Bioresource Technology 151 (2014) 258–264
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Enhanced synergistic denitrification and chemical precipitation in a modified BAF process by using Fe2+ Hongjie Wang a,b, Wengyi Dong a,b, Ting Li a, Tongzhou Liu a,b,⇑ a b
Harbin Institute of Technology Shenzhen Graduate School, Shenzhen 518055, PR China Shenzhen Key Laboratory of Water Resource Utilization and Environmental Pollution Control, Shenzhen 518055, China
h i g h l i g h t s Synergistic denitrification and chemical precipitation of P and TSS in BAF process. Fe
2+
is suitable for dosing in the pre-denitrification stage of a BAF process. 2+ increased the denitrification loading rate in BAF pre-denitrification. 2+ 3+ A more efficient controlling of residual P was exhibited by dosing Fe than Fe . 2+ Dosing Fe greatly improved the settleability of BAF pre-denitrification sludge.
Dosing Fe
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 10 October 2013 Accepted 18 October 2013 Available online 29 October 2013 Keywords: BAF Synergistic denitrification and chemical precipitation Ferrous iron Residue phosphorus control Sludge settleability
a b s t r a c t A series of laboratory-scale experiments for examining the feasibility and suitability of using Fe2+ as the precipitant dosed in the pre-denitrification stage of a modified BAF process employing simultaneous chemical precipitation of TSS and phosphorus were carried out. The effects of dosing Fe2+ on effluent quality and sludge characteristics of the pre-denitrification stage were assessed with comparing to the cases of no additional chemical dosing and dosing Fe3+. Results obtained demonstrated a sound performance of synergistic denitrification and chemical precipitation in pre-denitrification of the modified BAF process when dosing Fe salts, which showed enhanced by using Fe2+ as the dosed precipitant in increasing the denitrification loading rate, exhibiting a better controlling of the residual phosphorus in pre-denitrification effluent, and improving sludge settleability. Dosing Fe salt showed no adverse impact in removing COD, but resulted in a relatively higher SS content in the pre-denitrification effluent. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Biological aerated filter (BAF) process, which relies on attached growth of biomass on inert granular medium within compact units, is an alternative to the traditional activated sludge process (Mendoza-Espinosa and Stephenson, 1999). Since it possesses a number of advantages, including achieving solids, organic matter, and nutrient removal in a same reactor, higher pollutants loading, and hence reduced space requirement, researches, and applications of BAF have grown with great popularity in recent years (Gilbert et al., 2008; Zhao et al., 2009; González-Martínez et al., 2010; Bai et al., 2011; Abu Hasan et al., 2012; Li et al., 2013). Because solids removal in BAF is achieved mainly through filtration, total suspended solids (TSS) in BAF influent is usually below ⇑ Corresponding author at: Harbin Institute of Technology Shenzhen Graduate School, Shenzhen 518055, China. Tel./fax: +86 755 2603 2718. E-mail address: [email protected] (T. Liu). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.10.066
100 mg/L (Canler and Perret, 1994) to avoid filter bed clogging and associated frequent backwashing. To deal with the high suspended solids content in influent streams, chemical coagulation/ precipitation is an often applied pre-treatment step for removing TSS (Farabegoli et al., 2009). Additionally, due to the lack of alternate anaerobic and aerobic conditions, BAF is incapable of enhanced biological phosphorus removal (Lee et al., 2005), and supplementary physio-chemical precipitation of phosphorus is often used in BAF plants for meeting stringent effluent standards (Rogalla et al., 1990; Clark et al., 1997). Though an enhanced chemical pre-treatment step can reduce both TSS and total phosphorus in the influent stream fed into BAF, removal of considerable amount of organic matter can also be resulted at the same time, which reportedly could reach up to over 50% COD removal (Rogalla et al., 1990). The simultaneous organic matter removal in the chemical pre-treatment step might lead to the shortage of carbon sources in the post-denitrification process for BAF nitrified effluent, particularly when treating weak sewage with relatively low C/N
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
ratio. In order to meet the total nitrogen discharge requirement in the effluent, an external carbon source has to be added in the denitrification process, and hence operating cost is increased. In this study, a modified BAF process employing simultaneous chemical precipitation in pre-denitrification is proposed. A schematic diagram of the proposed process is shown in Fig. 1. When dosing chemicals in the pre-denitrification tank, synergistic denitrification, and removal of phosphorus and TSS through chemical precipitation is anticipated. Simultaneous chemical precipitation in biological processes can utilize the advantages of biological and chemical methods for removing nutrients (Philips et al., 2003; Caravelli et al., 2010), and offer benefits in ease of operation, flexibility to changing conditions, and relatively small additional solids production (De Haas et al., 2000). Furthermore, parts of organic matter content in the influent can be utilized by the denitrifiers in the denitrification/flocculation tank, avoiding the problem of carbon source shortage occurring in biological processes implemented with a normal chemical pre-treatment step. To obtain sound treatment effectiveness and control operating costs, the selection of suitable dosed precipitants is essential for this proposed synergistic denitrification and chemical precipitation process. Iron and aluminum salts are currently the most used precipitants in sewage treatment. Al salts are relatively costly, and residual Al in effluent would arouse concerns on the risk of neurologic diseases potentially associated with long-term exposure to Al in environments (Willhite et al., 2012). Ferric chloride and ferrous sulfate are the most used Fe salts. When dosed in oxic reactors, one principal disadvantage of using Fe2+ salt is the consumption of dissolved oxygen due to oxidation of Fe2+ to Fe3+. Fe3+ salt is however much more expensive than Fe2+ salt. Additionally, the pH value of the mixed liquor in denitrification process is around pH 7.5, and the optimum pH ranges for phosphorus precipitation with Fe3+ and Fe2+ are between pH 4 and 5 and close to pH 8, respectively. Hence, if dosing Fe salts in the proposed synergistic denitrification and chemical precipitation process, pH condition would be favorable for Fe2+ rather than Fe3+ to precipitate phosphorus. Despite Fe2+ is considered potentially a suitable precipitant being used in the proposed process due to its relatively low cost and high phosphorus removal efficiency under the typical pH condition of denitrification, an uncertainty still lies in the possible smaller size of Fe2+ dosed sludge flocs compared to Fe3+ dosed ones (Oikonomidis et al., 2010), which might adversely affect liquid–solid separation performance in the settling tank. To date, report on applying Fe2+ salt in the pre-denitrification stage of BAF process to achieve simultaneous chemical precipitation of TSS and phosphorus remains few. To provide basic information for modifying a
Effluent
Recycled Effluent
M
Denitrification/Flocculation Tank
259
BAF process with employing simultaneous pre-nitrification and chemical precipitation, this study aims to examine the feasibility and suitability of using Fe2+ salt as the precipitant dosed in the pre-denitrification stage of a BAF process. The effects of dosing Fe2+ salt on effluent quality and sludge characteristics were monitored, and comparisons with the cases of no additional chemical dosing and dosing Fe3+ salt were conducted to investigate any adverse effect introduced by dosing Fe2+ salt. 2. Methods 2.1. Experimental setup A series of laboratory scale experiments were conducted in 25 L rectors to examine the performance of dosing chemicals in the denitrification/flocculation tank shown in Fig. 1. Three sets of experiments were carried out (see Table 1). The reference experiment (R1) was performed to examine the performance of pre-denitrification without dosing any additional chemical. Another two sets of experiments were dosed with Fe2+ salt (R2) and Fe3+ salt (R3), respectively, to investigate the effects of applying chemical precipitation in the pre-denitrification process. A synthetic domestic wastewater was used as the experimental influent. It was prepared by diluting methanol, KNO3, and KH2PO4 stock solutions with deoxygenated tap water to simulate the mixed liquor of sewage influent and recycled BAF nitrified effluent, whose chemical oxygen demand (COD), NO x —NðNO3 — N þ NO2 —NÞ, and total phosphorus (TP) concentrations, and pH value were within their typical ranges. Initial activated sludge used in the reactor was collected in a municipal wastewater treatment plant (Luofang Wastewater Treatment Plant, Shenzhen, China). The experiments were continuously carried out in cycles for more than 70 days. In each cycle, 18 L of the simulating solution was firstly fed into and mixed with the sludge already inside the reactor. For the experiments dosed with Fe2+ or Fe3+ salt, certain amount of FeSO4 or FeCl3 stock solution was added, respectively, to achieve a final Fe concentration of 24 mg/L. In a preliminary experiment (data not shown), this Fe concentration was justified sufficient to control the residual phosphorus concentration in the effluents of Fe dosed reactors (R2 and R3) to below 1 mg/L. The mixed liquor in the reactor was mechanically stirred for 9 h, and then settled down for 0.5 h for liquid–solid separation. Supernatant after liquid–solid separation was discharged. Prior to starting next cycle, the sludge remained in the reactor was rinsed by deoxygenated tap water for 5 times to minimize residual nutrient accumulation. The whole operation process for each cycle took about 12 h. About 1 L of the mixed liquor in the reactors, or the supernatant after liquid–solid separation were regularly sampled for monitoring effluent quality and sludge characteristics, respectively. Initial sludge concentrations of the experiments were about 1000 mg/L. Within the time span of the experiment, no excess sludge was discharged. 2.2. Materials
Settling Tank
BAF
Chemical stock solutions were prepared by dissolving reagent grade chemicals, methanol, KNO3, KH2PO4, FeCl36H2O, and FeSO47H2O, into ultrapure water (>18.1 MX cm) (deoxygenated for preparing Fe2+ stock solution). All stock solutions were stored in 4 °C cold room. Fe2+ stock solution was acidified by sulfuric acid to pH 3 prior to storage.
Influent
Return Sludge
Chemical Dosing
2.3. Analytical methods Fig. 1. The schematic diagram of a modified BAF process with applying simultaneous chemical precipitation in pre-denitrification.
Water quality parameters are analyzed following APHA Standard Method (APHA, 1998). COD and alkalinity were analyzed by
260
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
Table 1 Effluent quality of the rectors without dosing addition chemical and with dosing Fe2+ and Fe3+.a Parameters monitored
NO x —N Total phosphorous COD Suspended solid pH Alkalinity
Unit
mg/L mg/L mg/L mg/L – mg/L (as CaCO3)
Influent
41.7 ± 5.0 4.93 ± 0.30 239.8 ± 40.0 – 7.13 ± 0.30 90 ± 20
Effluent R1 (Reference experiment)
R2 (Experiment dosed with Fe2+)
R3 (Experiment dosed with Fe3+)
0.20 ± 0.10 4.47 ± 0.32 35.5 ± 25.4 11.1 ± 7.0 7.83 ± 0.22 254 ± 25
0.10 ± 0.10 0.50 ± 0.25 36.1 ± 28.2 24.5 ± 10.8 7.40 ± 0.31 219 ± 19
0.11 ± 0.10 0.93 ± 0.25 43.3 ± 21.8 15.7 ± 9.8 7.38 ± 0.33 209 ± 21
a The data showed in the table correspond to mean values of 13 events of determination of wastewater quality parameters during the experimental time, and the associated uncertainties are the standard deviations.
The specific denitrification rates of the mixed liquors in the three reactors (R1, R2, and R3) were determined regularly. 1 L of the mixed liquor withdrawn from the reactor was mixed with a certain amount of KNO3 stock solution to achieve a final NO 3 —N concentration of 40 mg/L, and kept stirred using a magnetic stirrer. Duplicate aliquots of the mixed liquor were sampled periodically for determining NO x —N concentration. The specific denitrification rate of the mixed liquor was obtained by dividing the slope of the plot of NO x —N concentration versus time by its MLVSS content. 2.5. Statistical analysis The student’s t-test was used to statistically evaluate the significance of a difference of two data sets. A confidence level of 95% was used as the cutoff for statistical significance. 3. Results and discussion 3.1. Effects of dosing Fe2+ on effluent quality 3.1.1. Denitrification Effluent quality parameters reflecting the performance of R1 (reference reactor), R2 (Fe2+ dosed), and R3 (Fe3+ dosed) are shown in Table 1. Sound denitrification performance was recorded in all the three reactors. Effluent NO x —N concentrations of the three reactors kept below 0.5 mg/L throughout the course of the experiment. Study of the specific denitrification rates (with the unit of
(a) Specific Denitrification Rate (mg/h.gVSS)
2.4. Determination of denitrification rate
mg/h gVSS) of the mixed liquors in different reactors showed that, in the steady-state period (after day 25 of the experiment) and comparing with the specific denitrification rate of R1, R2 showed no significant difference (Fig. 2a), whereas, R3 appeared a little bit slower (student’s t-test at 95% confidence). A conversion of specific denitrification rate to denitrification loading rate (with the unit of mg/h L) was conducted by multiplying the data of the specific denitrification rate with the MLVSS concentration in each reactor. As shown in Fig. 2b, the converted denitrification loading rate in R2 was found apparently faster than that in R1. In contrasting, little difference was observed between R3 and R1. The
80
60
40
R1 R2 R3
20
0
0
10
20
30
40
50
60
70
Experiment Time (day)
(b) 80 Denitrification Loading Rate (mg/h.L)
titrimetric method; NO 3 —N, NO2 —N, and TP were determined by colorimetric method using a UV/visible spectrophotometer (UV 2450, Shimadzu); TSS, mixed liquor suspended solids (MLSS), and mixed liquor volatile suspended solids (MLVSS) concentrations were determined following method outlined in Section 2540 D of the Standard Method (APHA, 1998), and pH was measured using a pH meter. All the determinations of the examined parameters were carried out for duplicated samples. On day 40 of the experiment, samples of the mixed liquor in the three reactors (R1, R2, and R3) were withdrawn and subject to particle size distribution analysis by a laser diffraction particle size analyzer (Malvern Mastersizer 2000), and SEM-EDS (Hitachi S-4700) analysis after freeze drying. Sludge settling characteristics were assessed periodically during the experiment through zone settling test and settled sludge volume test (expressed in the term of SV30) according to methods outlined in Section 2710 C and E of the Standard Method (APHA, 1998). The dewaterability of sludge was measured regularly by the Buchner funnel test, and expressed in the term of specific resistance in filtration (SRF). The Particle size distribution analysis and the sludge zone settling test were performed for a single sample. The settled sludge volume test and the dewaterability test were conducted in duplicate.
60
40
R1 R2 R3
20
0
0
10
20
30
40
50
60
70
Experiment Time (day) Fig. 2. Changes of (a) the specific denitrification rate and (b) the denitrification loading rate along with experiment time in the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
magnitude of denitrification loading rate was mathematically determined by the values of specific denitrification rate and MLVSS concentration. An elevation of MLVSS concentrations in R2 (average 32.9% increase) and R3 (average 18.1% increase) were observed compared to that in R1 (Fig. 3a), it was however very likely resulted from increased adsorption/enmeshment of organic matters in colloidal form by Fe hydroxides complexes formed in the mixed liquor (De Haas et al., 2001). Such adsorped/enmeshed organic matters are not expected to contribute for biotic denitrification. Thus, even the MLVSS concentration in R3 was higher, the denitrification loading rate of R3 remained almost unchanged comparing to that of R1. In contrast to R3, the denitrification loading rate of R2 showed apparently increased. In the steady-state period (after day 25 of the experiment) and compared to R1, an average 34.8% increase in denitrification loading rate was recorded in R2, indicating an enhanced denitrification when dosing Fe2+. A predominantly microbial Fe2+-dependent nitrate reduction process with high rate is believed to be responsible for the enhanced denitrification resulting from dosing Fe2+. Straub et al. (1996) reported that microbes might be important for anoxic oxidation of Fe2+ with nitrate in environmental systems. Their observation was further confirmed by Nielsen and Nielsen (1998), whose work demonstrated the occurrence of microbial Fe2+dependent nitrate removal in activated sludge simultaneously with the oxidation of Fe2+ to Fe3+, which resulted in an increase
(a) 1400 1200
MLVSS (mg/L)
1000 800 600 400
R1 R2 R3
200 0
0
10
20
30
40
50
60
70
Time (day)
(b) 3000
MLSS (mg/L)
2500
2000
1500
1000
R1 R2 R3
500
0
0
10
20
30
40
50
60
70
Time (day) Fig. 3. Changes of the concentrations of (a) MLVSS and (b) MLSS along with experiment time in the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
261
in denitrification rate, ranged from 3% to 107%, compared to background organic matter-dependent denitrification. The results of the presented work together with previous studies clearly indicated the contribution of the microbial Fe2+-dependent nitrate reduction to the denitrification process in addition to the organic matter-dependent nitrate removal. Seen from an engineering view, the enhancement on denitrification signified an advantage of using Fe2+ over dosing no chemical or Fe3+ in the proposed synergistic denitrification and chemical precipitation process in increasing denitrification loading, and hence reducing the volume of the pre-denitrification/flocculation tank of a modified BAF treatment facility. But the detailed mechanism of dosing Fe2+ on the enhanced denitrification is not well described so far, and needs further study. 3.1.2. Phosphorus removal Since there was no alternating anaerobic/aerobic process, enhanced biological phosphorus removal cannot be achieved in a normal denitrification reactor, and hence poor phosphorus removal was recorded in R1 (reference reactor), with an average phosphorus removal rate of only 9.3%. On the contrary, good phosphorus removal was observed in both R2 (Fe2+ dosed) and R3 (Fe3+ dosed), where the average phosphorus removal rates were 89.9% and 81.1%, respectively. Fe:P molar ratio applied in both R2 and R3 in this study was 2.7:1, which is higher than the reported required molar ratios of Fe2+:P (2–2.4:1) (Clark et al., 2000; Banu et al., 2008) and Fe3+:P (1.5–1.9:1) (Clark et al., 2000; Caravelli et al., 2010) applied in activated sludge systems where the residual phosphorus concentrations in the effluents were decreased to below 1 mg/L. The high phosphorus removal rates recorded in R2 and R3 should mainly result from chemical precipitation of sparingly soluble phosphates by dosed Fe salts. It was evidenced by SEM-EDS results that atomic concentrations of both Fe and P were much higher in the dried sludge collected in R2 and R3 than that collected in R1 (Supplementary data Fig. S1). Though high phosphorus removal rates were achieved in R2 and R3, R2 showed a better controlling of the effluent residual phosphorus than R3 (student’s t-test at 95% confidence, data not shown). The average effluent phosphorus concentration in R3 was 0.93 mg/L; whereas, that in R2 was 0.50 mg/L. Such difference in controlling of the residual phosphorus in R2 and R3 was considered reflecting the pH dependence of phosphorus removal by chemically precipitation with Fe2+ or Fe3+. pH values of the mixed liquors in R2 and R3 were in the range of 7.1–7.4. The optimum pH range for phosphorus precipitation with Fe3+ is between pH 4 and 5 (Stumm and Morgan, 1996); whereas, maximum phosphorus removal by Fe2+ was reportedly obtained in the vicinity to pH 8 (USEPA, 1971). The higher efficiency of phosphorus precipitation with Fe2+ at a higher pH should attribute to the lower tendency to hydrolyze of Fe2+ than Fe3+ (Stumm and Morgan, 1996), and thereby diminishing the competition between hydroxide and orthophosphate ions for reacting with Fe2+. The better performance in controlling the residue phosphorus in R2 than R3 announced another advantage of using Fe2+ over Fe3+ in the pre-denitrification stage in keeping effluent phosphorus at a lower level, where alkalinity production occurs and typical pH ranges from 7 to 8. 3.1.3. COD, SS, pH, and alkalinity As shown in Table 1, average COD removal rates of 87.9%, 84.9%, and 81.9% were achieved in R1, R2, and R3, respectively. The similar good COD removal performance in the three reactors showed that dosing Fe salts has no adverse impact in removing organic matters. The effluent suspended solids (SS) content in R2 was observed apparently greater than that in R1, and the effluent SS in R3 was slightly higher than that in R1. Flocs formed in activate sludge systems dosed with Fe2+ often appear compact and in
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
smaller size, and may partly be washed out to the effluent, resulting in relatively poor performance in solids removal (Thistleton et al., 2001; Oikonomidis et al., 2010). In this study, the reduction in sludge floc size caused by dosing Fe salts was evidenced through particle size distribution analysis of the mixed liquor collected from the three rectors (Supplementary data Fig. S2). Dosing Fe salts caused d50 of sludge flocs to decrease. d50 of sludge flocs in R1 was 72.3 lm, and in R2 and R3, d50 decreased to 12.9 and 36.2 lm, respectively. Thus, the descending variation in sludge flocs particle size resulted from dosing Fe2+ (R2) or Fe3+ (R3) should be responsible for the higher effluent SS content compared to the experiment without additional chemical dosing (R1). The effluent pH and alkalinity in R1, R2, and R3 were increased compared to the influent (Table 1) due to alkalinity production in the denitrification process. But the elevations of pH and alkalinity in R2 and R3 were not as high as that in R1. It should be attributed to the hydrolysis of Fe took place in R2 and R3, which leads to the formation of Fe hydroxides resulting in alkalinity consumption and an observed pH lowering in R2 and R3 compared to R1.
(a)
60
0 9 25 42 70
40
0
0
5
10
15
20
25
30
25
30
Time (min)
(b)
100 Day Day Day Day Day
Height of Interface (%)
80
60
0 9 25 42 70
40
20
0
0
5
10
15
20
Time (min)
(c) 100 Day Day Day Day Day
80
Height of Interface (%)
3.2.2. Sludge settleability The settling characteristics of sludge in the three reactors were assessed by zone settling test and settled sludge volume test (expressed in the term of SV30) in this study, and the results are shown in Figs. 4 and 5, respectively. Sludge settling process can be generally divided into three stages, those are hindered settling stage, transitional settling stage, and compression settling stage. The duration of the former two stages is crucial for determining the residence time in a sludge settling tank. As shown in Fig. 4a, there was no significant change in the settleability of the sludge in R1 along with experimental time, where the hindered settling stage ended at about 4 min, and the transitional settling stage lasted till around 10 min. In contrasting, sludge settleability was observed significantly improved in R2 (Fe2+ dosed) (Fig. 4b) in the steady-state period (after day 25 of the experiment). The time span for the hindered settling stage of the sludge in R2 was shortened to about 2 min, and the transitional settling stage ended at about 5 min. Zone settling test curves of the sludge in R3 (Fe3+ dosed) (Fig. 4c) were similar as those of R1. As shown in Fig. 5, SV30 of the sludge in R1 did not changed significantly throughout the experiment with an average value of 26.8%. Compared to that of R1, in the steady-state period (after day 25 of the experiment), SV30 of the sludge in R3 showed a little bit decrease, and an average SV30 value of 23.9% was recorded; whereas, SV30 of the sludge in R2 was observed remarkably lower, and its average value was only 10.4%. The much better settleability exhibited by Fe2+ dosed sludge in this study was in agreement with what was reported in literatures (Clark et al., 2000; Muller, 2001; Oikonomidis et al., 2010), and is suspected related to more compact flocs of this sludge. SEM analysis on the freeze-dried sludge revealed that Fe2+ dosed sludge floc
Day Day Day Day Day
20
3.2. Effects of dosing Fe2+ on sludge characteristic 3.2.1. Sludge production As discussed in the previous section, average MLVSS concentrations in R2 and R3 were 32.9% and 18.1% greater than that in R1 (Fig. 3a), which should result from increased adsorption/enmeshment of organic matters in colloidal form by Fe hydroxides complexes formed after dosing Fe salts (De Haas et al., 2001). In terms of TSS (i.e. MLSS), a significant increase (more than one fold increase in average) in sludge production occurred in R2 and R3 compared to R1 (Fig. 3b). These results were expected for additional phosphorus removal through the way of chemical precipitation. As a result, the chemical precipitates would also contribute to sludge production in the form of inorganic suspended solids.
100
80
Height of Interface (%)
262
60
0 9 25 42 70
40
20
0
0
5
10
15
20
25
30
Time (min) Fig. 4. Zone settling test curves of the mixed liquors collected from (a) the reactor without dosing addition chemical (R1), (b) the reactor dosed with Fe2+ (R2), and (c) the reactor dosed with Fe3+ (R3).
displayed more compact and was featured by relatively smooth surface; whilst, rough and open structures were observed on the surfaces of sludge flocs either without dosing additional chemical or dosed with Fe3+ (Supplementary data Fig. S1). The smooth surface features (Supplementary data Fig. S1b) is believed to be condensed and less hydrated extracellular polymeric substances (EPS), and the rough and open structures (Supplementary data Fig. S1a and c) is attributed to more hydrated EPS matrix having a relative high water-binding capacity (Liss et al., 2002). One
263
H. Wang et al. / Bioresource Technology 151 (2014) 258–264 60
35 30
SRF (10 m/kg)
20
12
SV30(% )
25
15 10 R1 R2 R3
5 0
R1 R2 R3
50
0
10
40
30
20
10
20
30
40
50
60
70
0
0
20
40
60
Time (day)
Time (day) Fig. 5. Changes of the settled sludge volume in 30 min (SV30) along with experiment time of the mixed liquors collected from the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
Fig. 6. Changes of the sludge specific resistance in filtration (SRF) along with experiment time of the mixed liquors collected from the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
mechanism suggested to explain the formation of the compact sludge flocs by dosing Fe2+ is that the dosed Fe2+ reacts faster than Fe3+ with the negatively charged sludge colloidal materials; and subsequently, Fe3+ engendered in situ from Fe2+ oxidation would form tight binding with the colloidal materials and serve to decrease the specific area of sludge floc and make the flocs more compact and denser (Muller, 2001). It was also proposed that the major interaction mechanism of the dosed Fe2+ with sludge flocs is cation bridging which would stabilize the negatively charged groups present on the EPS and promote sludge flocculation (Oikonomidis et al., 2010). On the contrary, when the dosed precipitant is Fe3+, the formation of ferric hydroxide (Fe(OH)3) would occur rapidly under the pH condition (around pH 7.5) of the denitrification process. Fe(OH)3 present in colloidal form would have low neutralization capacity towards the negatively charged biopolymers in sludge flocs, and hence result in a more open floc structure and limited improvement on the sludge settleability compared to dosing Fe2+ (Oikonomidis et al., 2010). As for its engineering implication, the remarkably improved settleability of sludge in R2 displayed one more advantage of using Fe2+ over Fe3+ in the proposed synergistic denitrification and chemical precipitation process in shortening the residence time, and in turn, reducing the volume of a sludge settling tank.
(Liu and Fang, 2003; Park et al., 2006), and the bound EPS showed improving sludge dewaterability (Higgins and Novak, 1997; Murthy and Novak, 1999). Though a better settleability was observed for Fe2+ dosed sludge than Fe3+ dosed one in this study, the sludge dewatering performance was not found affected by which Fe species was dosed. Employing either Fe2+ or Fe3+ in the proposed synergistic denitrification and chemical precipitation process would finally result in an enhancement in sludge dewatering and conduce to subsequent sludge processing compared to no additional chemical dosing.
3.2.3. Sludge dewaterability Dewatering performance of activated sludge is commonly characterized in the term of specific resistance in filtration (SRF). The larger SRF value indicates the worse sludge dewatering performance. Fig. 6 shows the changes of SRF values of the sludge in the three reactors (R1, R2, and R3) along with experimental time. SRF values of the sludge in R1 kept relatively high throughout the experimental time, and with an average value of 38.8 1012 m/kg. In the steady-state period (after day 25 of the experiment), SRF values of the sludge in both R2 (Fe2+ dosed) and R3 (Fe3+ dosed) kept around 2 1012 m/kg which was significantly decreased compared to in R1. These results are of expectation. Despite the lack of a clear understanding, sludge dewatering could be affected by a number of EPS-related parameters, such as floc size, structure, and bound water content (Liu and Fang, 2003). The increasing in bound EPS resulting from the addition of multivalent cations has been reported in numerous studies
4. Conclusion The results of this study demonstrate the feasibility and suitability of using Fe2+ as the precipitant dosed in the pre-nitrification stage of a modified BAF process for synergistic denitrification and chemical precipitation of TSS and phosphorus. Dosing Fe2+ exhibited strengths over dosing Fe3+ or no chemical in increasing the denitrification loading rate, having a better controlling of residual phosphorus in the pre-denitrification effluent, and improving sludge settleability. Findings obtained provide engineering implications in reducing the volumes of the pre-denitrification tank and the associated settling tank, and ultimately, the space and capital costs required for a modified BAF process employing predenitrification. Acknowledgements This work is supported by research projects from National science and technology major project of China for water pollution control (Grant No. 2012ZX07206002-02), Ministry of Housing and Urban-Rural Development of China (Grant No. 2012-K7-18), and Shenzhen Technology Innovation Plan (Grant No. CXZZ201303 19100941767). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2013. 10.066.
264
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
References Abu Hasan, H., Sheikh Abdullah, S.R., Tan Kofli, N., Kamarudin, S.K., 2012. Effective microbes for simultaneous bio-oxidation of ammonia and manganese in biological aerated filter system. Bioresour. Technol. 124, 335–363. APHA, 1998. Standard Methods for the Examination of Water and Wastewater, twentieth ed. American Public Health Association, Washington, DC. Bai, Y., Sun, Q., Sun, R., Wen, D., Tang, X., 2011. Bioaugmentation and adsorption treatment of coking wastewater containing pyridine and quinoline using zeolite-biological aerated filters. Environ. Sci. Technol. 45, 1940–1948. Banu, J.R., Do, K., Yeom, I., 2008. Effect of ferrous sulphate on nitrification during simultaneous phosphorus removal from domestic wastewater using a laboratory scale anoxic/oxic reactor. World J. Microb. Biot. 24, 2981–2986. Canler, J.P., Perret, J.M., 1994. Biological aerated filters: assessment of the process based on 12 sewage treatment plants. Water Sci. Technol. 29, 13–22. Caravelli, A.H., Contreras, E.M., Zaritzky, N.E., 2010. Phosphorous removal in batch systems using ferric chloride in the presence of activated sludges. J. Hazard. Mater. 177, 199–208. Clark, T., Stephenson, T., Pearce, P.A., 1997. Phosphorus removal by chemical precipitation in a biological aerated filter. Water Sci. Technol. 31, 2557–2563. Clark, T., Burgess, J.E., Stephenson, T., Arnold-Smith, A.K., 2000. The influence of iron-based co-precipitant on activated sludge biomass. Process Saf. Environ. 78, 405–410. De Haas, D.W., Wentzel, M.C., Ekama, G.A., 2000. The use of simultaneous chemical precipitation in modified activated sludge systems exhibiting biological excess phosphate removal. Part 1: literature review. Water SA 26, 439–452. De Haas, D.W., Wentzel, M.C., Ekama, G.A., 2001. The use of simultaneous chemical precipitation in modified activated sludge systems exhibiting biological excess phosphate removal. Part 5: experimental periods using a ferrous-ferric chloride blend. Water SA 27, 117–134. Farabegoli, G., Chiavola, A., Rolle, E., 2009. The biological aerated filter (BAF) as alternative treatment for domestic sewage. Optimization of plant performance. J. Hazard. Mater. 171, 1126–1132. Gilbert, Y., Bihan, Y.L., Aubry, G., Veillette, M., Duchaine, C., Lessard, P., 2008. Microbiological and molecular characterization of denitrification in biofilters treating pig manure. Bioresour. Technol. 99, 4495–4502. González-Martínez, S., Piña-Mondragón, S., González-Barceló, Ó., 2010. Treatment of the azo dye direct blue 2 in a biological aerated filter under anaerobic/aerobic conditions. Water Sci. Technol. 61, 789–796. Higgins, M.J., Novak, J.T., 1997. The effect of cations on the settling and dewatering of activated sludges: laboratory results. Water Environ. Res. 69, 215–224. Lee, J., Kim, J., Lee, C., Yun, Z., Choi, E., 2005. Biological phosphorus and nitrogen removal with biological aerated filter using denitrifying phosphorus accumulating organism. Water Sci. Technol. 52, 569–578.
Li, Q., Sun, S., Guo, T., Yang, C., Song, C., Geng, W., Zhang, W., Feng, J., Wang, S., 2013. Short-cut nitrification in biological aerated filters with modified zeolite and nitrifying sludge. Bioresour. Technol. 136, 148–154. Liss, S.N., Liao, B.Q., Droppo, I.G., Allen, D.G., Leppard, G.G., 2002. Effect of solids retention time on floc structure. Water Sci. Technol. 46, 431–438. Liu, Y., Fang, H.H.P., 2003. Influence of extracellular polymeric substances (EPS) on flocculation, settling, and dewatering of activated sludge. Crit. Rev. Env. Sci. Technol. 33, 237–273. Mendoza-Espinosa, L., Stephenson, T., 1999. A review of biological aerated filters (BAFs) for wastewater treatment. Environ. Eng. Sci. 16, 201–216. Muller, C.D., 2001. High-Intensity Shear as a Wet Sludge Disintegration Technology and a Mechanism for Floc Structure Analysis (M.S. thesis). Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Murthy, S.N., Novak, J.T., 1999. Factors affecting floc properties during aerobic digestion: implication for dewatering. Water Environ. Res. 71, 197–202. Nielsen, J.L., Nielsen, P.H., 1998. Microbial nitrate-dependent oxidation of ferrous iron in activated sludge. Environ. Sci. Technol. 32, 3556–3561. Oikonomidis, I., Burrows, L.J., Carliell-Marquet, C.M., 2010. Mode of action of ferric and ferrous iron salts in activated sludge. J. Chem. Technol. Biotechnol. 85, 1067–1076. Park, C., Muller, C.D., Mohammad, M.A., Novak, J.T., 2006. The effect of wastewater cations on activated sludge characteristics: effects of aluminum and iron in floc. Water Environ. Res. 78, 31–40. Philips, S., Rabaey, K., Verstraete, W., 2003. Impact of iron salts on activated sludge and interaction with nitrite or nitrate. Bioresour. Technol. 88, 229–239. Rogalla, F., Payraudeau, M., Bacquet, G., Bourbigot, M., Sibony, J., Gilles, P., 1990. Nitrification and phosphorus precipitation with biological aerated filters. Res. J. Water Pollut. Control Fed. 62, 169–176. Straub, K.L., Benz, M., Schink, B., Widdel, F., 1996. Anaerobic, nitrate-dependent microbial oxidation of ferrous iron. Appl. Environ. Microbiol. 62, 1458–1460. Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, third ed. Wiley-Interscience, New York. Thistleton, J., Clark, T., Pearce, P., Parsons, S.A., 2001. Mechanisms of chemical phosphorus removal 1 – Iron(II) salts. Process Saf. Environ. 79, 339–344. USEPA, 1971. Water pollution control research series: phosphate precipitation with ferrous iron. 17010 EKI 09/71. Willhite, C.C., Ball, G.L., McLellan, C.J., 2012. Total allowable concentrations of monomeric inorganic aluminum and hydrated aluminum silicates in drinking water. Crit. Rev. Toxicol. 42, 358–442. Zhao, Y., Yue, Q., Li, R., Yue, M., Han, S., Gao, B., Li, Q., Yu, H., 2009. Research on sludge-fly ash ceramic particles (SFCP) for synthetic and municipal wastewater treatment in biological aerated filter (BAF). Bioresour. Technol. 100, 4955–4962.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Enhanced synergistic denitrification and chemical precipitation in a modified BAF process by using Fe2+ Hongjie Wang a,b, Wengyi Dong a,b, Ting Li a, Tongzhou Liu a,b,⇑ a b
Harbin Institute of Technology Shenzhen Graduate School, Shenzhen 518055, PR China Shenzhen Key Laboratory of Water Resource Utilization and Environmental Pollution Control, Shenzhen 518055, China
h i g h l i g h t s Synergistic denitrification and chemical precipitation of P and TSS in BAF process. Fe
2+
is suitable for dosing in the pre-denitrification stage of a BAF process. 2+ increased the denitrification loading rate in BAF pre-denitrification. 2+ 3+ A more efficient controlling of residual P was exhibited by dosing Fe than Fe . 2+ Dosing Fe greatly improved the settleability of BAF pre-denitrification sludge.
Dosing Fe
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 10 October 2013 Accepted 18 October 2013 Available online 29 October 2013 Keywords: BAF Synergistic denitrification and chemical precipitation Ferrous iron Residue phosphorus control Sludge settleability
a b s t r a c t A series of laboratory-scale experiments for examining the feasibility and suitability of using Fe2+ as the precipitant dosed in the pre-denitrification stage of a modified BAF process employing simultaneous chemical precipitation of TSS and phosphorus were carried out. The effects of dosing Fe2+ on effluent quality and sludge characteristics of the pre-denitrification stage were assessed with comparing to the cases of no additional chemical dosing and dosing Fe3+. Results obtained demonstrated a sound performance of synergistic denitrification and chemical precipitation in pre-denitrification of the modified BAF process when dosing Fe salts, which showed enhanced by using Fe2+ as the dosed precipitant in increasing the denitrification loading rate, exhibiting a better controlling of the residual phosphorus in pre-denitrification effluent, and improving sludge settleability. Dosing Fe salt showed no adverse impact in removing COD, but resulted in a relatively higher SS content in the pre-denitrification effluent. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Biological aerated filter (BAF) process, which relies on attached growth of biomass on inert granular medium within compact units, is an alternative to the traditional activated sludge process (Mendoza-Espinosa and Stephenson, 1999). Since it possesses a number of advantages, including achieving solids, organic matter, and nutrient removal in a same reactor, higher pollutants loading, and hence reduced space requirement, researches, and applications of BAF have grown with great popularity in recent years (Gilbert et al., 2008; Zhao et al., 2009; González-Martínez et al., 2010; Bai et al., 2011; Abu Hasan et al., 2012; Li et al., 2013). Because solids removal in BAF is achieved mainly through filtration, total suspended solids (TSS) in BAF influent is usually below ⇑ Corresponding author at: Harbin Institute of Technology Shenzhen Graduate School, Shenzhen 518055, China. Tel./fax: +86 755 2603 2718. E-mail address: [email protected] (T. Liu). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.10.066
100 mg/L (Canler and Perret, 1994) to avoid filter bed clogging and associated frequent backwashing. To deal with the high suspended solids content in influent streams, chemical coagulation/ precipitation is an often applied pre-treatment step for removing TSS (Farabegoli et al., 2009). Additionally, due to the lack of alternate anaerobic and aerobic conditions, BAF is incapable of enhanced biological phosphorus removal (Lee et al., 2005), and supplementary physio-chemical precipitation of phosphorus is often used in BAF plants for meeting stringent effluent standards (Rogalla et al., 1990; Clark et al., 1997). Though an enhanced chemical pre-treatment step can reduce both TSS and total phosphorus in the influent stream fed into BAF, removal of considerable amount of organic matter can also be resulted at the same time, which reportedly could reach up to over 50% COD removal (Rogalla et al., 1990). The simultaneous organic matter removal in the chemical pre-treatment step might lead to the shortage of carbon sources in the post-denitrification process for BAF nitrified effluent, particularly when treating weak sewage with relatively low C/N
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
ratio. In order to meet the total nitrogen discharge requirement in the effluent, an external carbon source has to be added in the denitrification process, and hence operating cost is increased. In this study, a modified BAF process employing simultaneous chemical precipitation in pre-denitrification is proposed. A schematic diagram of the proposed process is shown in Fig. 1. When dosing chemicals in the pre-denitrification tank, synergistic denitrification, and removal of phosphorus and TSS through chemical precipitation is anticipated. Simultaneous chemical precipitation in biological processes can utilize the advantages of biological and chemical methods for removing nutrients (Philips et al., 2003; Caravelli et al., 2010), and offer benefits in ease of operation, flexibility to changing conditions, and relatively small additional solids production (De Haas et al., 2000). Furthermore, parts of organic matter content in the influent can be utilized by the denitrifiers in the denitrification/flocculation tank, avoiding the problem of carbon source shortage occurring in biological processes implemented with a normal chemical pre-treatment step. To obtain sound treatment effectiveness and control operating costs, the selection of suitable dosed precipitants is essential for this proposed synergistic denitrification and chemical precipitation process. Iron and aluminum salts are currently the most used precipitants in sewage treatment. Al salts are relatively costly, and residual Al in effluent would arouse concerns on the risk of neurologic diseases potentially associated with long-term exposure to Al in environments (Willhite et al., 2012). Ferric chloride and ferrous sulfate are the most used Fe salts. When dosed in oxic reactors, one principal disadvantage of using Fe2+ salt is the consumption of dissolved oxygen due to oxidation of Fe2+ to Fe3+. Fe3+ salt is however much more expensive than Fe2+ salt. Additionally, the pH value of the mixed liquor in denitrification process is around pH 7.5, and the optimum pH ranges for phosphorus precipitation with Fe3+ and Fe2+ are between pH 4 and 5 and close to pH 8, respectively. Hence, if dosing Fe salts in the proposed synergistic denitrification and chemical precipitation process, pH condition would be favorable for Fe2+ rather than Fe3+ to precipitate phosphorus. Despite Fe2+ is considered potentially a suitable precipitant being used in the proposed process due to its relatively low cost and high phosphorus removal efficiency under the typical pH condition of denitrification, an uncertainty still lies in the possible smaller size of Fe2+ dosed sludge flocs compared to Fe3+ dosed ones (Oikonomidis et al., 2010), which might adversely affect liquid–solid separation performance in the settling tank. To date, report on applying Fe2+ salt in the pre-denitrification stage of BAF process to achieve simultaneous chemical precipitation of TSS and phosphorus remains few. To provide basic information for modifying a
Effluent
Recycled Effluent
M
Denitrification/Flocculation Tank
259
BAF process with employing simultaneous pre-nitrification and chemical precipitation, this study aims to examine the feasibility and suitability of using Fe2+ salt as the precipitant dosed in the pre-denitrification stage of a BAF process. The effects of dosing Fe2+ salt on effluent quality and sludge characteristics were monitored, and comparisons with the cases of no additional chemical dosing and dosing Fe3+ salt were conducted to investigate any adverse effect introduced by dosing Fe2+ salt. 2. Methods 2.1. Experimental setup A series of laboratory scale experiments were conducted in 25 L rectors to examine the performance of dosing chemicals in the denitrification/flocculation tank shown in Fig. 1. Three sets of experiments were carried out (see Table 1). The reference experiment (R1) was performed to examine the performance of pre-denitrification without dosing any additional chemical. Another two sets of experiments were dosed with Fe2+ salt (R2) and Fe3+ salt (R3), respectively, to investigate the effects of applying chemical precipitation in the pre-denitrification process. A synthetic domestic wastewater was used as the experimental influent. It was prepared by diluting methanol, KNO3, and KH2PO4 stock solutions with deoxygenated tap water to simulate the mixed liquor of sewage influent and recycled BAF nitrified effluent, whose chemical oxygen demand (COD), NO x —NðNO3 — N þ NO2 —NÞ, and total phosphorus (TP) concentrations, and pH value were within their typical ranges. Initial activated sludge used in the reactor was collected in a municipal wastewater treatment plant (Luofang Wastewater Treatment Plant, Shenzhen, China). The experiments were continuously carried out in cycles for more than 70 days. In each cycle, 18 L of the simulating solution was firstly fed into and mixed with the sludge already inside the reactor. For the experiments dosed with Fe2+ or Fe3+ salt, certain amount of FeSO4 or FeCl3 stock solution was added, respectively, to achieve a final Fe concentration of 24 mg/L. In a preliminary experiment (data not shown), this Fe concentration was justified sufficient to control the residual phosphorus concentration in the effluents of Fe dosed reactors (R2 and R3) to below 1 mg/L. The mixed liquor in the reactor was mechanically stirred for 9 h, and then settled down for 0.5 h for liquid–solid separation. Supernatant after liquid–solid separation was discharged. Prior to starting next cycle, the sludge remained in the reactor was rinsed by deoxygenated tap water for 5 times to minimize residual nutrient accumulation. The whole operation process for each cycle took about 12 h. About 1 L of the mixed liquor in the reactors, or the supernatant after liquid–solid separation were regularly sampled for monitoring effluent quality and sludge characteristics, respectively. Initial sludge concentrations of the experiments were about 1000 mg/L. Within the time span of the experiment, no excess sludge was discharged. 2.2. Materials
Settling Tank
BAF
Chemical stock solutions were prepared by dissolving reagent grade chemicals, methanol, KNO3, KH2PO4, FeCl36H2O, and FeSO47H2O, into ultrapure water (>18.1 MX cm) (deoxygenated for preparing Fe2+ stock solution). All stock solutions were stored in 4 °C cold room. Fe2+ stock solution was acidified by sulfuric acid to pH 3 prior to storage.
Influent
Return Sludge
Chemical Dosing
2.3. Analytical methods Fig. 1. The schematic diagram of a modified BAF process with applying simultaneous chemical precipitation in pre-denitrification.
Water quality parameters are analyzed following APHA Standard Method (APHA, 1998). COD and alkalinity were analyzed by
260
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
Table 1 Effluent quality of the rectors without dosing addition chemical and with dosing Fe2+ and Fe3+.a Parameters monitored
NO x —N Total phosphorous COD Suspended solid pH Alkalinity
Unit
mg/L mg/L mg/L mg/L – mg/L (as CaCO3)
Influent
41.7 ± 5.0 4.93 ± 0.30 239.8 ± 40.0 – 7.13 ± 0.30 90 ± 20
Effluent R1 (Reference experiment)
R2 (Experiment dosed with Fe2+)
R3 (Experiment dosed with Fe3+)
0.20 ± 0.10 4.47 ± 0.32 35.5 ± 25.4 11.1 ± 7.0 7.83 ± 0.22 254 ± 25
0.10 ± 0.10 0.50 ± 0.25 36.1 ± 28.2 24.5 ± 10.8 7.40 ± 0.31 219 ± 19
0.11 ± 0.10 0.93 ± 0.25 43.3 ± 21.8 15.7 ± 9.8 7.38 ± 0.33 209 ± 21
a The data showed in the table correspond to mean values of 13 events of determination of wastewater quality parameters during the experimental time, and the associated uncertainties are the standard deviations.
The specific denitrification rates of the mixed liquors in the three reactors (R1, R2, and R3) were determined regularly. 1 L of the mixed liquor withdrawn from the reactor was mixed with a certain amount of KNO3 stock solution to achieve a final NO 3 —N concentration of 40 mg/L, and kept stirred using a magnetic stirrer. Duplicate aliquots of the mixed liquor were sampled periodically for determining NO x —N concentration. The specific denitrification rate of the mixed liquor was obtained by dividing the slope of the plot of NO x —N concentration versus time by its MLVSS content. 2.5. Statistical analysis The student’s t-test was used to statistically evaluate the significance of a difference of two data sets. A confidence level of 95% was used as the cutoff for statistical significance. 3. Results and discussion 3.1. Effects of dosing Fe2+ on effluent quality 3.1.1. Denitrification Effluent quality parameters reflecting the performance of R1 (reference reactor), R2 (Fe2+ dosed), and R3 (Fe3+ dosed) are shown in Table 1. Sound denitrification performance was recorded in all the three reactors. Effluent NO x —N concentrations of the three reactors kept below 0.5 mg/L throughout the course of the experiment. Study of the specific denitrification rates (with the unit of
(a) Specific Denitrification Rate (mg/h.gVSS)
2.4. Determination of denitrification rate
mg/h gVSS) of the mixed liquors in different reactors showed that, in the steady-state period (after day 25 of the experiment) and comparing with the specific denitrification rate of R1, R2 showed no significant difference (Fig. 2a), whereas, R3 appeared a little bit slower (student’s t-test at 95% confidence). A conversion of specific denitrification rate to denitrification loading rate (with the unit of mg/h L) was conducted by multiplying the data of the specific denitrification rate with the MLVSS concentration in each reactor. As shown in Fig. 2b, the converted denitrification loading rate in R2 was found apparently faster than that in R1. In contrasting, little difference was observed between R3 and R1. The
80
60
40
R1 R2 R3
20
0
0
10
20
30
40
50
60
70
Experiment Time (day)
(b) 80 Denitrification Loading Rate (mg/h.L)
titrimetric method; NO 3 —N, NO2 —N, and TP were determined by colorimetric method using a UV/visible spectrophotometer (UV 2450, Shimadzu); TSS, mixed liquor suspended solids (MLSS), and mixed liquor volatile suspended solids (MLVSS) concentrations were determined following method outlined in Section 2540 D of the Standard Method (APHA, 1998), and pH was measured using a pH meter. All the determinations of the examined parameters were carried out for duplicated samples. On day 40 of the experiment, samples of the mixed liquor in the three reactors (R1, R2, and R3) were withdrawn and subject to particle size distribution analysis by a laser diffraction particle size analyzer (Malvern Mastersizer 2000), and SEM-EDS (Hitachi S-4700) analysis after freeze drying. Sludge settling characteristics were assessed periodically during the experiment through zone settling test and settled sludge volume test (expressed in the term of SV30) according to methods outlined in Section 2710 C and E of the Standard Method (APHA, 1998). The dewaterability of sludge was measured regularly by the Buchner funnel test, and expressed in the term of specific resistance in filtration (SRF). The Particle size distribution analysis and the sludge zone settling test were performed for a single sample. The settled sludge volume test and the dewaterability test were conducted in duplicate.
60
40
R1 R2 R3
20
0
0
10
20
30
40
50
60
70
Experiment Time (day) Fig. 2. Changes of (a) the specific denitrification rate and (b) the denitrification loading rate along with experiment time in the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
magnitude of denitrification loading rate was mathematically determined by the values of specific denitrification rate and MLVSS concentration. An elevation of MLVSS concentrations in R2 (average 32.9% increase) and R3 (average 18.1% increase) were observed compared to that in R1 (Fig. 3a), it was however very likely resulted from increased adsorption/enmeshment of organic matters in colloidal form by Fe hydroxides complexes formed in the mixed liquor (De Haas et al., 2001). Such adsorped/enmeshed organic matters are not expected to contribute for biotic denitrification. Thus, even the MLVSS concentration in R3 was higher, the denitrification loading rate of R3 remained almost unchanged comparing to that of R1. In contrast to R3, the denitrification loading rate of R2 showed apparently increased. In the steady-state period (after day 25 of the experiment) and compared to R1, an average 34.8% increase in denitrification loading rate was recorded in R2, indicating an enhanced denitrification when dosing Fe2+. A predominantly microbial Fe2+-dependent nitrate reduction process with high rate is believed to be responsible for the enhanced denitrification resulting from dosing Fe2+. Straub et al. (1996) reported that microbes might be important for anoxic oxidation of Fe2+ with nitrate in environmental systems. Their observation was further confirmed by Nielsen and Nielsen (1998), whose work demonstrated the occurrence of microbial Fe2+dependent nitrate removal in activated sludge simultaneously with the oxidation of Fe2+ to Fe3+, which resulted in an increase
(a) 1400 1200
MLVSS (mg/L)
1000 800 600 400
R1 R2 R3
200 0
0
10
20
30
40
50
60
70
Time (day)
(b) 3000
MLSS (mg/L)
2500
2000
1500
1000
R1 R2 R3
500
0
0
10
20
30
40
50
60
70
Time (day) Fig. 3. Changes of the concentrations of (a) MLVSS and (b) MLSS along with experiment time in the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
261
in denitrification rate, ranged from 3% to 107%, compared to background organic matter-dependent denitrification. The results of the presented work together with previous studies clearly indicated the contribution of the microbial Fe2+-dependent nitrate reduction to the denitrification process in addition to the organic matter-dependent nitrate removal. Seen from an engineering view, the enhancement on denitrification signified an advantage of using Fe2+ over dosing no chemical or Fe3+ in the proposed synergistic denitrification and chemical precipitation process in increasing denitrification loading, and hence reducing the volume of the pre-denitrification/flocculation tank of a modified BAF treatment facility. But the detailed mechanism of dosing Fe2+ on the enhanced denitrification is not well described so far, and needs further study. 3.1.2. Phosphorus removal Since there was no alternating anaerobic/aerobic process, enhanced biological phosphorus removal cannot be achieved in a normal denitrification reactor, and hence poor phosphorus removal was recorded in R1 (reference reactor), with an average phosphorus removal rate of only 9.3%. On the contrary, good phosphorus removal was observed in both R2 (Fe2+ dosed) and R3 (Fe3+ dosed), where the average phosphorus removal rates were 89.9% and 81.1%, respectively. Fe:P molar ratio applied in both R2 and R3 in this study was 2.7:1, which is higher than the reported required molar ratios of Fe2+:P (2–2.4:1) (Clark et al., 2000; Banu et al., 2008) and Fe3+:P (1.5–1.9:1) (Clark et al., 2000; Caravelli et al., 2010) applied in activated sludge systems where the residual phosphorus concentrations in the effluents were decreased to below 1 mg/L. The high phosphorus removal rates recorded in R2 and R3 should mainly result from chemical precipitation of sparingly soluble phosphates by dosed Fe salts. It was evidenced by SEM-EDS results that atomic concentrations of both Fe and P were much higher in the dried sludge collected in R2 and R3 than that collected in R1 (Supplementary data Fig. S1). Though high phosphorus removal rates were achieved in R2 and R3, R2 showed a better controlling of the effluent residual phosphorus than R3 (student’s t-test at 95% confidence, data not shown). The average effluent phosphorus concentration in R3 was 0.93 mg/L; whereas, that in R2 was 0.50 mg/L. Such difference in controlling of the residual phosphorus in R2 and R3 was considered reflecting the pH dependence of phosphorus removal by chemically precipitation with Fe2+ or Fe3+. pH values of the mixed liquors in R2 and R3 were in the range of 7.1–7.4. The optimum pH range for phosphorus precipitation with Fe3+ is between pH 4 and 5 (Stumm and Morgan, 1996); whereas, maximum phosphorus removal by Fe2+ was reportedly obtained in the vicinity to pH 8 (USEPA, 1971). The higher efficiency of phosphorus precipitation with Fe2+ at a higher pH should attribute to the lower tendency to hydrolyze of Fe2+ than Fe3+ (Stumm and Morgan, 1996), and thereby diminishing the competition between hydroxide and orthophosphate ions for reacting with Fe2+. The better performance in controlling the residue phosphorus in R2 than R3 announced another advantage of using Fe2+ over Fe3+ in the pre-denitrification stage in keeping effluent phosphorus at a lower level, where alkalinity production occurs and typical pH ranges from 7 to 8. 3.1.3. COD, SS, pH, and alkalinity As shown in Table 1, average COD removal rates of 87.9%, 84.9%, and 81.9% were achieved in R1, R2, and R3, respectively. The similar good COD removal performance in the three reactors showed that dosing Fe salts has no adverse impact in removing organic matters. The effluent suspended solids (SS) content in R2 was observed apparently greater than that in R1, and the effluent SS in R3 was slightly higher than that in R1. Flocs formed in activate sludge systems dosed with Fe2+ often appear compact and in
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
smaller size, and may partly be washed out to the effluent, resulting in relatively poor performance in solids removal (Thistleton et al., 2001; Oikonomidis et al., 2010). In this study, the reduction in sludge floc size caused by dosing Fe salts was evidenced through particle size distribution analysis of the mixed liquor collected from the three rectors (Supplementary data Fig. S2). Dosing Fe salts caused d50 of sludge flocs to decrease. d50 of sludge flocs in R1 was 72.3 lm, and in R2 and R3, d50 decreased to 12.9 and 36.2 lm, respectively. Thus, the descending variation in sludge flocs particle size resulted from dosing Fe2+ (R2) or Fe3+ (R3) should be responsible for the higher effluent SS content compared to the experiment without additional chemical dosing (R1). The effluent pH and alkalinity in R1, R2, and R3 were increased compared to the influent (Table 1) due to alkalinity production in the denitrification process. But the elevations of pH and alkalinity in R2 and R3 were not as high as that in R1. It should be attributed to the hydrolysis of Fe took place in R2 and R3, which leads to the formation of Fe hydroxides resulting in alkalinity consumption and an observed pH lowering in R2 and R3 compared to R1.
(a)
60
0 9 25 42 70
40
0
0
5
10
15
20
25
30
25
30
Time (min)
(b)
100 Day Day Day Day Day
Height of Interface (%)
80
60
0 9 25 42 70
40
20
0
0
5
10
15
20
Time (min)
(c) 100 Day Day Day Day Day
80
Height of Interface (%)
3.2.2. Sludge settleability The settling characteristics of sludge in the three reactors were assessed by zone settling test and settled sludge volume test (expressed in the term of SV30) in this study, and the results are shown in Figs. 4 and 5, respectively. Sludge settling process can be generally divided into three stages, those are hindered settling stage, transitional settling stage, and compression settling stage. The duration of the former two stages is crucial for determining the residence time in a sludge settling tank. As shown in Fig. 4a, there was no significant change in the settleability of the sludge in R1 along with experimental time, where the hindered settling stage ended at about 4 min, and the transitional settling stage lasted till around 10 min. In contrasting, sludge settleability was observed significantly improved in R2 (Fe2+ dosed) (Fig. 4b) in the steady-state period (after day 25 of the experiment). The time span for the hindered settling stage of the sludge in R2 was shortened to about 2 min, and the transitional settling stage ended at about 5 min. Zone settling test curves of the sludge in R3 (Fe3+ dosed) (Fig. 4c) were similar as those of R1. As shown in Fig. 5, SV30 of the sludge in R1 did not changed significantly throughout the experiment with an average value of 26.8%. Compared to that of R1, in the steady-state period (after day 25 of the experiment), SV30 of the sludge in R3 showed a little bit decrease, and an average SV30 value of 23.9% was recorded; whereas, SV30 of the sludge in R2 was observed remarkably lower, and its average value was only 10.4%. The much better settleability exhibited by Fe2+ dosed sludge in this study was in agreement with what was reported in literatures (Clark et al., 2000; Muller, 2001; Oikonomidis et al., 2010), and is suspected related to more compact flocs of this sludge. SEM analysis on the freeze-dried sludge revealed that Fe2+ dosed sludge floc
Day Day Day Day Day
20
3.2. Effects of dosing Fe2+ on sludge characteristic 3.2.1. Sludge production As discussed in the previous section, average MLVSS concentrations in R2 and R3 were 32.9% and 18.1% greater than that in R1 (Fig. 3a), which should result from increased adsorption/enmeshment of organic matters in colloidal form by Fe hydroxides complexes formed after dosing Fe salts (De Haas et al., 2001). In terms of TSS (i.e. MLSS), a significant increase (more than one fold increase in average) in sludge production occurred in R2 and R3 compared to R1 (Fig. 3b). These results were expected for additional phosphorus removal through the way of chemical precipitation. As a result, the chemical precipitates would also contribute to sludge production in the form of inorganic suspended solids.
100
80
Height of Interface (%)
262
60
0 9 25 42 70
40
20
0
0
5
10
15
20
25
30
Time (min) Fig. 4. Zone settling test curves of the mixed liquors collected from (a) the reactor without dosing addition chemical (R1), (b) the reactor dosed with Fe2+ (R2), and (c) the reactor dosed with Fe3+ (R3).
displayed more compact and was featured by relatively smooth surface; whilst, rough and open structures were observed on the surfaces of sludge flocs either without dosing additional chemical or dosed with Fe3+ (Supplementary data Fig. S1). The smooth surface features (Supplementary data Fig. S1b) is believed to be condensed and less hydrated extracellular polymeric substances (EPS), and the rough and open structures (Supplementary data Fig. S1a and c) is attributed to more hydrated EPS matrix having a relative high water-binding capacity (Liss et al., 2002). One
263
H. Wang et al. / Bioresource Technology 151 (2014) 258–264 60
35 30
SRF (10 m/kg)
20
12
SV30(% )
25
15 10 R1 R2 R3
5 0
R1 R2 R3
50
0
10
40
30
20
10
20
30
40
50
60
70
0
0
20
40
60
Time (day)
Time (day) Fig. 5. Changes of the settled sludge volume in 30 min (SV30) along with experiment time of the mixed liquors collected from the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
Fig. 6. Changes of the sludge specific resistance in filtration (SRF) along with experiment time of the mixed liquors collected from the reactors without dosing addition chemical (R1), dosed with Fe2+ (R2), and dosed with Fe3+ (R3). Error bars indicate the standard deviations of duplicate determinations of the examined parameter.
mechanism suggested to explain the formation of the compact sludge flocs by dosing Fe2+ is that the dosed Fe2+ reacts faster than Fe3+ with the negatively charged sludge colloidal materials; and subsequently, Fe3+ engendered in situ from Fe2+ oxidation would form tight binding with the colloidal materials and serve to decrease the specific area of sludge floc and make the flocs more compact and denser (Muller, 2001). It was also proposed that the major interaction mechanism of the dosed Fe2+ with sludge flocs is cation bridging which would stabilize the negatively charged groups present on the EPS and promote sludge flocculation (Oikonomidis et al., 2010). On the contrary, when the dosed precipitant is Fe3+, the formation of ferric hydroxide (Fe(OH)3) would occur rapidly under the pH condition (around pH 7.5) of the denitrification process. Fe(OH)3 present in colloidal form would have low neutralization capacity towards the negatively charged biopolymers in sludge flocs, and hence result in a more open floc structure and limited improvement on the sludge settleability compared to dosing Fe2+ (Oikonomidis et al., 2010). As for its engineering implication, the remarkably improved settleability of sludge in R2 displayed one more advantage of using Fe2+ over Fe3+ in the proposed synergistic denitrification and chemical precipitation process in shortening the residence time, and in turn, reducing the volume of a sludge settling tank.
(Liu and Fang, 2003; Park et al., 2006), and the bound EPS showed improving sludge dewaterability (Higgins and Novak, 1997; Murthy and Novak, 1999). Though a better settleability was observed for Fe2+ dosed sludge than Fe3+ dosed one in this study, the sludge dewatering performance was not found affected by which Fe species was dosed. Employing either Fe2+ or Fe3+ in the proposed synergistic denitrification and chemical precipitation process would finally result in an enhancement in sludge dewatering and conduce to subsequent sludge processing compared to no additional chemical dosing.
3.2.3. Sludge dewaterability Dewatering performance of activated sludge is commonly characterized in the term of specific resistance in filtration (SRF). The larger SRF value indicates the worse sludge dewatering performance. Fig. 6 shows the changes of SRF values of the sludge in the three reactors (R1, R2, and R3) along with experimental time. SRF values of the sludge in R1 kept relatively high throughout the experimental time, and with an average value of 38.8 1012 m/kg. In the steady-state period (after day 25 of the experiment), SRF values of the sludge in both R2 (Fe2+ dosed) and R3 (Fe3+ dosed) kept around 2 1012 m/kg which was significantly decreased compared to in R1. These results are of expectation. Despite the lack of a clear understanding, sludge dewatering could be affected by a number of EPS-related parameters, such as floc size, structure, and bound water content (Liu and Fang, 2003). The increasing in bound EPS resulting from the addition of multivalent cations has been reported in numerous studies
4. Conclusion The results of this study demonstrate the feasibility and suitability of using Fe2+ as the precipitant dosed in the pre-nitrification stage of a modified BAF process for synergistic denitrification and chemical precipitation of TSS and phosphorus. Dosing Fe2+ exhibited strengths over dosing Fe3+ or no chemical in increasing the denitrification loading rate, having a better controlling of residual phosphorus in the pre-denitrification effluent, and improving sludge settleability. Findings obtained provide engineering implications in reducing the volumes of the pre-denitrification tank and the associated settling tank, and ultimately, the space and capital costs required for a modified BAF process employing predenitrification. Acknowledgements This work is supported by research projects from National science and technology major project of China for water pollution control (Grant No. 2012ZX07206002-02), Ministry of Housing and Urban-Rural Development of China (Grant No. 2012-K7-18), and Shenzhen Technology Innovation Plan (Grant No. CXZZ201303 19100941767). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2013. 10.066.
264
H. Wang et al. / Bioresource Technology 151 (2014) 258–264
References Abu Hasan, H., Sheikh Abdullah, S.R., Tan Kofli, N., Kamarudin, S.K., 2012. Effective microbes for simultaneous bio-oxidation of ammonia and manganese in biological aerated filter system. Bioresour. Technol. 124, 335–363. APHA, 1998. Standard Methods for the Examination of Water and Wastewater, twentieth ed. American Public Health Association, Washington, DC. Bai, Y., Sun, Q., Sun, R., Wen, D., Tang, X., 2011. Bioaugmentation and adsorption treatment of coking wastewater containing pyridine and quinoline using zeolite-biological aerated filters. Environ. Sci. Technol. 45, 1940–1948. Banu, J.R., Do, K., Yeom, I., 2008. Effect of ferrous sulphate on nitrification during simultaneous phosphorus removal from domestic wastewater using a laboratory scale anoxic/oxic reactor. World J. Microb. Biot. 24, 2981–2986. Canler, J.P., Perret, J.M., 1994. Biological aerated filters: assessment of the process based on 12 sewage treatment plants. Water Sci. Technol. 29, 13–22. Caravelli, A.H., Contreras, E.M., Zaritzky, N.E., 2010. Phosphorous removal in batch systems using ferric chloride in the presence of activated sludges. J. Hazard. Mater. 177, 199–208. Clark, T., Stephenson, T., Pearce, P.A., 1997. Phosphorus removal by chemical precipitation in a biological aerated filter. Water Sci. Technol. 31, 2557–2563. Clark, T., Burgess, J.E., Stephenson, T., Arnold-Smith, A.K., 2000. The influence of iron-based co-precipitant on activated sludge biomass. Process Saf. Environ. 78, 405–410. De Haas, D.W., Wentzel, M.C., Ekama, G.A., 2000. The use of simultaneous chemical precipitation in modified activated sludge systems exhibiting biological excess phosphate removal. Part 1: literature review. Water SA 26, 439–452. De Haas, D.W., Wentzel, M.C., Ekama, G.A., 2001. The use of simultaneous chemical precipitation in modified activated sludge systems exhibiting biological excess phosphate removal. Part 5: experimental periods using a ferrous-ferric chloride blend. Water SA 27, 117–134. Farabegoli, G., Chiavola, A., Rolle, E., 2009. The biological aerated filter (BAF) as alternative treatment for domestic sewage. Optimization of plant performance. J. Hazard. Mater. 171, 1126–1132. Gilbert, Y., Bihan, Y.L., Aubry, G., Veillette, M., Duchaine, C., Lessard, P., 2008. Microbiological and molecular characterization of denitrification in biofilters treating pig manure. Bioresour. Technol. 99, 4495–4502. González-Martínez, S., Piña-Mondragón, S., González-Barceló, Ó., 2010. Treatment of the azo dye direct blue 2 in a biological aerated filter under anaerobic/aerobic conditions. Water Sci. Technol. 61, 789–796. Higgins, M.J., Novak, J.T., 1997. The effect of cations on the settling and dewatering of activated sludges: laboratory results. Water Environ. Res. 69, 215–224. Lee, J., Kim, J., Lee, C., Yun, Z., Choi, E., 2005. Biological phosphorus and nitrogen removal with biological aerated filter using denitrifying phosphorus accumulating organism. Water Sci. Technol. 52, 569–578.
Li, Q., Sun, S., Guo, T., Yang, C., Song, C., Geng, W., Zhang, W., Feng, J., Wang, S., 2013. Short-cut nitrification in biological aerated filters with modified zeolite and nitrifying sludge. Bioresour. Technol. 136, 148–154. Liss, S.N., Liao, B.Q., Droppo, I.G., Allen, D.G., Leppard, G.G., 2002. Effect of solids retention time on floc structure. Water Sci. Technol. 46, 431–438. Liu, Y., Fang, H.H.P., 2003. Influence of extracellular polymeric substances (EPS) on flocculation, settling, and dewatering of activated sludge. Crit. Rev. Env. Sci. Technol. 33, 237–273. Mendoza-Espinosa, L., Stephenson, T., 1999. A review of biological aerated filters (BAFs) for wastewater treatment. Environ. Eng. Sci. 16, 201–216. Muller, C.D., 2001. High-Intensity Shear as a Wet Sludge Disintegration Technology and a Mechanism for Floc Structure Analysis (M.S. thesis). Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Murthy, S.N., Novak, J.T., 1999. Factors affecting floc properties during aerobic digestion: implication for dewatering. Water Environ. Res. 71, 197–202. Nielsen, J.L., Nielsen, P.H., 1998. Microbial nitrate-dependent oxidation of ferrous iron in activated sludge. Environ. Sci. Technol. 32, 3556–3561. Oikonomidis, I., Burrows, L.J., Carliell-Marquet, C.M., 2010. Mode of action of ferric and ferrous iron salts in activated sludge. J. Chem. Technol. Biotechnol. 85, 1067–1076. Park, C., Muller, C.D., Mohammad, M.A., Novak, J.T., 2006. The effect of wastewater cations on activated sludge characteristics: effects of aluminum and iron in floc. Water Environ. Res. 78, 31–40. Philips, S., Rabaey, K., Verstraete, W., 2003. Impact of iron salts on activated sludge and interaction with nitrite or nitrate. Bioresour. Technol. 88, 229–239. Rogalla, F., Payraudeau, M., Bacquet, G., Bourbigot, M., Sibony, J., Gilles, P., 1990. Nitrification and phosphorus precipitation with biological aerated filters. Res. J. Water Pollut. Control Fed. 62, 169–176. Straub, K.L., Benz, M., Schink, B., Widdel, F., 1996. Anaerobic, nitrate-dependent microbial oxidation of ferrous iron. Appl. Environ. Microbiol. 62, 1458–1460. Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, third ed. Wiley-Interscience, New York. Thistleton, J., Clark, T., Pearce, P., Parsons, S.A., 2001. Mechanisms of chemical phosphorus removal 1 – Iron(II) salts. Process Saf. Environ. 79, 339–344. USEPA, 1971. Water pollution control research series: phosphate precipitation with ferrous iron. 17010 EKI 09/71. Willhite, C.C., Ball, G.L., McLellan, C.J., 2012. Total allowable concentrations of monomeric inorganic aluminum and hydrated aluminum silicates in drinking water. Crit. Rev. Toxicol. 42, 358–442. Zhao, Y., Yue, Q., Li, R., Yue, M., Han, S., Gao, B., Li, Q., Yu, H., 2009. Research on sludge-fly ash ceramic particles (SFCP) for synthetic and municipal wastewater treatment in biological aerated filter (BAF). Bioresour. Technol. 100, 4955–4962.
