- Email: [email protected]
Granule development and performance in sucrose fed anaerobic baffled reactors
Journal of Biotechnology 122 (2006) 198–208
Granule development and performance in sucrose fed anaerobic baffled reactors Zonglian She ∗ , Xilai Zheng, Bairen Yang, Chunji Jin, Mengchun Gao Laboratory of Marine Environmental Science and Ecology, College of the Environmental Science and Engineering, Ocean University of China, No. 5 Yushan Road, Qingdao 266003, China Received 27 May 2005; received in revised form 27 August 2005; accepted 13 September 2005
Abstract Two 90 L anaerobic baffled reactors were used to study the granulation of sludge and the effect of the organic loading rate and NaHCO3 /COD ratios on reactor performance. Furthermore, it was determined whether an anaerobic baffled reactor would promote phase separation and if additive of bentonite or granular active carbon was capable of enhancing granule formation. In order to minimize feed variations, and have a totally biodegradable substrate, a synthetic sucrose substrate was used. Granulation was achieved in both reactors within 75 days. However, the granules from the granular active carbon amended reactor appeared earlier and were larger and more compact. The reactors were maintained at a hydraulic retention time of 20 h during performance study stage. The results showed that when organic loading rate were changed from 2.15 to 6.29 kg COD m−3 day−1 , chemical oxygen demand (COD) removal was not decreased (91–93%), but a slight increase in effluent COD was observed. It was found that the COD removals were generally good (87–92%) and had not obviously change with the decreasing NaHCO3 /COD ratios. From the bacterial distribution and the concentration of volatile fatty acids in four compartments, it was concluded that a separation of phases occurred within the anaerobic baffled reactors. © 2005 Published by Elsevier B.V. Keywords: Anaerobic baffled reactor; Granulation of sludge; Performance; Organic loading rate; NaHCO3 /COD ratios
1. Introduction The anaerobic baffled reactor (ABR) is one of the many types of high-rate anaerobic reactors and able to separate hydraulic retention time (HRT) from solid ∗ Corresponding author. Tel.: +86 532 2032571; fax: +86 532 2032102. E-mail address: [email protected] (Z. She).
0168-1656/$ – see front matter © 2005 Published by Elsevier B.V. doi:10.1016/j.jbiotec.2005.09.004
retention time (SRT). This allows anaerobic microorganism to remain within the reactor and produces higher volumetric loads and significantly enhanced removal efficiencies. The ABR was initially developed by McCarty and co-workers at Stanford University. In the ABR, a series of vertical baffles forces the wastewater to flow under and over them as it passes from inlet to outlet. This configuration has been shown to result in a high degree
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
of chemical oxygen demand (COD) removal at high organic loading rate (OLR) (Grobicki and Stuckey, 1991; Uyanik et al., 2002). The main advantage of using an ABR comes from its compartmentalized structure. This leads to separating acidogenesis and methanogenesis longitudinally down the reactor, allowing the different bacterial groups to develop under most favorable conditions. This characteristic is similar to two-phase digestion process. A number of advantages of the twophase digestion process have been summarized in the literature (Jeyaseelan and Matsuo, 1995; Bhattacharya et al., 1996; Ince, 1998). Granulation is a complex process, in which suspended biomass agglutinates to form discrete macroscopic aggregates. This enhances the settleability of the biomass and leads to an effective retention of bacteria in the reactor. It is widely assumed that granular biomass have operational advantages over the use of flocculent biomass (Wirtz and Dague, 1996; Pereboom, 1997; Batstone and Keller, 2001). The most significant advantage of granulation is high biomass concentration in continuous reactors, and then high treatment efficiencies can be expected. Upflow anaerobic sludge blanket (UASB) reactors are generally considered to be the best design for promoting granulation. The ABR may be classified as a series of UASB reactors. So, the ABR has shown the potential to produce granular sludge (Barber and Stuckey, 1999). The objectives of this study were: (1) to examine effects with addition of granular active carbon (GAC) or bentonite and polyacrylamide on the granulation of sludge in ABR; (2) to investigate the performance of ABR during various OLR and (3) to investigate the performance of ABR at different NaHCO3 /COD ratios in influent. In order to achieve these goals, two ABRs having identical dimensions and configurations were used.
2. Materials and methods 2.1. The reactors The ABRs were made of clear acrylic plastic with total effective volume of 90 L. Every reactor contained three vertical standing baffles that divided it into four identical compartments (as shown schematically in Fig. 1). Within each compartment, downcomer and
199
Fig. 1. Schematic diagram of ABR. 1, Feed tank; 2, heater; 3, rotameter; 4, influent; 5, gas collection; 6, effluent; 7, supernatant sampling port; 8, sludge sampling port.
riser regions were created by a further slanted edge (45◦ ) vertical baffle in order to direct the flow evenly through the riser. The width of the downcomer was 6 cm and the riser was 17 cm. Each compartment was equipped with sampling ports that allowed biological solids and liquid samples to be withdrawn. The influent was feed from troughs in higher position and rotameters were used to control the influent feed rate to the ABR. The temperature of influent water in the troughs was heated to about 50 ◦ C in order to maintain the operating temperature of the first compartment at constant 34 ◦ C. The produced gas was collected via portholes in the top of the reactors and was recorded by wet gas meters. 2.2. Sampling and analyses Supernatant liquor and sludge samples were taken separately from each compartment for analysis to characterise the behaviour of the reactors. Analysis of supernatant liquor included COD, pH, temperature, alkalinity and volatile fatty acids (VFA). COD, pH and temperature were monitored every day. Alkalinity and VFA were analysed weekly. COD and pH of the influent and effluent were analysed every day, too. Analysis of sludge samples included suspended solids (SS), volatile suspended solids (VSS) and particle size distribution at the end of every stage. COD, SS, VSS, pH, VFA and alkalinity were determined as described in Standard Methods for Examination of Water and Wastewater published by American Public Health Association (1995).
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
200
2.3.1. The development of granular sludge The reactors were seeded with anaerobically digested sewage sludge taken from a primary anaerobic digester at a local municipal sewage treatment works. The sludge was first sieved (<3 mm) to remove debris and large particles and was then introduced into all four compartments of each reactor. Each 22.5 L compartment contained 11.2 L sludge having a suspended solids composition of 28,700 mg SS L−1 and 14,050 mg VSS L−1 giving a total of 629 g VSS in each reactor. This value (6.99 g VSS L−1 ) is in accordance with the initial VSS values used in other studies on ABR (Barber and Stuckey, 1999; Sallis and Uyanik, 2003). At the same time of seeding sludge, reactor A was fed with GAC (Ø0.2–0.4 mm, 134.4 g) and reactor B was fed with bentonite (5‰, 9 L) and polyacrylamide (0.5‰, 9 L) for accelerating formation of granular sludge. Mean particle size of the bentonite was 5.2 m and the molecular weight of polyacrylamide was 1200 × 104 . After seeding the reactors, the covers were sealed and then synthetic wastewater was introduced into the reactors. Initially, the reactors were fed with low COD concentration (1500 mg L−1 ) and had long hydraulic retention time (30 h) in order to encourage the growth of granular biomass. Thereafter, influent feed strength was increased and at the same time the HRT was reduced in a stepwise fashion. Operational parameters during the stage of granulation were given in Table 1.
For estimating the size distribution, the sludge samples taken from the bottom sampling points of each chamber were classified into four fractions using laboratory sieves with various openings (0.5, 1.0 and 2.0 mm). The sludge particles were first placed in the sieve with the biggest opening (2.0 mm). The particles were gently submerged in water and shaken to let the smaller particles pass through. The procedures were repeated until the three sieves were used. After granulation of sludge, sludge samples were taken from each compartment of both reactors and were examined by scanning electron microscope (SEM). The samples were first fixed with 2.5% glutaric dialdehyde solution, and dehydrated through a graded series of water–ethanol mixtures. The samples were dried by the critical-piont drying method before sputter-coating with gold particles. The samples were then examined in a SEM. 2.3. Operational methods The ABRs were fed with a sucrose based synthetic medium, which contained nutrition (COD:N:P = 200:5:1) and trace metals. The trace metals was added into the influent as the following (mg per gram COD): CaCl2 , 10.0; MgSO4 ·7H2 O, 20.0; FeSO4 ·7H2 O, 1.5; MnSO4 ·H2 O, 0.5; H3 BO3 , 0.5; ZnCl2 , 0.5; CuCl2 ·2H2 O, 0.5; (NH4 )2 Mo7 O2 ·4H2 O, 0.5; CoCl2 ·6H2 O, 0.3 and NiCl2 ·6H2 O, 0.5. The experiment had two stages. The first stage is the development of granular sludge (start-up stage) and the second is performance study stage. During the second stage, the HRT was kept constant at 20 h while influent NaHCO3 was reduced in one reactor (A) and OLR was increased in another reactor (B). For reactor A, NaHCO3 concentrations in feed varied from 970 to 100 mg L−1 .
2.3.2. Performance study of varying NaHCO3 /COD ratios or OLR After granulation of sludge, the investigation in this study was carried out with a constant HRT (20 h) and varying NaHCO3 /COD ratios in influent in reactor A, and varying influent COD concentration (i.e. varying OLR) in reactor B. Reactor A was fed with sucrose synthetic influent containing about 2000 mg L−1 COD at
Table 1 Operational parameters during granulation process Steps
Times (day)
Influent COD (mg L−1 )
HRT (h)
OLR (kg COD m−3 day−1 )
Liquid upflow velocities (m h−1 )
1 2 3
1–25 26–55 56–75
1500 2000 3000
30 20 16
1.2 2.4 4.5
0.074 0.11 0.14
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
201
Table 2 The size distribution (by weight) of granular sludge by the end of start-up The size distribution of granular sludge (%)
Particle size (mm)
<0.5 0.5–1.0 1.0–2.0 >2.0
A1
A2
A3
A4
B1
B2
B3
B4
29.2 35.1 22.8 12.9
39.2 28.7 24.8 8.3
59.8 22.9 14.5 2.8
55.7 28.8 11.7 3.8
36.9 33.8 18.6 10.7
44.5 32.3 15.2 8.0
70.2 18.9 9.3 1.6
66.5 25.9 6.8 0.8
different NaHCO3 /COD ratios (0.5, 0.35, 0.25, 0.2, 0.1 and 0.05 g NaHCO3 /gCOD). Reactor B was fed with sucrose synthetic influent (containing 0.5 g NaHCO3 /g COD) at different values of COD (1793, 3008, 4053, 5245, 5868 and 8012 mg L−1 ) corresponding to OLR of about 2.15, 3.61, 4.86, 6.29, 7.04 and 9.61 kg COD m−3 day−1 . For each NaHCO3 /COD ratio and OLR studied, steady state performance was marked by near constant effluent pH values and COD values with less than 10% variation. Six samples of influent and effluent were usually collected and analysed during steady state. The average values of the six samples were regarded as the steady state performance data. During the studies (including start-up), dilution of the influent was carried out with tap water. It showed a steady decrease in temperature down the reactors (A1 34 ◦ C, A2 32 ◦ C, A3 30 ◦ C, A4 29 ◦ C, B1 34 ◦ C, B2 31 ◦ C, B3 29 ◦ C and B4 26 ◦ C).
3. Results and discussion 3.1. Granulation of sludge The development of granule sludge had three steps (Table 1). Granulation was first noted on day 45 in compartment 1 and 2 of reactor A and compartment 1 of reactor B, and after further 10 days granule-like flocs started to appear in all compartments of both reactors. By day 75 of the experiment all compartments in both reactors contained predominantly granular sludge.
Sludge samples were taken from each compartment of both reactors on day 75 of operation in order to study performance of the granular sludge. The size distribution (by weight) of the granules in both reactors was shown in Table 2 on day 75 of operation. The maximum size of granules was 6.0 mm in both reactors. The granules from reactor A appeared earlier and were larger and more compact. It seemed to be more effective in promoting the formation of anaerobic granules with addition of GAC than bentonite and polyacrylamide into inoculated sludge. According to the inert nuclei model for anaerobic granulation, the presence of micro-size biocarrier for bacterial attachment is a first step towards anaerobic granulation and the contact between particles and biomass could be improved using nuclei of lower density than sand and other inert materials (Liu et al., 2003; Imai et al., 1997). The percentages of granules in different sizes were determined through calculating the ratios between the weight of the granules in a range of size and the weight of granules in all sizes. It was showed a steady decrease in particle size down both reactors. About 35.7 and 29.3% (on weight basis, w/w) of the granules from chamber 1 of reactor A and B were larger than 1.0 mm and only 15.5 and 7.6% of granules from chamber 4 were above 1.0 mm. Orozco (1988) reported a similar decrease in granule size treating dilute carbohydrate waste. The solids concentration and VSS/TSS values appeared a steady decrease down both reactors, too (Table 3). The maximum sludge concentrations were 50.1 and 45.1 g TSS L−1 in compartment 1 of reactor A and B.
Table 3 The solids concentrations in each compartment by the end of start-up Solids
A1
A2
A3
A4
B1
B2
B3
B4
TSS (g L−1 ) VSS (g L−1 ) VSS/TSS (%)
50.1 37.8 75.4
45.6 31.5 69.1
40.5 25.3 62.5
35.4 19.5 55.1
45.1 31.4 69.6
40.1 26.5 66.1
35.9 20.1 56.0
32.2 17.4 54.0
202
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 2. Scanning electron photomicrographs showing the external surfaces of granular sludge in both reactors.
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 3. Scanning electron photomicrographs showing the bacterium of granular sludge.
203
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
204
showed that microbial selection and zoning are encouraged inside the ABRs.
The seed sludge was black, and the surface of granular sludge was grey in compartment 1 of reactor A while black in others compartments of both reactors. The SEM photomicrographs of granule samples taken from every compartment of both reactors are shown in Figs. 2 and 3. The granules had compact structure and naturally occurring cavities on the surface, as illustrated in Fig. 2. All granules examined were composed of various types of bacteria including bacilli, cocci and filaments. The granules in the initial compartments of the reactors were mainly composed of acidogenic short rod-shaped and filamentous bacteria, as illustrated in Fig. 3A1, B1 and B2. On the basis of appearance, in chambers 2, 3 and 4 of reactor A, the preponderant bacteria were similar to Methanobacterium. In compartment 3 of reactor B, granules were dominated presumably by Methanococcus and Methanobacterium. In compartment 4 of reactor B, the preponderant bacteria were similar to Methanobacterium. This result
3.2. Effect of OLR on the performance of ABR In order to determine the treatment capacity, the OLR in reactor B was progressively increased by increasing the strength of the feed. The mean performance data of reactor B throughout the entire length of the reactor during each OLR period is given in Table 4. The VFA concentrations decreased longitudally down the reactor. The highest VFA concentration was found in the first compartment with average values of 612–1548 mg L−1 . The VFA concentrations in every compartment increased with the increasing OLR. The VFA data demonstrated that hydrolysis and acidogenesis were the main biochemical activities occurring in
Table 4 Mean reactor performance characteristics at each OLR OLR (kg COD m−3 day−1 )
Influent COD (mg L−1 )
COD (mg L−1 )
pH
2.15
1793
B1 B2 B3 B4
1014 434 221 59
B1 B2 B3 B4
6.08 6.55 6.61 6.61
B1 B2 B3 B4
462 703 736 774
B1 B2 B3 B4
612 251 152 88
3.61
3008
B1 B2 B3 B4
1808 566 274 196
B1 B2 B3 B4
5.47 6.6 6.65 6.64
B1 B2 B3 B4
323 728 946 1064
B1 B2 B3 B4
1245 498 319 193
4.86
4053
B1 B2 B3 B4
2888 1304 534 404
B1 B2 B3 B4
4.56 5.98 6.39 6.44
B1 B2 B3 B4
0 514 790 855
B1 B2 B3 B4
1036 564 432 224
6.29
5245
B1 B2 B3 B4
3587 1473 411 356
B1 B2 B3 B4
3.95 5.56 6.26 6.33
B1 B2 B3 B4
0 359 674 718
B1 B2 B3 B4
1526 714 334 122
7.04
5868
B1 B2 B3 B4
4654 2994 1874 1187
B1 B2 B3 B4
4.33 5.16 5.78 6.06
B1 B2 B3 B4
0 174 455 656
B1 B2 B3 B4
1305 1191 940 574
9.61
8012
B1 B2 B3 B4
6153 3254 2362 1818
B1 B2 B3 B4
3.81 4.88 6.03 6.15
B1 B2 B3 B4
0 85 442 612
B1 B2 B3 B4
1548 1524 804 552
VFA (mg HAc L−1 )
Alkalinity (mg CaCO3 L−1 )
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 4. Variation with OLR of relative COD removal in compartments.
the first and first two compartments at low and high OLR. The methanogenesis appeared to be dominant in the third and fourth compartments. These observations suggest that the ABR system promoted a systematic selection in the different compartments in such a manner as to bring about phase separation. Similar observations have been reported in other studies (Akunna and Clark, 2000; Wang et al., 2004). The alkalinity values appeared a significant increasing step by step from compartment 1 to 4. Furthermore, the increasing of OLR resulted in the significant reduction of alkalinity in compartment 1–2. The COD removals were generally very good (91–93%) when OLR was lower than 6.29 kg COD m−3 day−1 . The best performance was observed with influent COD of 3008 mg L−1 (or a loading of 3.61 kg COD m−3 day−1 ). When the influent COD increased to 5245 mg L−1 (or a loading of 6.29 kg COD m−3 day−1 ) the efficiency of COD removals began to drop. It was found that most organic matter was removed in the first two chambers, and COD removal in the four compartments was 35.0, 41.6, 22.5 and 1.2% based on the overall removal at OLR of 6.29 kg COD m−3 day−1 (Fig. 4). Vossoughi et al. (2003) who used a five-chamber ABR to treat synthetic wastewater at an HRT of 1 day, have also reported that over 65% COD removal was obtained in the first chamber. Manariotis and Grigoropoulos (2002) also showed that 56.1% COD removal occurred in the first chamber. Biogas production rate increased with increasing load, from 1.27 to 4.46 L L−1 day−1 (litres per litre of reactor per day), for OLR of 2.15 and 6.29 kg COD
205
Fig. 5. Variation of biogas production with OLR.
m−3 day−1 . Biogas production rate began to decrease when OLR was higher than 6.29 kg COD m−3 day−1 (Fig. 5). The methane content of the biogas varied between 59 and 73% throughout the experiment. 3.3. Effect of NaHCO3 /COD ratios on the performance of ABR The alkalinity in influent of reactor A was progressively decreased through decreasing additive of NaHCO3 in inflow and the treatment capacity was determined. The main monitoring parameters observed during each alkalinity period were summarized in Table 5. During these periods, the OLR and HRT were constant at about 2.4 kg COD m−3 day−1 and 20 h, respectively. The influent NaHCO3 /COD ratios decreased from 0.5 to 0.05. From the data presented in Table 5, it can be observed that the pH and alkalinity values in each compartment were generally decreased with the decreasing of NaHCO3 /COD ratio. Information on the effect of different NaHCO3 concentration on the performance of ABR shows that the ABR could be successfully anaerobic operated when the NaHCO3 concentration decreased to 100 mg L−1 (NaHCO3 /COD ratio 0.05). When the COD concentration was about 2000 mg L−1 , over 87% COD removal was obtained and influent NaHCO3 /COD ratios ranging from 0.5 to 0.05. No significant variations in the overall COD removal were observed due to the decrease of influent NaHCO3 /COD ratios. Fig. 6 shows the effect of influent NaHCO3 on relative COD removals in each compartment, as can be seen there were no obvious trends of the relative COD
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
206
Table 5 Mean reactor performance characteristics at each influent alkalinity Influent NaHCO3 / COD (g g−1 )
Influent COD (mg L−1 )
COD (mg L−1 )
0.5
1937
A1 A2 A3 A4
867 577 353 247
A1 A2 A3 A4
6.63 6.72 6.78 6.81
A1 A2 A3 A4
1025 1311 1375 1472
A1 A2 A3 A4
570 477 276 375
0.35
2168
A1 A2 A3 A4
1035 495 242 260
A1 A2 A3 A4
6.23 6.44 6.54 6.56
A1 A2 A3 A4
557 749 856 832
A1 A2 A3 A4
648 307 210 131
0.25
1837
A1 A2 A3 A4
1009 503 263 185
A1 A2 A3 A4
5.85 6.38 6.45 6.43
A1 A2 A3 A4
214 462 527 524
A1 A2 A3 A4
632 195 121 90
0.2
2155
A1 A2 A3 A4
758 516 213 215
A1 A2 A3 A4
5.72 6.19 6.38 6.44
A1 A2 A3 A4
364 507 594 626
A1 A2 A3 A4
390 156 90 18
0.1
2110
A1 A2 A3 A4
958 449 213 220
A1 A2 A3 A4
5.11 5.81 6.2 6.21
A1 A2 A3 A4
27 192 318 319
A1 A2 A3 A4
302 204 164 104
0.05
2024
A1 A2 A3 A4
1349 458 134 88
A1 A2 A3 A4
4.88 5.96 6.22 6.26
A1 A2 A3 A4
0 237 374 402
A1 A2 A3 A4
736 280 144 142
pH
removals with the decrease in influent NaHCO3 /COD ratios. Most of the influent COD was removed in compartment 1 (51–72%), with smaller removals occurring in other compartments (Fig. 6).
Fig. 6. Variation of relative COD removal with NaHCO3 /COD ratio in compartments.
Alkalinity (mg CaCO3 L−1 )
VFA (mg HAc L−1 )
3.4. The sludge characteristics At the end of the experiment, sludge samples from each compartment were taken across the length of the reactors to examine the characteristics and distribution of biomass. The TSS and VSS changes showed different trends down the two reactors (Table 6). For reactor A, compartment 2 has the maximal biomass concentration. However, in reactor B, the first chamber produced more biomass than others chamber. This may be attributed to the fact that the growth rate of acidogenic bacteria is greater than that of methanogenic bacteria at higher OLR. Uyanik et al. (2002) have also reported that the first two compartments have much more VSS than the final two chambers at 3.04–14.4 kg COD m−3 day−1 of OLR in a four-chamber ABR. The average ratios of VSS/TSS were 83.8 and 81.0% for reactor A and B, much more than the ratios (65.5 and 61.4% for reactor A and B) by the end of start-up stage (Table 3). As can be seen from TSS and VSS values
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
207
Table 6 The biomass concentrations by the end of the experiment Solids
A1
A2
A3
A4
B1
B2
B3
B4
TSS (g L−1 ) VSS (g L−1 ) VSS/TSS (%)
58.6 49.7 84.8
79.2 71.8 90.8
77.3 66.0 85.5
33.4 24.8 74.3
85.1 63.8 74.9
62.5 50.6 80.9
34.3 29.0 84.6
30.9 25.8 83.7
Table 7 The particle size distribution of granule by the end of experiment Particle size (mm)
<0.5 0.5–1.0 1.0–2.0 >2.0
The particle size distribution of granule (%) A1
A2
A3
A4
B1
B2
B3
B4
10.4 7.9 18.3 63.4
10.8 11.2 23.0 55.0
27.1 16.7 13.0 43.2
29.0 25.0 13.5 32.5
24.8 32.5 23.5 19.2
18.3 27.4 24.3 30.0
17.7 24.6 33.1 24.6
13.7 17.1 33.2 36.0
in Tables 3 and 6, the solids concentration increased greatly in all compartments during the performance study stage. Table 7 shows the particle size distribution of granular sludge. Most granules ranged in size between 0.5 and 4 mm, with compartment 1 and 2 containing mostly larger sized granules in reactor A, and compartment 4 containing mostly larger granules in reactor B. This pattern is different from the distribution of granules by the end of the start-up stage.
4. Conclusion Granulation of sludge was studied in continuous operation of two-laboratory scale ABRs with addition of GAC or bentonite and polyacrylamide into inoculated sludge, respectively. Performance of ABR at constant HRT, different OLR (or different influent COD concentrations) and NaHCO3 /COD ratios were investigated, and the following results were obtained. (1) Granulation of sludge was effectively performed and granulation was achieved in each compartment of two ABRs by day 75. It seemed to be more effective in promoting the formation of anaerobic granules with addition of GAC than bentonite and polyacrylamide. Granules had a maximum diameter of 6.0 mm at the end of start-up period. The granules size and the biomass concentration in each compartment increased significantly during
the performance experiment period. By the end of the experiment, the average VSS concentrations were 53.1 and 42.3 g L−1 in reactor A and B. (2) The ABRs had partial phase separation through compartmentalisation. And microbial selection and zoning are encouraged inside the reactors, with the acidogens in chambers closer to the inlet. (3) The ABR proved to be an efficient reactor configuration for the treatment of a sucrose based synthetic wastewater. With organic loading rate of 2.15–6.29 kg COD m−3 day−1 , the COD removals were more than 91%. (4) The ABR could be successfully anaerobic operated when the NaHCO3 concentration in influent decreased to 100 mg L−1 (NaHCO3 /COD ratio 0.05). When the COD concentration was about 2000 mg L−1 , over 87% COD removal was obtained, and influent NaHCO3 /COD ratios ranging from 0.5 to 0.05. No significant variations in the overall COD removal were observed due to the decrease of influent NaHCO3 /COD ratios from 0.5 to 0.05.
Acknowledgment The authors would like to express their thanks to NSFC (National Natural Science Foundation of China) for financial support by the foundation item No. 40272108.
208
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
References Akunna, J.C., Clark, M., 2000. Performance of a granular-bed anaerobic baffled reactor (GRABBR) treating whisky distillery wastewater. Bioresour. Technol. 74, 257–261. American Public Health Association, 1995. Standard Methods for Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC. Barber, W.P., Stuckey, D.C., 1999. The use of the anaerobic baffled reactor (ABR) for wastewater treatment: a review. Water Res. 33 (7), 1559–1578. Batstone, D.J., Keller, J., 2001. Variation of bulk properties of anaerobic granules with wastewater type. Water Res. 35 (7), 1723–1729. Bhattacharya, S.K., Madura, R.L., Walling, D.A., Farrell, J.B., 1996. Volatile solids reduction in two-phase and conventional anaerobic sludge digestion. Water Res. 30 (5), 1041–1048. Grobicki, A.M.W., Stuckey, D.C., 1991. Performance of the anaerobic baffled reactor under steady state and shock loading conditions. Biotechnol. Bioeng. 37, 344–355. Imai, T., Ukita, M., Liu, J., Sekine, M., Nakanishi, H., Fukagawa, M., 1997. Advanced start-up of UASB reactors by adding of water absorbing polymer. Water Sci. Technol. 36 (6–7), 399–406. Ince, O., 1998. Performance of a two-phase anaerobic digestion system when treating dairy wastewater. Water Res. 32 (9), 2707–2713. Jeyaseelan, S., Matsuo, T., 1995. Effects of phase separation in anaerobic digestion on different substrates. Water Sci. Technol. 31 (9), 153–162.
Liu, Y., Xu, H.L., Yang, S.F., Tay, J.H., 2003. Mechanisms and models for anaerobic granulation in upflow anaerobic sludge blanket reactor. Water Res. 37, 661–673. Manariotis, I.D., Grigoropoulos, S.G., 2002. Low-strength wastewater treatment using an anaerobic baffled reactor. Water Environ. Res. 74 (2), 170–176. Orozco, A., 1988. Anaerobic wastewater treatment using an open plug flow baffled reactor at low temperature. In: 5th International Symposium on Anaerobic Digestion, Bologna, Italy, pp. 759–762. Pereboom, J.H.F., 1997. Strength characterization of microbial granules. Water Sci. Technol. 36 (6–7), 141–148. Sallis, P.J., Uyanik, S., 2003. Granule development in a splitfed anaerobic baffled reactor. Bioresour. Technol. 89, 255– 265. Uyanik, S., Sallis, P.J., Anderson, G.K., 2002. The effect of polymer addition on granulation in an anaerobic baffled reactor (ABR). Part I: process performance. Water Res. 36, 933–943. Vossoughi, M., Shakeri, M., Alemzadeh, I., 2003. Performance of anaerobic baffled reactor treating synthetic wastewater influenced by decreasing COD/SO4 ratios. Chem. Eng. Process. 42, 811–816. Wang, J.L., Huang, Y.H., Zhao, X., 2004. Performance and characteristics of an anaerobic baffled reactor. Bioresour. Technol. 93, 205–208. Wirtz, R.A., Dague, R.R., 1996. Enhancement of granulation and start-up in the anaerobic sequencing batch reactor. Water Environ. Res. 68 (5), 883–892.
Granule development and performance in sucrose fed anaerobic baffled reactors Zonglian She ∗ , Xilai Zheng, Bairen Yang, Chunji Jin, Mengchun Gao Laboratory of Marine Environmental Science and Ecology, College of the Environmental Science and Engineering, Ocean University of China, No. 5 Yushan Road, Qingdao 266003, China Received 27 May 2005; received in revised form 27 August 2005; accepted 13 September 2005
Abstract Two 90 L anaerobic baffled reactors were used to study the granulation of sludge and the effect of the organic loading rate and NaHCO3 /COD ratios on reactor performance. Furthermore, it was determined whether an anaerobic baffled reactor would promote phase separation and if additive of bentonite or granular active carbon was capable of enhancing granule formation. In order to minimize feed variations, and have a totally biodegradable substrate, a synthetic sucrose substrate was used. Granulation was achieved in both reactors within 75 days. However, the granules from the granular active carbon amended reactor appeared earlier and were larger and more compact. The reactors were maintained at a hydraulic retention time of 20 h during performance study stage. The results showed that when organic loading rate were changed from 2.15 to 6.29 kg COD m−3 day−1 , chemical oxygen demand (COD) removal was not decreased (91–93%), but a slight increase in effluent COD was observed. It was found that the COD removals were generally good (87–92%) and had not obviously change with the decreasing NaHCO3 /COD ratios. From the bacterial distribution and the concentration of volatile fatty acids in four compartments, it was concluded that a separation of phases occurred within the anaerobic baffled reactors. © 2005 Published by Elsevier B.V. Keywords: Anaerobic baffled reactor; Granulation of sludge; Performance; Organic loading rate; NaHCO3 /COD ratios
1. Introduction The anaerobic baffled reactor (ABR) is one of the many types of high-rate anaerobic reactors and able to separate hydraulic retention time (HRT) from solid ∗ Corresponding author. Tel.: +86 532 2032571; fax: +86 532 2032102. E-mail address: [email protected] (Z. She).
0168-1656/$ – see front matter © 2005 Published by Elsevier B.V. doi:10.1016/j.jbiotec.2005.09.004
retention time (SRT). This allows anaerobic microorganism to remain within the reactor and produces higher volumetric loads and significantly enhanced removal efficiencies. The ABR was initially developed by McCarty and co-workers at Stanford University. In the ABR, a series of vertical baffles forces the wastewater to flow under and over them as it passes from inlet to outlet. This configuration has been shown to result in a high degree
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
of chemical oxygen demand (COD) removal at high organic loading rate (OLR) (Grobicki and Stuckey, 1991; Uyanik et al., 2002). The main advantage of using an ABR comes from its compartmentalized structure. This leads to separating acidogenesis and methanogenesis longitudinally down the reactor, allowing the different bacterial groups to develop under most favorable conditions. This characteristic is similar to two-phase digestion process. A number of advantages of the twophase digestion process have been summarized in the literature (Jeyaseelan and Matsuo, 1995; Bhattacharya et al., 1996; Ince, 1998). Granulation is a complex process, in which suspended biomass agglutinates to form discrete macroscopic aggregates. This enhances the settleability of the biomass and leads to an effective retention of bacteria in the reactor. It is widely assumed that granular biomass have operational advantages over the use of flocculent biomass (Wirtz and Dague, 1996; Pereboom, 1997; Batstone and Keller, 2001). The most significant advantage of granulation is high biomass concentration in continuous reactors, and then high treatment efficiencies can be expected. Upflow anaerobic sludge blanket (UASB) reactors are generally considered to be the best design for promoting granulation. The ABR may be classified as a series of UASB reactors. So, the ABR has shown the potential to produce granular sludge (Barber and Stuckey, 1999). The objectives of this study were: (1) to examine effects with addition of granular active carbon (GAC) or bentonite and polyacrylamide on the granulation of sludge in ABR; (2) to investigate the performance of ABR during various OLR and (3) to investigate the performance of ABR at different NaHCO3 /COD ratios in influent. In order to achieve these goals, two ABRs having identical dimensions and configurations were used.
2. Materials and methods 2.1. The reactors The ABRs were made of clear acrylic plastic with total effective volume of 90 L. Every reactor contained three vertical standing baffles that divided it into four identical compartments (as shown schematically in Fig. 1). Within each compartment, downcomer and
199
Fig. 1. Schematic diagram of ABR. 1, Feed tank; 2, heater; 3, rotameter; 4, influent; 5, gas collection; 6, effluent; 7, supernatant sampling port; 8, sludge sampling port.
riser regions were created by a further slanted edge (45◦ ) vertical baffle in order to direct the flow evenly through the riser. The width of the downcomer was 6 cm and the riser was 17 cm. Each compartment was equipped with sampling ports that allowed biological solids and liquid samples to be withdrawn. The influent was feed from troughs in higher position and rotameters were used to control the influent feed rate to the ABR. The temperature of influent water in the troughs was heated to about 50 ◦ C in order to maintain the operating temperature of the first compartment at constant 34 ◦ C. The produced gas was collected via portholes in the top of the reactors and was recorded by wet gas meters. 2.2. Sampling and analyses Supernatant liquor and sludge samples were taken separately from each compartment for analysis to characterise the behaviour of the reactors. Analysis of supernatant liquor included COD, pH, temperature, alkalinity and volatile fatty acids (VFA). COD, pH and temperature were monitored every day. Alkalinity and VFA were analysed weekly. COD and pH of the influent and effluent were analysed every day, too. Analysis of sludge samples included suspended solids (SS), volatile suspended solids (VSS) and particle size distribution at the end of every stage. COD, SS, VSS, pH, VFA and alkalinity were determined as described in Standard Methods for Examination of Water and Wastewater published by American Public Health Association (1995).
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
200
2.3.1. The development of granular sludge The reactors were seeded with anaerobically digested sewage sludge taken from a primary anaerobic digester at a local municipal sewage treatment works. The sludge was first sieved (<3 mm) to remove debris and large particles and was then introduced into all four compartments of each reactor. Each 22.5 L compartment contained 11.2 L sludge having a suspended solids composition of 28,700 mg SS L−1 and 14,050 mg VSS L−1 giving a total of 629 g VSS in each reactor. This value (6.99 g VSS L−1 ) is in accordance with the initial VSS values used in other studies on ABR (Barber and Stuckey, 1999; Sallis and Uyanik, 2003). At the same time of seeding sludge, reactor A was fed with GAC (Ø0.2–0.4 mm, 134.4 g) and reactor B was fed with bentonite (5‰, 9 L) and polyacrylamide (0.5‰, 9 L) for accelerating formation of granular sludge. Mean particle size of the bentonite was 5.2 m and the molecular weight of polyacrylamide was 1200 × 104 . After seeding the reactors, the covers were sealed and then synthetic wastewater was introduced into the reactors. Initially, the reactors were fed with low COD concentration (1500 mg L−1 ) and had long hydraulic retention time (30 h) in order to encourage the growth of granular biomass. Thereafter, influent feed strength was increased and at the same time the HRT was reduced in a stepwise fashion. Operational parameters during the stage of granulation were given in Table 1.
For estimating the size distribution, the sludge samples taken from the bottom sampling points of each chamber were classified into four fractions using laboratory sieves with various openings (0.5, 1.0 and 2.0 mm). The sludge particles were first placed in the sieve with the biggest opening (2.0 mm). The particles were gently submerged in water and shaken to let the smaller particles pass through. The procedures were repeated until the three sieves were used. After granulation of sludge, sludge samples were taken from each compartment of both reactors and were examined by scanning electron microscope (SEM). The samples were first fixed with 2.5% glutaric dialdehyde solution, and dehydrated through a graded series of water–ethanol mixtures. The samples were dried by the critical-piont drying method before sputter-coating with gold particles. The samples were then examined in a SEM. 2.3. Operational methods The ABRs were fed with a sucrose based synthetic medium, which contained nutrition (COD:N:P = 200:5:1) and trace metals. The trace metals was added into the influent as the following (mg per gram COD): CaCl2 , 10.0; MgSO4 ·7H2 O, 20.0; FeSO4 ·7H2 O, 1.5; MnSO4 ·H2 O, 0.5; H3 BO3 , 0.5; ZnCl2 , 0.5; CuCl2 ·2H2 O, 0.5; (NH4 )2 Mo7 O2 ·4H2 O, 0.5; CoCl2 ·6H2 O, 0.3 and NiCl2 ·6H2 O, 0.5. The experiment had two stages. The first stage is the development of granular sludge (start-up stage) and the second is performance study stage. During the second stage, the HRT was kept constant at 20 h while influent NaHCO3 was reduced in one reactor (A) and OLR was increased in another reactor (B). For reactor A, NaHCO3 concentrations in feed varied from 970 to 100 mg L−1 .
2.3.2. Performance study of varying NaHCO3 /COD ratios or OLR After granulation of sludge, the investigation in this study was carried out with a constant HRT (20 h) and varying NaHCO3 /COD ratios in influent in reactor A, and varying influent COD concentration (i.e. varying OLR) in reactor B. Reactor A was fed with sucrose synthetic influent containing about 2000 mg L−1 COD at
Table 1 Operational parameters during granulation process Steps
Times (day)
Influent COD (mg L−1 )
HRT (h)
OLR (kg COD m−3 day−1 )
Liquid upflow velocities (m h−1 )
1 2 3
1–25 26–55 56–75
1500 2000 3000
30 20 16
1.2 2.4 4.5
0.074 0.11 0.14
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
201
Table 2 The size distribution (by weight) of granular sludge by the end of start-up The size distribution of granular sludge (%)
Particle size (mm)
<0.5 0.5–1.0 1.0–2.0 >2.0
A1
A2
A3
A4
B1
B2
B3
B4
29.2 35.1 22.8 12.9
39.2 28.7 24.8 8.3
59.8 22.9 14.5 2.8
55.7 28.8 11.7 3.8
36.9 33.8 18.6 10.7
44.5 32.3 15.2 8.0
70.2 18.9 9.3 1.6
66.5 25.9 6.8 0.8
different NaHCO3 /COD ratios (0.5, 0.35, 0.25, 0.2, 0.1 and 0.05 g NaHCO3 /gCOD). Reactor B was fed with sucrose synthetic influent (containing 0.5 g NaHCO3 /g COD) at different values of COD (1793, 3008, 4053, 5245, 5868 and 8012 mg L−1 ) corresponding to OLR of about 2.15, 3.61, 4.86, 6.29, 7.04 and 9.61 kg COD m−3 day−1 . For each NaHCO3 /COD ratio and OLR studied, steady state performance was marked by near constant effluent pH values and COD values with less than 10% variation. Six samples of influent and effluent were usually collected and analysed during steady state. The average values of the six samples were regarded as the steady state performance data. During the studies (including start-up), dilution of the influent was carried out with tap water. It showed a steady decrease in temperature down the reactors (A1 34 ◦ C, A2 32 ◦ C, A3 30 ◦ C, A4 29 ◦ C, B1 34 ◦ C, B2 31 ◦ C, B3 29 ◦ C and B4 26 ◦ C).
3. Results and discussion 3.1. Granulation of sludge The development of granule sludge had three steps (Table 1). Granulation was first noted on day 45 in compartment 1 and 2 of reactor A and compartment 1 of reactor B, and after further 10 days granule-like flocs started to appear in all compartments of both reactors. By day 75 of the experiment all compartments in both reactors contained predominantly granular sludge.
Sludge samples were taken from each compartment of both reactors on day 75 of operation in order to study performance of the granular sludge. The size distribution (by weight) of the granules in both reactors was shown in Table 2 on day 75 of operation. The maximum size of granules was 6.0 mm in both reactors. The granules from reactor A appeared earlier and were larger and more compact. It seemed to be more effective in promoting the formation of anaerobic granules with addition of GAC than bentonite and polyacrylamide into inoculated sludge. According to the inert nuclei model for anaerobic granulation, the presence of micro-size biocarrier for bacterial attachment is a first step towards anaerobic granulation and the contact between particles and biomass could be improved using nuclei of lower density than sand and other inert materials (Liu et al., 2003; Imai et al., 1997). The percentages of granules in different sizes were determined through calculating the ratios between the weight of the granules in a range of size and the weight of granules in all sizes. It was showed a steady decrease in particle size down both reactors. About 35.7 and 29.3% (on weight basis, w/w) of the granules from chamber 1 of reactor A and B were larger than 1.0 mm and only 15.5 and 7.6% of granules from chamber 4 were above 1.0 mm. Orozco (1988) reported a similar decrease in granule size treating dilute carbohydrate waste. The solids concentration and VSS/TSS values appeared a steady decrease down both reactors, too (Table 3). The maximum sludge concentrations were 50.1 and 45.1 g TSS L−1 in compartment 1 of reactor A and B.
Table 3 The solids concentrations in each compartment by the end of start-up Solids
A1
A2
A3
A4
B1
B2
B3
B4
TSS (g L−1 ) VSS (g L−1 ) VSS/TSS (%)
50.1 37.8 75.4
45.6 31.5 69.1
40.5 25.3 62.5
35.4 19.5 55.1
45.1 31.4 69.6
40.1 26.5 66.1
35.9 20.1 56.0
32.2 17.4 54.0
202
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 2. Scanning electron photomicrographs showing the external surfaces of granular sludge in both reactors.
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 3. Scanning electron photomicrographs showing the bacterium of granular sludge.
203
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
204
showed that microbial selection and zoning are encouraged inside the ABRs.
The seed sludge was black, and the surface of granular sludge was grey in compartment 1 of reactor A while black in others compartments of both reactors. The SEM photomicrographs of granule samples taken from every compartment of both reactors are shown in Figs. 2 and 3. The granules had compact structure and naturally occurring cavities on the surface, as illustrated in Fig. 2. All granules examined were composed of various types of bacteria including bacilli, cocci and filaments. The granules in the initial compartments of the reactors were mainly composed of acidogenic short rod-shaped and filamentous bacteria, as illustrated in Fig. 3A1, B1 and B2. On the basis of appearance, in chambers 2, 3 and 4 of reactor A, the preponderant bacteria were similar to Methanobacterium. In compartment 3 of reactor B, granules were dominated presumably by Methanococcus and Methanobacterium. In compartment 4 of reactor B, the preponderant bacteria were similar to Methanobacterium. This result
3.2. Effect of OLR on the performance of ABR In order to determine the treatment capacity, the OLR in reactor B was progressively increased by increasing the strength of the feed. The mean performance data of reactor B throughout the entire length of the reactor during each OLR period is given in Table 4. The VFA concentrations decreased longitudally down the reactor. The highest VFA concentration was found in the first compartment with average values of 612–1548 mg L−1 . The VFA concentrations in every compartment increased with the increasing OLR. The VFA data demonstrated that hydrolysis and acidogenesis were the main biochemical activities occurring in
Table 4 Mean reactor performance characteristics at each OLR OLR (kg COD m−3 day−1 )
Influent COD (mg L−1 )
COD (mg L−1 )
pH
2.15
1793
B1 B2 B3 B4
1014 434 221 59
B1 B2 B3 B4
6.08 6.55 6.61 6.61
B1 B2 B3 B4
462 703 736 774
B1 B2 B3 B4
612 251 152 88
3.61
3008
B1 B2 B3 B4
1808 566 274 196
B1 B2 B3 B4
5.47 6.6 6.65 6.64
B1 B2 B3 B4
323 728 946 1064
B1 B2 B3 B4
1245 498 319 193
4.86
4053
B1 B2 B3 B4
2888 1304 534 404
B1 B2 B3 B4
4.56 5.98 6.39 6.44
B1 B2 B3 B4
0 514 790 855
B1 B2 B3 B4
1036 564 432 224
6.29
5245
B1 B2 B3 B4
3587 1473 411 356
B1 B2 B3 B4
3.95 5.56 6.26 6.33
B1 B2 B3 B4
0 359 674 718
B1 B2 B3 B4
1526 714 334 122
7.04
5868
B1 B2 B3 B4
4654 2994 1874 1187
B1 B2 B3 B4
4.33 5.16 5.78 6.06
B1 B2 B3 B4
0 174 455 656
B1 B2 B3 B4
1305 1191 940 574
9.61
8012
B1 B2 B3 B4
6153 3254 2362 1818
B1 B2 B3 B4
3.81 4.88 6.03 6.15
B1 B2 B3 B4
0 85 442 612
B1 B2 B3 B4
1548 1524 804 552
VFA (mg HAc L−1 )
Alkalinity (mg CaCO3 L−1 )
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
Fig. 4. Variation with OLR of relative COD removal in compartments.
the first and first two compartments at low and high OLR. The methanogenesis appeared to be dominant in the third and fourth compartments. These observations suggest that the ABR system promoted a systematic selection in the different compartments in such a manner as to bring about phase separation. Similar observations have been reported in other studies (Akunna and Clark, 2000; Wang et al., 2004). The alkalinity values appeared a significant increasing step by step from compartment 1 to 4. Furthermore, the increasing of OLR resulted in the significant reduction of alkalinity in compartment 1–2. The COD removals were generally very good (91–93%) when OLR was lower than 6.29 kg COD m−3 day−1 . The best performance was observed with influent COD of 3008 mg L−1 (or a loading of 3.61 kg COD m−3 day−1 ). When the influent COD increased to 5245 mg L−1 (or a loading of 6.29 kg COD m−3 day−1 ) the efficiency of COD removals began to drop. It was found that most organic matter was removed in the first two chambers, and COD removal in the four compartments was 35.0, 41.6, 22.5 and 1.2% based on the overall removal at OLR of 6.29 kg COD m−3 day−1 (Fig. 4). Vossoughi et al. (2003) who used a five-chamber ABR to treat synthetic wastewater at an HRT of 1 day, have also reported that over 65% COD removal was obtained in the first chamber. Manariotis and Grigoropoulos (2002) also showed that 56.1% COD removal occurred in the first chamber. Biogas production rate increased with increasing load, from 1.27 to 4.46 L L−1 day−1 (litres per litre of reactor per day), for OLR of 2.15 and 6.29 kg COD
205
Fig. 5. Variation of biogas production with OLR.
m−3 day−1 . Biogas production rate began to decrease when OLR was higher than 6.29 kg COD m−3 day−1 (Fig. 5). The methane content of the biogas varied between 59 and 73% throughout the experiment. 3.3. Effect of NaHCO3 /COD ratios on the performance of ABR The alkalinity in influent of reactor A was progressively decreased through decreasing additive of NaHCO3 in inflow and the treatment capacity was determined. The main monitoring parameters observed during each alkalinity period were summarized in Table 5. During these periods, the OLR and HRT were constant at about 2.4 kg COD m−3 day−1 and 20 h, respectively. The influent NaHCO3 /COD ratios decreased from 0.5 to 0.05. From the data presented in Table 5, it can be observed that the pH and alkalinity values in each compartment were generally decreased with the decreasing of NaHCO3 /COD ratio. Information on the effect of different NaHCO3 concentration on the performance of ABR shows that the ABR could be successfully anaerobic operated when the NaHCO3 concentration decreased to 100 mg L−1 (NaHCO3 /COD ratio 0.05). When the COD concentration was about 2000 mg L−1 , over 87% COD removal was obtained and influent NaHCO3 /COD ratios ranging from 0.5 to 0.05. No significant variations in the overall COD removal were observed due to the decrease of influent NaHCO3 /COD ratios. Fig. 6 shows the effect of influent NaHCO3 on relative COD removals in each compartment, as can be seen there were no obvious trends of the relative COD
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
206
Table 5 Mean reactor performance characteristics at each influent alkalinity Influent NaHCO3 / COD (g g−1 )
Influent COD (mg L−1 )
COD (mg L−1 )
0.5
1937
A1 A2 A3 A4
867 577 353 247
A1 A2 A3 A4
6.63 6.72 6.78 6.81
A1 A2 A3 A4
1025 1311 1375 1472
A1 A2 A3 A4
570 477 276 375
0.35
2168
A1 A2 A3 A4
1035 495 242 260
A1 A2 A3 A4
6.23 6.44 6.54 6.56
A1 A2 A3 A4
557 749 856 832
A1 A2 A3 A4
648 307 210 131
0.25
1837
A1 A2 A3 A4
1009 503 263 185
A1 A2 A3 A4
5.85 6.38 6.45 6.43
A1 A2 A3 A4
214 462 527 524
A1 A2 A3 A4
632 195 121 90
0.2
2155
A1 A2 A3 A4
758 516 213 215
A1 A2 A3 A4
5.72 6.19 6.38 6.44
A1 A2 A3 A4
364 507 594 626
A1 A2 A3 A4
390 156 90 18
0.1
2110
A1 A2 A3 A4
958 449 213 220
A1 A2 A3 A4
5.11 5.81 6.2 6.21
A1 A2 A3 A4
27 192 318 319
A1 A2 A3 A4
302 204 164 104
0.05
2024
A1 A2 A3 A4
1349 458 134 88
A1 A2 A3 A4
4.88 5.96 6.22 6.26
A1 A2 A3 A4
0 237 374 402
A1 A2 A3 A4
736 280 144 142
pH
removals with the decrease in influent NaHCO3 /COD ratios. Most of the influent COD was removed in compartment 1 (51–72%), with smaller removals occurring in other compartments (Fig. 6).
Fig. 6. Variation of relative COD removal with NaHCO3 /COD ratio in compartments.
Alkalinity (mg CaCO3 L−1 )
VFA (mg HAc L−1 )
3.4. The sludge characteristics At the end of the experiment, sludge samples from each compartment were taken across the length of the reactors to examine the characteristics and distribution of biomass. The TSS and VSS changes showed different trends down the two reactors (Table 6). For reactor A, compartment 2 has the maximal biomass concentration. However, in reactor B, the first chamber produced more biomass than others chamber. This may be attributed to the fact that the growth rate of acidogenic bacteria is greater than that of methanogenic bacteria at higher OLR. Uyanik et al. (2002) have also reported that the first two compartments have much more VSS than the final two chambers at 3.04–14.4 kg COD m−3 day−1 of OLR in a four-chamber ABR. The average ratios of VSS/TSS were 83.8 and 81.0% for reactor A and B, much more than the ratios (65.5 and 61.4% for reactor A and B) by the end of start-up stage (Table 3). As can be seen from TSS and VSS values
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
207
Table 6 The biomass concentrations by the end of the experiment Solids
A1
A2
A3
A4
B1
B2
B3
B4
TSS (g L−1 ) VSS (g L−1 ) VSS/TSS (%)
58.6 49.7 84.8
79.2 71.8 90.8
77.3 66.0 85.5
33.4 24.8 74.3
85.1 63.8 74.9
62.5 50.6 80.9
34.3 29.0 84.6
30.9 25.8 83.7
Table 7 The particle size distribution of granule by the end of experiment Particle size (mm)
<0.5 0.5–1.0 1.0–2.0 >2.0
The particle size distribution of granule (%) A1
A2
A3
A4
B1
B2
B3
B4
10.4 7.9 18.3 63.4
10.8 11.2 23.0 55.0
27.1 16.7 13.0 43.2
29.0 25.0 13.5 32.5
24.8 32.5 23.5 19.2
18.3 27.4 24.3 30.0
17.7 24.6 33.1 24.6
13.7 17.1 33.2 36.0
in Tables 3 and 6, the solids concentration increased greatly in all compartments during the performance study stage. Table 7 shows the particle size distribution of granular sludge. Most granules ranged in size between 0.5 and 4 mm, with compartment 1 and 2 containing mostly larger sized granules in reactor A, and compartment 4 containing mostly larger granules in reactor B. This pattern is different from the distribution of granules by the end of the start-up stage.
4. Conclusion Granulation of sludge was studied in continuous operation of two-laboratory scale ABRs with addition of GAC or bentonite and polyacrylamide into inoculated sludge, respectively. Performance of ABR at constant HRT, different OLR (or different influent COD concentrations) and NaHCO3 /COD ratios were investigated, and the following results were obtained. (1) Granulation of sludge was effectively performed and granulation was achieved in each compartment of two ABRs by day 75. It seemed to be more effective in promoting the formation of anaerobic granules with addition of GAC than bentonite and polyacrylamide. Granules had a maximum diameter of 6.0 mm at the end of start-up period. The granules size and the biomass concentration in each compartment increased significantly during
the performance experiment period. By the end of the experiment, the average VSS concentrations were 53.1 and 42.3 g L−1 in reactor A and B. (2) The ABRs had partial phase separation through compartmentalisation. And microbial selection and zoning are encouraged inside the reactors, with the acidogens in chambers closer to the inlet. (3) The ABR proved to be an efficient reactor configuration for the treatment of a sucrose based synthetic wastewater. With organic loading rate of 2.15–6.29 kg COD m−3 day−1 , the COD removals were more than 91%. (4) The ABR could be successfully anaerobic operated when the NaHCO3 concentration in influent decreased to 100 mg L−1 (NaHCO3 /COD ratio 0.05). When the COD concentration was about 2000 mg L−1 , over 87% COD removal was obtained, and influent NaHCO3 /COD ratios ranging from 0.5 to 0.05. No significant variations in the overall COD removal were observed due to the decrease of influent NaHCO3 /COD ratios from 0.5 to 0.05.
Acknowledgment The authors would like to express their thanks to NSFC (National Natural Science Foundation of China) for financial support by the foundation item No. 40272108.
208
Z. She et al. / Journal of Biotechnology 122 (2006) 198–208
References Akunna, J.C., Clark, M., 2000. Performance of a granular-bed anaerobic baffled reactor (GRABBR) treating whisky distillery wastewater. Bioresour. Technol. 74, 257–261. American Public Health Association, 1995. Standard Methods for Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC. Barber, W.P., Stuckey, D.C., 1999. The use of the anaerobic baffled reactor (ABR) for wastewater treatment: a review. Water Res. 33 (7), 1559–1578. Batstone, D.J., Keller, J., 2001. Variation of bulk properties of anaerobic granules with wastewater type. Water Res. 35 (7), 1723–1729. Bhattacharya, S.K., Madura, R.L., Walling, D.A., Farrell, J.B., 1996. Volatile solids reduction in two-phase and conventional anaerobic sludge digestion. Water Res. 30 (5), 1041–1048. Grobicki, A.M.W., Stuckey, D.C., 1991. Performance of the anaerobic baffled reactor under steady state and shock loading conditions. Biotechnol. Bioeng. 37, 344–355. Imai, T., Ukita, M., Liu, J., Sekine, M., Nakanishi, H., Fukagawa, M., 1997. Advanced start-up of UASB reactors by adding of water absorbing polymer. Water Sci. Technol. 36 (6–7), 399–406. Ince, O., 1998. Performance of a two-phase anaerobic digestion system when treating dairy wastewater. Water Res. 32 (9), 2707–2713. Jeyaseelan, S., Matsuo, T., 1995. Effects of phase separation in anaerobic digestion on different substrates. Water Sci. Technol. 31 (9), 153–162.
Liu, Y., Xu, H.L., Yang, S.F., Tay, J.H., 2003. Mechanisms and models for anaerobic granulation in upflow anaerobic sludge blanket reactor. Water Res. 37, 661–673. Manariotis, I.D., Grigoropoulos, S.G., 2002. Low-strength wastewater treatment using an anaerobic baffled reactor. Water Environ. Res. 74 (2), 170–176. Orozco, A., 1988. Anaerobic wastewater treatment using an open plug flow baffled reactor at low temperature. In: 5th International Symposium on Anaerobic Digestion, Bologna, Italy, pp. 759–762. Pereboom, J.H.F., 1997. Strength characterization of microbial granules. Water Sci. Technol. 36 (6–7), 141–148. Sallis, P.J., Uyanik, S., 2003. Granule development in a splitfed anaerobic baffled reactor. Bioresour. Technol. 89, 255– 265. Uyanik, S., Sallis, P.J., Anderson, G.K., 2002. The effect of polymer addition on granulation in an anaerobic baffled reactor (ABR). Part I: process performance. Water Res. 36, 933–943. Vossoughi, M., Shakeri, M., Alemzadeh, I., 2003. Performance of anaerobic baffled reactor treating synthetic wastewater influenced by decreasing COD/SO4 ratios. Chem. Eng. Process. 42, 811–816. Wang, J.L., Huang, Y.H., Zhao, X., 2004. Performance and characteristics of an anaerobic baffled reactor. Bioresour. Technol. 93, 205–208. Wirtz, R.A., Dague, R.R., 1996. Enhancement of granulation and start-up in the anaerobic sequencing batch reactor. Water Environ. Res. 68 (5), 883–892.