However, NH+4-N removal performance was limited when the NH+4-N concentration increased to 2.62 ± 0.25 mg L1 (P41, Fig. 2a and b). This phenomenon was caused by the lacking of DO. After increasing the DO level in bulk liquid phase, the NH+4-N removal was enhanced (P42 and P43). As shown in Table 4 and Fig. 2(a and b), we concluded that a higher DO concentration does not always mean a higher NH+4-N removal. It was reported that algae could be used as a novel biotechnology to remove N, P and metal in

30

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

wastewater because photosynthetic oxygen production by algae could replace mechanical aeration and save energy simultaneously (Mallick, 2002). In this study, the oxygen content exceeded 10 mg L1 during the daytime before algae removed, because of the photosynthetic oxygen production by algae. However, the oxygen attached on the surface of algae could not be effectively used by aerobic microorganisms such as nitrifying bacteria without full gas–liquid mixing. Aeration intensity could not only affect the oxygen transport, but also impose shear stress to affect the biofilm thickness on the carrier (Han et al., 2013). The thickness of the biofilm and distribution of the bacteria affected oxygen transfer (Asiedu, 2001; Leung et al., 2006). When DO concentrations in both reactors sharply decreased to a level of lower than 0.5 mg L1, which may not maintain the activity of AOB and NOB. When the best ammonia removal performance was observed, the corresponding DO level was the optimum DO level. There is an optimum DO level ranged 1.0 to 2.6 mg L1 for nitrification in this study. It was reported that the DO concentrations varied from 2 mg L1 to saturation level in the biofilm raw water treatment system (Qin et al., 2008; de Vet et al., 2009; Han et al., 2013; Zhang et al., 2013). A lower DO level for NH+4-N removal was observed in this study, and the nitrite accumulation was not obvious. In addition, we also found that aeration could favor the removal of UV254. Temperature is one of the important environmental factors that affect the biofilm performance (Qin et al., 2008; Warneke et al., 2011; Zhang et al., 2013). Q10 is a useful indicator to quantify the effects of temperature on the biological reaction (Warneke et al., 2011), which indicating the increasing of reaction rate with 10 °C raised. Biochemical reaction roughly doubles for every 10 °C increasing (Rittmann and McCarty, 2001; Lee et al., 2011). However, the Q10 values of NH+4-N removal efficiency in R1 and R2 were -1.14 (R2 = 0.8401) and -0.84 (R2 = 0.7161) between the temperature range of 20–30 °C (P4). These negative values showed that the NH+4-N removal was negative correlated to the temperature rise, and higher temperature could inhibit nitrification. The optimal growth temperature for AOB was 28–29 °C (Fdz-Polanco et al., 1994) or 30 °C (Lee et al., 2011). Beyond 30 or lower than 10 °C, nitrification could be inhibited (Jiao et al., 2011; Lee et al., 2011). The optimum temperature for the ammonia removal in biofilm raw water pretreatment system was 21–22 °C. At this temperature, the maximum ammonia oxidation ability in both reactors was observed in this study. As shown in Fig. 4(a–c), the daily periodic variation of operation performance was significantly different. The main factor is that the temperature was changed obviously, even shift from 20 to 30 °C. The daily periodic variation of NH+4-N removal also provides that nitrification had a better performance at around 21 °C in oligotrophic niche. In addition, the performance of raw water biological pretreatment of raw water was affected by the change in influent quality (Qin et al., 2008). Influent NH+4-N rate affected the removal performance of other pollutants. In stage P4, the removal ability of UV254 was improved comparing with the previous three periods (P1–P3). The average UV254 removal efficiencies of R1 and R2 were 7.1 ± 10.9% and 8.7 ± 10.6%, respectively, with maximum removal efficiencies of 26.2% and 27.0%, respectively. This phenomenon may be caused by DO limitation, and a low influent NH+4-N rate consumes little DO, and more DO could be used to removal UV254. Generally, a high influent NH+4-N rate is advantage to the growth of biomass. Zhang et al. (2013) reported that increasing influent NH+4-N rate could accelerate the accumulation rate of volatile solids in biofilm. In this study the biomass increase was also observed, but the attached biomass decreased rapidly in biofilm reactor even the NH+4-N rate increased to 2.4 times after long-term operation (Zhang et al., 2013). In the study, this phenomenon was not observed.

Table 6 The estimated THMs concentrations based on empiric models. THMs Concentration (lg L1)

Reactors

R1

R2

Influent Effluent Influent Effluent Influent Effluent Influent Effluent

Model

P1

P2

P31

P32

P4

25.6 23.8 22.7 22.8 25.6 16.9 22.7 23.5

19.9 24.0 24.2 25.0 19.9 19.2 24.2 24.3

16.4 21.9 24.0 27.9 16.4 25.3 24.0 29.7

16.3 16.5 24.9 23.5 16.3 15.0 24.9 23.3

23.8 18.6 25.1 22.8 23.8 19.3 25.1 20.3

Eq. Eq. Eq. Eq. Eq. Eq. Eq. Eq.

(4) (4) (5) (5) (4) (4) (5) (5)

4.3. Potential risk assessment of algae and it’s control strategy It was reported that photosynthetic oxygen production could replace mechanical aeration for saving of energy in wastewater treatment systems (Mallick, 2002). In this study, both reactors were oxygen saturation in the presence of algae. This caused the maximum NO 3 -N removal efficiency decreased from 77.1% to 69.4% for R1 and from 78.4% to 69.4% for R2 (Fig. 2c and d). Aeration could not be replaced by photosynthetic oxygen production of algae in raw water pretreatment due to the fact that aeration could reduce the DBPs precursor DOC and UV254. The organics level in drinking water had direct relationship to the formation of DBPs (Ates et al., 2007). Organic matter, especially the dissolved organic matter (DOM) was common in polluted raw water, which was considered as the precursors of DBPs (Panyapinyopol et al., 2005; Platikanov et al., 2010). Trihalomethane (THMs) are the common DBPs presented in chlorination (Xu et al., 2007), which could cause acute or chronic effects to human health (Villanueva et al., 2004; Chowdhury et al., 2009; Platikanov et al., 2010). Thus, the removal of organic matter could effectively reduce the risk of THMs formation. It was recommended that the chlorophyll-a level in raw water should be below 50 mg m3 (Bartram et al., 1999; Chen et al., 2009). In this study, the chlorophyll-a concentration in effluent were always below 12 mg m3, even when the chlorophyll-a level in both reactors accumulated to more than 200 mg m3 (Fig. 3b). However, the presentence of algae in the raw water pretreatment system increased the risk of more DBPs production. According to the empiric models for predicting THMs formation based on DOC and UV254 (Eq. (4) and (5), respectively), the predicted THMs levels of samples obtained from R1 and R2 were calculated at residue chlorine-Cl concentration of more than 1 mg L1 (Table 6). The simulation results of both empiric models showed that the algae growth in the reactors could enhance the risk of DBPs production. The methods of manual removing algae and light-shading were useful to reduce this risk. On day 101, the chlorophyll of effluent in both reactors was lower than detection limit (Fig. 3c). Chen et al. (2009) reported that algal in raw water could be controlled by light-shading within 6–9 days. In our study, this period is longer than the reported days by Chen et al. (2009). This may be because the effluent water quality in this study was significantly improved (p > 0.01). After the algae removal, the potential THMs production was obviously reduced (P4, Table 6). When both reactors operated at a higher DO level, the reduce amount was higher. Compared with the predicted THMs values reported in previous studies (Hong et al., 2008; Feng et al., 2013), the optimized process could successfully reduce the THMs risk. 5. Conclusions The application of carrier precoated biofilm improved the NH+4N removal and reduced NO 2 -N accumulation in the raw water

G.-f. Yang et al. / Bioresource Technology 172 (2014) 22–31

pretreatment process, which is advantage to the biomass enrichment. The removal efficiencies of NH+4-N, DOC and UV254 increased under the aeration condition, and the optimum DO level for the nitrification was 1.0–2.6 mg L1 with suitable temperature range of 21–22 °C. Study on the trihalomethane prediction model indicated that the presentence of algae increased the risk of disinfection by-products production, which could be controlled by the manual algae removing and light shading. Acknowledgements This work was funded by the National Key Technologies Research and Development Program of China (No. 2012BAJ25B07/2013BAC16B04). References Asiedu, K., 2001. Evaluating Biological Treatment Systems (Master thesis). Virginia Polytechnic Institute and State University, USA, 18–23. Ates, N., Kitis, M., Yetis, U., 2007. Formation of chlorination by-products in waters with low SUVA-correlations with SUVA and differential UV spectroscopy. Water Res. 41 (18), 4139–4148. Bartram, J., Burch, M., Falconer, I., Jones, G., Kuiper-Godman, T., 1999. Situation assessment, planning and management. In: World Health Organization (Ed.), Toxic Cyano Bacteria in Water. A Guide to their Public Health Consequences, Monitoring and Management. E & FN SPON, London. Benner, J., Helbling, D.E., Kohler, H.P.E., Wittebol, J., Kaiser, E., Prasse, C., Ternes, T.A., Albers, C.N., Aamand, J., Horemans, B., Springael, D., Walravens, E., Boon, N., 2013. Is biological treatment a viable alternative for micropollutant removal in drinking water treatment processes? Water Res. 47 (16), 5955–5976. Bruce, E.R., Douglas, S., 2002. Modelling biological processes in water treatment: the integrated biofiltration model. J. Water Supply: Res. Technol-AQUA 51 (1), 1– 14. Chen, X.C., Kong, H.N., He, S.B., Wu, D.Y., Li, C.J., Huang, X.C., 2009. Reducing harmful algae in raw water by light-shading. Process Biochem. 44, 357–360. Chinese, S.E.P.A., 2002. Water and Wastewater Monitoring Methods, fourth ed. Chinese Environmental Science Publishing House, Beijing, China. Chowdhury, S., Champagne, P., McLellan, P.J., 2009. Models for predicting disinfection byproduct (DBP) formation in drinking waters: a chronological review. Sci. Total Environ. 407, 4189–4206. Chu, H.Q., Cao, D.W., Dong, B.Z., Qiang, Z.M., 2010. Bio-diatomite dynamic membrane reactor for micro-polluted surface water treatment. Water Res. 44, 1573–1579. Chu, W.H., Gao, N.Y., Deng, Y., Templeton, M.R., Yin, D.Q., 2011. Impacts of drinking water pretreatments on the formation of nitrogenous disinfection by-products. Bioresour. Technol. 102 (24), 11161–11166. de Vet, W.W.J.M., Dinkla, I.J.T., Muyzer, G., Rietveld, L.C., van Loosdrecht, M.C.M., 2009. Molecular characterization of microbial populations in groundwater sources and sand filters for drinking water production. Water Res. 43, 182–194. Egli, 2010. How to live at very low substrate concentration. Water Res. 44, 4826– 4837. Fdz-Polanco, F., Villaverde, S., García, P.A., 1994. Temperature effect on nitrifying bacteria activity in biofilters: activation and free ammonia inhibition. Water Sci. Technol. 30 (11), 121–130. Feng, L.J., Xu, J., Xu, X.Y., Zhu, L., Xu, J., Ding, W., Luan, J., 2012. Enhanced biological nitrogen removal via dissolved oxygen partitioning and step feeding in a

31

simulated river bioreactor for contaminated source water remediation. Int. Biodeterior. Biodegradation 71, 72–79. Feng, L.J., Zhu, L., Yang, Q., Yang, G.F., Xu, J., Xu, X.Y., 2013. Simultaneous enhancement of organics and nitrogen removal in drinking water biofilm pretreatment system with reed addition. Bioresour. Technol. 129, 274–280. Han, M., Zhao, Z.W., Gao, W., Cui, F.Y., 2013. Study on the factors affecting simultaneous removal of ammonia and manganese by pilot-scale biological aerated filter (BAF) for drinking water pre-treatment. Bioresour. Technol. 145, 17–24. Hong, H.C., Wong, M.H., Mazumder, A., Liang, Y., 2008. Trophic state, natural organic matter content, and disinfection by-product formation potential of six drinking water reservoirs in the Pearl River Delta, China. J. Hydrol. 359 (1–2), 164–173. Jiao, Y., Zhao, Q.L., Jin, W.B., Hao, X.D., You, S.J., 2011. Bioaugmentation of a biological contact oxidation ditch with indigenous nitrifying bacteria for in situ remediation of nitrogen-rich stream water. Bioresour. Technol. 102 (2), 990– 995. Lee, S.Y., Cho, K.J., Lim, J.T., Kim, W., Hwang, S.H., 2011. Acclimation and activity of ammonia-oxidizing bacteria with respect to variations in zinc concentration, temperature, and microbial population. Bioresour. Technol. 102 (5), 4196–4203. Leung, S.M., Little, J.C., Holst, T., Love, N.G., 2006. Air/water oxygen transfer in a biological aerated filter. J. Environ. Eng. 132, 181–189. Mallick, N., 2002. Biotechnological potential of immobilized algae for wastewater N, P and metal removal: a review. Biometals 15, 377–390. Panyapinyopol, B., Marhaba, T.F., Kanokkantapong, V., Pavasant, P., 2005. Characterization of precursors to trihalomethanes formation in Bangkok source water. J. Hazard. Mater. 120, 229–236. Platikanov, S., Tauler, R., Rodrigues, P.M., Antunes, M.C., Pereira, D., Esteves da Silva, J.C., 2010. Factorial analysis of the trihalomethane formation in the reaction of colloidal, hydrophobic, and transphilic fractions of DOM with free chlorine. Environ. Sci. Pollut. Res. 17 (8), 1389–1400. Qian, J.Z., Wang, Z.P., Jin, S., Liu, Y., Chen, T.H., Fallgren, P.H., 2011. Nitrate removal from groundwater in columns packed with reed and rice stalks. Environ. Technol. 32 (14), 1589–1595. Qin, Y.Y., Zhang, X.W., Ren, H.Q., Li, D.T., Yang, H., 2008. Population dynamics of ammonia-oxidizing bacteria in an aerated submerged biofilm reactor for micropolluted raw water pretreatment. Appl. Microbiol. Biotechnol. 79, 135– 145. Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles and Applications. McGraw-Hill, New York, NY. Villanueva, C.M., Cantor, K.P., Cordier, S., Jaakkola, J.J.K., King, W.D., Lynch, C.F., Porru, S., Kogevinas, M., 2004. Disinfection byproducts and bladder cancer: a pooled analysis. Epidemiology 15 (3), 357–367. Warneke, S., Schipper, L.A., Matiasek, M.G., Scow, K.M., Cameron, S., Bruesewitz, D.A., McDonald, I.R., 2011. Nitrate removal, communities of denitrifiers and adverse effects in different carbon substrates for use in denitrification beds. Water Res. 45 (17), 5463–5475. Xu, B., Gao, N.Y., Sun, X.F., 2007. Characteristics of organic material in Huangpu River and treatability with the O3-BAC process. Sep. Purif. Technol. 57, 348– 355. Yu, X., Qi, Z.H., Zhang, X.J., Yu, P., Liu, B., Zhang, L.M., Fu, L., 2007. Nitrogen loss and oxygen paradox in full-scale biofiltration for drinking water treatment. Water Res. 41 (7), 1455–1464. Yu, Z.X., Chu, H.Q., Cao, D.W., Ma, Y.Q., Dong, B.Z., Wei, Y., 2012. Pilot-scale hybrid bio-diatomite/dynamic membrane reactor for slightly polluted raw water purification. Desalination 285, 73–82. Zhang, S.F., Wang, Y.Y., He, W.T., Wu, M., Xing, M.Y., Yang, J., Gao, N.Y., Daqiang Yin, D.Q., 2013. Responses of biofilm characteristics to variations in temperature and NH4+-N loading in a moving-bed biofilm reactor treating micro-polluted raw water. Bioresour. Technol. 131, 365–373. Zhu, L., Ding, W., Feng, L.J., Kong, Y., Xu, J., Xu, X.Y., 2012. Isolation of aerobic denitrifiers and characterization for their potential application in the bioremediation of oligotrophic ecosystem. Bioresour. Technol. 108, 1–7.