- Email: [email protected]
Effects of feeding time and organic loading in an anaerobic sequencing batch biofilm reactor (ASBBR) treating diluted whey
ARTICLE IN PRESS
Journal of Environmental Management 85 (2007) 927–935 www.elsevier.com/locate/jenvman
Effects of feeding time and organic loading in an anaerobic sequencing batch biofilm reactor (ASBBR) treating diluted whey Leonardo H.S. Damascenoa, Jose´ A.D. Rodriguesb,, Suzana M. Ratuszneib, Marcelo Zaiata, Eugeˆnio Forestia a
Departamento de Hidra´ulica e Saneamento, Escola de Engenharia de Sa˜o Carlos, Universidade de Sa˜o Paulo (USP), Sa˜o Carlos-SP, Brazil Departamento de Engenharia Quı´mica e de Alimentos, Escola de Engenharia Maua´, Instituto Maua´ de Tecnologia, Caetano do Sul-SP, Brazil
b
Received 28 June 2005; received in revised form 23 October 2006; accepted 1 November 2006 Available online 20 December 2006
Abstract An investigation was carried out on the performance of an anaerobic sequencing batch biofilm reactor (ASBBR) treating diluted cheese whey when submitted to different feed strategies and volumetric organic loads (VOL). Polyurethane foam cubes were used as support for biomass immobilization and stirring was provided by helix impellers. The reactor with a working volume of 3 L treated 2 L of wastewater in 8-h cycles at 500 rpm and 30 C. The organic loads applied were 2, 4, 8 and 12 gCOD L1 d1 , obtained by increasing the feed concentration. Alkalinity was supplemented at a ratio of 50% NaHCO3 /COD. For each organic load applied three feed strategies were tested: (a) batch operation with 8-h cycle; (b) 2-h fed-batch operation followed by 6-h batch; and (c) 4-h fed-batch followed by 4-h batch. The 2-h fed-batch operation followed by 6-h batch presented the best results for the organic loads of 2 and 4 gCOD L1 d1 , whereas the 4-h fed-batch operation followed by 4-h batch presented results slightly inferior for the same organic loads and the best results at organic loads of 8 and 12 gCOD L1 d1 . The concentration of total volatile acids varied with fill time. For the higher fill times maximum concentrations were obtained at the end of the cycle. Moreover, no significant difference was detected in the maximum concentration of total volatile acids for any of the investigated conditions. However, the maximum values of propionic acid tended to decrease with increasing fill time considering the same organic load. Microbiological analyses revealed the presence of Methanosaeta-like structures and methanogenic hydrogenotrophic-like fluorescent bacilli. No Methanosarcina-like structures were observed in the samples. r 2006 Elsevier Ltd. All rights reserved. Keywords: ASBBR; Feeding time; Organic loading; Whey
1. Introduction Cheese whey possesses a high organic load with concentrations of approximately 60–80 gCOD L1 . This characteristic, combined with others such as low alkalinity and high biodegradability, make anaerobic treatment in high rate reactors difficult (Malaspina et al., 1996; Yan et al., 1988). Alkalinity present in the system is quickly consumed due to rapid conversion of lactose into shortchain volatile acids, making it necessary to constantly monitor alkalinity as well as pH (Backus et al., 1988). Supplemental addition of alkalinity as bicarbonate, carboCorresponding author. Tel.: +55 11 4239 3148; fax: +55 11 4239 3131.
E-mail address: [email protected] (J.A.D. Rodrigues). 0301-4797/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvman.2006.11.001
nate or hydroxide may also be necessary (Lo and Liao, 1986; Wildenauer and Winter, 1985). Factors that may significantly affect ASBR performance according to Zaiat et al. (2001) include feed strategy, agitation, reactor configuration and initial ratio between substrate and biomass concentrations (F/M). In batch or fed-batch operated reactors the influence of feed strategy is related to the F/M ratio. According to Angenent and Dague (1995), increase in feed time would cause a reduction in total volatile acids (TVA) and consequently increase reactor performance, since by supplying substrate within a longer time, availability to microorganisms will be less, occasioning less accumulation of these acids. Bagley and Brodkorb (1999) corroborated the assumptions of Angenent and Dague (1995) in easily
ARTICLE IN PRESS 928
L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
Nomenclature Symbols BA bicarbonate alkalinity, mgCaCO3 L1 CI influent substrate concentration, mgCOD L1 CFS filtered substrate concentration in the effluent, mgCOD L1 C TS non-filtered substrate concentration in the effluent, mgCOD L1 tF =tC ratios between feed time and cycle time, dimensionless TVA total volatile acids concentration determined by titrimetric method, mgHAc L1 TVAC total volatile acids concentration determined by chromatography method, mg L1 V Cycle feed volume per cycle, L VOL volumetric organic load, gCOD L1 d1
degraded wastewaters, suggesting, however, investigations in complex substrates. Kennedy et al. (1991) investigated the effect of feed time on the performance of an ASBR treating synthetic sucrose substrate with organic rates ranging from 2.5–18:5 gCOD L1 d1 and concentration of 7 gCOD L1 , varying filland-react periods (fill-to-react ratios of 0.2, 0.5 and 2). The process was not affected by volumetric organic loads (VOL) below 9 gCOD L1 d1 and for higher loads removal efficiency was reduced more than 25% for low ratios between feed time and cycle time ðtF =tC Þ. Suthaker et al. (1991) submitted an ASBR treating glucose based wastewater with a concentration of 35 gCOD L1 to different operation conditions (temperature from 25 to 34 C and tF =tC ratio from 0 to 0.75), with a VOL of 1:6 gCOD L1 d1 . The maximum conversion achieved was 73% for filtered samples for a 16-day cycle time and 4-day feed time. They concluded that feed time has a stronger influence than tF =tC ratio, being of fundamental importance in defining feed strategy. Shizas and Bagley (2002) also investigated an ASBR treating glucose based wastewater. The system operating with a VOL of 2:1 gCOD L1 d1 suffered overload at 3:2 gCOD L1 d1 . Reactor performance increased at high tF =tC and low initial substrate concentrations, indicating that operation can be optimized by altering these operation parameters. Rodrigues et al. (2003), operating an ASBR with granulated biomass treating low-strength wastewater ð0:5 gCOD L1 Þ, concluded that long feed times (tF =tC greater than 0.5) affects system performance related to organic matter removal efficiency, settleability characteristics and extra-cellular polymer synthesis. Ratusznei et al. (2003a), operating the ASBR with biomass immobilized on polyurethane foam and treating the same wastewater, observed a drop in efficiency at long feed times. This behavior was attributed to the time the immobilized mass
V Working working volume, L Greek characters eFS substrate removal efficiency considering filtered samples, % eTS substrate removal efficiency considering nonfiltered samples, % Abbreviations ASBR anaerobic sequencing batch reactor ASBBR anaerobic sequencing batch biofilm reactor COD chemical oxygen demand F/M ratio between substrate and biomass concentrations TSS total suspended solids VSS volatile suspended solids TS total solids TVS total volatile solids UASB upflow anaerobic sludge blanket
was exposed to air, since during feeding the bed was gradually immersed, occasioning periods at which the biomass had no contact with the wastewater; at short feed times the bed was rapidly immersed in the substrate, which did not occur at longer feed times. The main objective of this investigation was to assess the behavior of an ASBBR containing microorganisms immobilized on polyurethane foam treating diluted cheese whey when submitted to different VOL (2, 4, 8 and 12 gCOD L1 d1 Þ and feed strategies (fill times of 10 min, 2 h and 4 h by a cycle time of 8 h). 2. Materials and methods The reactor used was made of acrylic with internal diameter and height of 200 mm. A 100-mm high basket was used in the reactor to contain the biomass, as shown in the scheme in Fig. 1. This basket occupied half of the reactor volume to allow permanent immersion of the bed during fed-batch operation. The system operated in a chamber which maintained the temperature at 30 1 C. Stirring at 500 rpm was provided by two helix propellers. The helix propellers with diameter of 60 mm consisted of three blades; one positioned at 20 mm from the reactor bottom and the other above the bed at 120 mm from the reactor bottom. The inert support used consisted of 5-mm polyurethane foam cubes and was inoculated, according to methodology proposed by Zaiat et al. (1994), i.e., (i) the biomass was squeezed through a 1 mm mesh sieve as a way to crush the granules present and facilitate attachment of the biomass in the foam; (ii) the 5-mm foam cubes were added to the crushed inoculum and mixed in a way to saturate the foam completely with the sludge; and (iii) after 2 h the foam cubes containing the immobilized biomass were added to the reactor, taking care to avoid compression of the cubes. The sludge used was from a UASB reactor treating
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
929
Fig. 1. Scheme of the anaerobic sequencing batch biofilm reactor (ASBBR) containing immobilized biomass [(1) bioreactor, (2) stainless steel basket containing particles with immobilized cells, (3) impeller, (4) feed pump, (5) discharge pump, (6) substrate, (7) treated effluent, (8) timers].
wastewater from a poultry slaughterhouse that had total volatile solids (TVS) and total solids (TS) of 51 and 62 g L1 , respectively. The wastewater used consisted of reconstituted dehydrated cheese whey for which the ratio between dry whey and amount of COD was experimentally determined as 1:1. Alkalinity supplementation was performed by adding sodium bicarbonate, utilizing g-NaHCO3 /g-COD ratios proposed by Ratusznei et al. (2003b), i.e., to guarantee that bicarbonate alkalinity (BA) was not a limiting factor. This way, at the start of each phase a g-NaHCO3 /g-COD ratio of 100% was used and after experimental verification of stability this ratio was reduced to 50%. The following organic matter concentrations were used: 1 gCOD L1 (assay 1), 2 gCOD L1 (assay 2), 4 gCOD L1 (assay 3) and 6 gCOD L1 (assay 4). Bioreactor operation consisted of initially feeding 3 L wastewater to be treated in 8-h cycles. In each cycle 2 L wastewater were discharged; the remaining 1 L residual volume in the reactor was used to dilute the following 2 L wastewater fed in the next cycle. Feeding and discharge were performed by means of diaphragm pumps; discharge was done in 10 min. The study of feed influence was performed in the same bioreactor in three different periods, which defined the feeding strategies investigated. Each assay using a different feed strategy was performed with the same biomass without an acclimation period. (i) 8-h batch operation: Feeding of 2 L wastewater to be treated in 10 min and a g-NaHCO3 /g-COD ratio of 100% for initial alkalinity supplementation (phase I) and subsequently of 50% (phase II); (ii) 2-h fed-batch operation followed by 6-h batch: Feeding of 2 L wastewater to be treated in 2 h and a g-NaHCO3 / g-COD ratio of 50% for initial alkalinity supplementation (phase III);
(iii) 4-h fed-batch followed by 4-h batch: Feeding of 2 L wastewater to be treated in 4 h and a g-NaHCO3 / g-COD ratio of 50% for initial alkalinity supplementation (phase IV). This way, the VOL were approximately 2 gCOD L1 d1 (assay 1), 4 gCOD L1 d1 (assay 2), 8 gCOD L1 d1 (assay 3) and 12 gCOD L1 d1 (assay 4), determined by VOL ¼
V Cycle N Cycles=Day C I . V Useful
(1)
Monitoring was performed according to Standard Methods for the Examination of Water and Wastewater (1995). The following parameters were analyzed: organic matter concentration (COD) for filtered ðC FS Þ and non-filtered samples ðC TS Þ, pH, BA, TVA, TS, TVS, total suspended solids (TSS) and volatile suspended solids (VSS). Moreover, intermediate volatile acids and biogas composition were analyzed by gas phase chromatography. Composition of the biogas generated by anaerobic degradation was analyzed by gas chromatography using a Hewlett Packards 6890 gas chromatograph equipped with a thermal conductivity detector. Sample volume was 1 mL, drag gas was hydrogen at a flow rate of 50:0 mL h1 , and column, injector and detector temperatures were 35, 60 and 160 C, respectively. Volatile fatty acid (VFA) samples were analyzed by gas chromatography, using a gas chromatograph HP6890 with flame ionization detector at 300 C and an HP-INNOWAX column ð30 m 0:25 mm 0:25 mmÞ. The injector temperature was kept at 250 C; the oven was held at 100 C for 3 min, after which it was heated at a rate of 5 C min1 to 180 C, and held at that temperature for 5 min. After attainment of operation stability in each assay and in each implemented feed strategy, profiles along cycles were taken of COD concentrations for filtered samples, C S , pH, BA, TVA, intermediate volatile acids and biogas
ARTICLE IN PRESS 930
L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
composition. Microbiological identification was performed by means of common optical and phase contrast microscopy employing an Olympus BH2 microscope. 3. Results and discussion Average values of the monitored parameters are listed in Tables 1–4. Figs. 2 and 5 show the variation in organic matter concentration in the filtered and non-filtered samples, conversion efficiency, BA and TVA concentration in effluent samples for VOLs of 2–12 gCOD L1 d1 . For a VOL of 2 gCOD L1 d1 (assay 1—see Table 1 and Fig. 2) high efficiency was obtained in all phases, presenting low values of TVA concentration and increase in BA values of the effluent in relation to the influent. The operation with 2-h fed-batch followed by 6-h batch (phase III) presented improved operation stability when compared to the batch operation.
At the condition with VOL of 4 gCOD L1 d1 (assay 2—see Table 2 and Fig. 3) organic matter conversions in all phases were high, with improved performance for the 2-h fed-batch followed by 6-h batch (phase III). The system was stable presenting BA generation, reduced values of TVA and pH close to 7. Extra-cellular polymer formation, which affected the system performance, was also observed. This material was removed in the 42nd cycle (see Fig. 3). During the assay with VOL of 8 gCOD L1 d1 (assay 3—see Table 3 and Fig. 4) the system was less stable than in other conditions, but considering the ‘‘organic matter efficiency removal’’ profiles, the system may be considered slightly stable in all phases, despite a decrease in system efficiency and increase in TVA, both presenting a tendency to improve with longer fill times (phase IV). The high TVA concentrations did not occasion reactor instability, as no accumulation of these acids occurred. This fact was a consequence of the cycle time used, which did not allow
Table 1 Average values of the monitored variables in the assay with VOL of 2 gCOD L1 d1 (assay 1) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS (mgCOD L1 Þ eTS (%) eFS (%)
976 32ð17Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
44 6ð9Þ 280 21ð9Þ 8:3 0:2ð9Þ 1272 76ð9Þ 974 69ð9Þ 45 11ð9Þ 40 9ð9Þ –
177 65ð10Þ 78 29ð10Þ 82 7ð10Þ 92 3ð10Þ 27 8ð8Þ 334 14ð6Þ 7:2 0:4ð6Þ 761 54ð3Þ 430 70ð3Þ 77 29ð3Þ 65 36ð3Þ 71:2ð1Þ
150 28ð5Þ 41 8ð5Þ 85 3ð5Þ 96 1ð5Þ 30 6ð5Þ 341 42ð5Þ 6:8 0:2ð5Þ 693 31ð3Þ 372 55ð3Þ 95 36ð3Þ 91 34ð3Þ 72:9ð1Þ
192 5ð6Þ 85 15ð6Þ 80 1ð6Þ 91 1ð6Þ 59 21ð6Þ 295 14ð6Þ 6:8 0:2ð6Þ 622 103ð3Þ 313 69ð3Þ 83 8ð3Þ 75 17ð3Þ 71:2ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
Table 2 Average values of the monitored variables in the assay with VOL of 4 gCOD L1 d1 (assay 2) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS ðmgCOD L1 Þ eTS (%) eFS (%)
2005 64ð14Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
83 13ð10Þ 553 13ð10Þ 8:4 0:2ð10Þ 2345 369ð8Þ 1726 363ð8Þ 78 31ð8Þ 72 30ð8Þ –
265 67ð8Þ 115 45ð8Þ 87 3ð8Þ 94 2ð8Þ 56 26ð9Þ 653 23ð5Þ 7:6 0:1ð5Þ 1249 163ð4Þ 476 158ð4Þ 71 35ð4Þ 65 29ð4Þ 67:8ð1Þ
104 36ð3Þ 59 18ð3Þ 95 2ð3Þ 97 1ð3Þ 46 20ð4Þ 672 12ð4Þ 6:8 0:1ð4Þ 1117 61ð3Þ 502 48ð3Þ 98 11ð3Þ 93 10ð3Þ 71:8ð1Þ
266 68ð5Þ 147 36ð5Þ 87 3ð5Þ 93 2ð5Þ 90 13ð6Þ 626 12ð6Þ 6:8 0:1ð6Þ 1116 8ð2Þ 433 64ð2Þ 101 7ð2Þ 83 10ð2Þ 71ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
931
Table 3 Average values of the monitored variables in the assay with VOL of 8 gCOD L1 d1 (assay 3) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ CFS ðmgCOD L1 Þ eTS (%) eFS (%)
4020 108ð26Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
134 10ð12Þ 1103 40ð12Þ 8:4 0:3ð12Þ 4858 216ð11Þ 3430 182ð11Þ 175 11ð11Þ 153 15ð11Þ –
879 136ð12Þ 650 135ð12Þ 78 3ð12Þ 84 3ð12Þ 321 80ð14Þ 1103 121ð10Þ 6:8 0:1ð10Þ 2681 563ð5Þ 1018 112ð5Þ 187 49ð5Þ 166 36ð5Þ 63:7ð1Þ
963 144ð6Þ 615 142ð6Þ 76 4ð6Þ 85 4ð6Þ 301 67ð7Þ 1084 56ð7Þ 6:9 0:2ð7Þ 2402 226ð2Þ 928 209ð2Þ 226 71ð2Þ 195 75ð2Þ 65:5ð1Þ
904 278ð9Þ 566 285ð9Þ 78 7ð9Þ 86 7ð9Þ 285 97ð9Þ 1096 92ð9Þ 6:9 0:1ð9Þ 2282 334ð4Þ 978 85ð4Þ 252 8ð4Þ 226 7ð4Þ 67:8ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
Table 4 Average values of the monitored variables in the assay with VOL of 12 gCOD L1 d1 (assay 4) for batch mode (phase II), 2-h fed-batch (phase III) and 4h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS (mgCOD L1 Þ eTS (%) eFS (%)
5969 165ð19Þ – – –
TVA (mgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
204 28ð9Þ 1689 77ð9Þ 8:4 0:3ð9Þ 7132 154ð8Þ 4960 164ð8Þ 302 71ð8Þ 266 73ð8Þ –
1782 399ð8Þ 1215 444ð8Þ 70 7ð8Þ 78 8ð8Þ 527 181ð7Þ 1582 173ð7Þ 7:1 0:1ð7Þ 3537 352ð4Þ 1431 381ð4Þ 295 55ð4Þ 248 50ð4Þ 66:4ð1Þ
2055 113ð5Þ 1626 141ð5Þ 66 2ð5Þ 73 2ð5Þ 750 82ð6Þ 1429 72ð6Þ 7:1 0:2ð6Þ 3800 57ð2Þ 1602 34ð2Þ 318 37ð2Þ 276 8ð2Þ 63:6ð1Þ
1687 237ð5Þ 1118 181ð5Þ 72 4ð5Þ 81 3ð5Þ 602 93ð6Þ 1571 49ð6Þ 7:1 0:1ð6Þ 3502 238ð2Þ 1318 141ð2Þ 442 21ð2Þ 403 19ð2Þ 63:9ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
III
IV
I
II
III
IV
700
100
400
80
560
80
300
60
420
60
200
40
280
40
100
20
140
20
0 0
10 20 30 40 50 60 70 80 90 Operation cycles
0
BA (mgCaCO3.l -1)
100
ε (%)
Cs (mgCOD.l -1)
II
0 0
10 20 30 40 50 60 70 80 90
TVA (mgHAc.l -1)
I
500
0
Operation cycles
Fig. 2. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 2 gCOD L1 d1 (assay 1) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV).
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935 I
II
III
IV
I
II
III
IV
1250
150
400
80
1000
125
300
60
200
40
100
20
0
BA (mgCaCO3.l -1)
100
ε (%)
Cs (mgCOD.l -1)
500
10
20 30 40 50 60 Operation cycles
70
75 500
50
250
25
0
0 0
100
750
80
TVA (mgHAc.l -1)
932
0
10 20 30 40 50 60 70 80 Operation cycles
0
Fig. 3. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 4 gCOD L1 d1 (assay 2) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV). III
IV
I 100
60 1000 40 500
20
0 0
20
BA (mgCaCO3.l -1)
80
1500
II
III
IV
3000
ε (%)
Cs (mgCOD.l -1)
II
500 375
2000
250 1000 125 0
0 40 60 80 100 120 140 Operation cycles
0
20
40
60
TVA (mgHAc.l -1)
I
2000
0 80 100 120 140
Operation cycles
Fig. 4. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’) of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 8 gCOD L1 d1 (assay 3) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV). IV
4000
80
3000
60
2000
40
1000
20
0 0
10 20 30 40 50 60 70 80 90 Operation cycles
I
100
II
III
IV
4000
0
1000 800
3000
600 2000 400 1000 0
TVA (mgHAc.l -1)
III
BA (mgCaCO3.l -1)
Cs (mgCOD.l -1)
II
ε (%)
I
5000
200 0 0 10 20 30 40 50 60 70 80 90 Operation cycles
Fig. 5. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 12 gCOD L1 d1 (assay 4) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fedbatch (phase IV).
adequate consumption of the acids formed. Moreover, the system managed to generate a certain amount of BA beyond that consumed, presenting values similar to those of the influent.
An increase in influent organic matter concentration at a VOL of 12 gCOD L1 d1 (assay 4—see Table 4 and Fig. 5) resulted in a reduction in conversion efficiency. The TVA concentration in the effluent was high; however, no
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
tendency to accumulate was observed, i.e., values were erratic, being again a result of insufficient time for consumption of the generated acids. Analyzing the behavior of the process variables monitored in assays 1 and 2 (at VOLs of 2 and 4 gCOD L1 d1 Þ indicated that an increase in feed time did not significantly alter reactor performance. For the process variables monitored in assays 3 and 4 (at VOLs of 8 and 12 gCOD L1 d1 Þ system stability tends to increase with increasing feed time, as can be seen by the increase in both BA generation and organic matter in filtered and nonfiltered samples. High concentrations of acetic and propionic acid were found at VOLs of 2 and 4 gCOD L1 d1 (assays 1 and 2). At conditions of 8 and 12 gCOD L1 d1 (assays 3 and 4) additionally butyric and valeric acid were detected. Assuming TVA as the sum of all acids encountered along the cycle (see Figs. 6 and 7), no significant variation was seen in terms of their maximum values with increasing fill times. The difference occurred in the time at which the maximum value was attained. This time was longer as fill time increased. Moreover, acid accumulation increased with increasing organic load. The effect of fill time (i.e., fed-batch operation time) on the behavior of the volatile acids profile, considering the maximum value of volatile acids and the time when this
maximum occurred, shows that the maximum value remained approximately the same (independently of fill time) and the time at which this occurred was shifted (approximately proportional to the increase in fill time). This fact may be explained by a possible need for a minimum concentration of volatile acids for the consumption rate of these acids (for instance by methanogenesis) to attain values sufficient to promote their reduction. This behavior contradicts the expectation that a reduction in volatile acids formation rate brought about by longer fill time (i.e., by the reduced availability of organic matter) would promote a reduction in the maximum value attained as the methanogenesis step facilitates consumption of these acids. However, the propionic acid presented a behavior related to the feed strategy used (see Figs. 8 and 9). Analysis of the profile along the cycle shows that the longer the fill time, the lower the maximum concentration of this acid. Hence, the longer fill time does not seem to favor formation of propionic acid or favor its conversion to acetic acid. All assays showed a big difference between organic matter concentration (COD) in the filtered and non-filtered samples, affected by the formation of polymer-like material which was removed during discharge. This also explains the higher concentration of VSS and TSS determined in the
b
150
TVAC (mg.l -1)
TVAC (mg.l -1)
a
100
50
0 0
120
240 360 Time (minutes)
933
300
200
100
0
480
0
120
240
360
480
Time (minutes)
Fig. 6. Profiles of total volatile acids concentration by chromatography at VOL of 2 gCOD L1 d1 (a—assay 1) and of 4 gCOD L1 d1 (b—assay 2) for batch mode ð—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
b
1000 750
TVAC (mg.l -1)
TVAC (mg.l -1)
a
500 250 0
1200
800
400
0 0
120
240 360 Time (minutes)
480
0
120
240
360
480
Time (minutes)
Fig. 7. Profiles of total volatile acids concentration by chromatography at VOL of 8 gCOD L1 d1 (a—assay 3) and of 12 gCOD L1 d1 (b—assay 4) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
934
100 Concentration (mg.l -1)
Concentration (mg.l -1)
75
50
25
0
0
120
240 Time (minutes)
360
75 50 25 0
480
0
120
240 360 Time (minutes)
480
Fig. 8. Profiles of propionic acid concentration at VOL of 2 gCOD L1 d1 (a—assay 1) and of 4 gCOD L1 d1 (b—assay 2) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
b
250
Concentration (mg.l -1)
Concentration (mg.l -1)
a
200 150 100 50
500 400 300 200 100 0
0 0
120
240
360
480
Time (minutes)
0
120
240 360 Time (minutes)
480
Fig. 9. Profiles of propionic acid concentration at VOL of 8 gCOD L1 d1 (a—assay 3) and of 12 gCOD L1 d1 (b—assay 4) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
effluent in all operation conditions. Another important fact is the significant reduction observed in polymer formation with increasing fill time. Microbiological analyses showed that the foam succeeded to immobilize the anaerobic microorganisms, evidenced by the large presence of Methanosaeta-like structures and methanogenic hydrogenotrophic-like fluorescent bacilli. No Methanosarcina-like structures were observed in the samples. Considering the average values obtained in all assays, after immobilization the bioparticles presented 1:4 0:1 gTVS =gfoam and 1:5 0:1 gTS =gfoam , with TVS in the reactor of approximately 69 g. 4. Conclusions At all investigated VOLs and feed times the ASBR containing immobilized biomass presented high organic matter conversions, indicating that this technology may be applied to the treatment of cheese whey at different operation conditions. It is worth to mention that the treatment of diluted whey may be the same as the treatment of raw whey considering as a design parameter the VOL and maintaining cycle length, volume treated per cycle and total volume of the medium in the bioreactor. However, since we deal with a biological process in which
intermediate metabolites may play an important role, an investigation should be undertaken with an appropriate experimental protocol to verify this behavior. Considering the organic matter removal efficiency behavior, for VOLs of 2 and 4 gCOD L1 d1 the 2-h feed time yielded improved conversion efficiency as well as better operation stability. For VOLs of 8 and 12 gCOD L1 d1 this behavior was observed at 4-h feed time. Furthermore, despite higher TVA concentrations found at the highest VOLs, there was no tendency toward acid accumulation, indicating that this behavior was due to cycle time which was not sufficient for complete volatile acids consumption. The profiles showed that in terms of TVA concentration, there was no significant difference between the maximum values at all conditions investigated. Yet, with increasing fill time these maximum values tended to occur at times near the end of the cycle. Moreover, a reduction was observed in the maximum values of propionic acid during a cycle for longer fill times, considering each applied organic load separately. The 4-h fed-batch operation followed by 4-h batch yielded lower maximum propionic acid concentrations, indicating that this technology may also be used for substrates that produce elevated concentrations of this acid and impede anaerobic treatment.
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
Acknowledgments This study was supported by the Fundac- a˜o de Amparo a` Pesquisa do Estado de Sa˜o Paulo–FAPESP (Sa˜o Paulo, Brazil), process number 02/06.154-5. The authors gratefully acknowledge Dr. Baltus C. Bonse for the revision of this paper. References Angenent, L.T., Dague, R.R., 1995. A laboratory-scale comparison of the UASB and ASBR processes. In: 50th Purdue Industrial Wastewater Conference Proceedings. Ann Arbor Press, Chelsea. Backus, B.D., Clanton, C.J., Goodrich, P.R., Morris, H.A., 1988. Carbonto-nitrogen ratio and hydraulic retention time effect on the anaerobic digestion of cheese whey. Transactions of the ASAE 31, 1274–1282. Bagley, D.M., Brodkorb, T.S., 1999. Modeling microbial kinetics in an anaerobic sequencing batch reactor—model development and experimental validation. Water Environmental Research 71, 1320–1332. Kennedy, K.J., Sanchez, W.A., Hamoda, M.F., Droste, R.L., 1991. Performance of anaerobic sludge blanket sequencing batch reactors. Research Journal of Water Pollution Control Federation. 63, 75–83. Lo, K.V., Liao, P.H., 1986. Digestion of cheese whey with anaerobic rotating biological contact reactor. Biomass 10, 243–252. Malaspina, F., Cellamare, C.M., Stante, L., Tilche, A., 1996. Anaerobic treatment of cheese whey with a downflow–upflow hybrid reactor. Bioresource Technology 55, 131–139. Ratusznei, S.M., Rodrigues, J.A.D., Camargo, E.F.M., Ribeiro, R., Zaiat, M., 2003a. Effects of feeding strategy on a stirred anaerobic sequential fed-batch reactor containing immobilized biomass. Bioresource Technology 90, 199–205.
935
Ratusznei, S.M., Rodrigues, J.A.D., Zaiat, M., 2003b. Operating feasibility of anaerobic whey treatment in a stirred sequencing batch reactor containing immobilized biomass. Water Science and Technology 48, 179–186. Rodrigues, J.A.D., Ratusznei, S.M., Zaiat, M., 2003. Fed-Batch and batch operating mode analysis of a stirred anaerobic sequencing reactor with self-immobilized biomass treating lowstrength wastewater. Journal of Environmental Management 69, 193–200. Shizas, I., Bagley, D.M., 2002. Improving anaerobic sequencing batch reactor performance by modifying operational parameters. Water Research 36, 363–367. Standard Methods for the Examination of Water and Wastewater, 1995. 19th ed. American Public Health Association/American Water Works Association/Water Environment Federation, Washington, DC, USA. Suthaker, S., Polprasert, C., Droste, R.L., 1991. Sequencing batch anaerobic reactors for treatment of a high-strength organic wastewater. Water Science and Technology 23, 1249–1257. Wildenauer, F.X., Winter, J., 1985. Anaerobic digestion of high strength acidic whey in a ph-controlled up-flow fixed-film loop reactor. Applied Biochemistry and Biotechnology. 22, 367–372. Yan, J.Q., Liao, P.H., Lo, K.V., 1988. Methane production from cheese whey. Biomass 17, 185–202. Zaiat, M., Cabral, A.K.A., Foresti, E., 1994. Horizontal-flow anaerobic immobilized sludge reactor for wastewater treatment: conception and performance evaluation. Brazilian Journal of Chemical Engineering 11, 33–42. Zaiat, M., Rodrigues, J.A.D., Ratusznei, S.M., Camargo, E.F.M., Borzani, W., 2001. Anaerobic sequencing batch reactors for wastewater treatment: a developing technology. Applied Microbiology and Biotechnology 55, 29–35.
Journal of Environmental Management 85 (2007) 927–935 www.elsevier.com/locate/jenvman
Effects of feeding time and organic loading in an anaerobic sequencing batch biofilm reactor (ASBBR) treating diluted whey Leonardo H.S. Damascenoa, Jose´ A.D. Rodriguesb,, Suzana M. Ratuszneib, Marcelo Zaiata, Eugeˆnio Forestia a
Departamento de Hidra´ulica e Saneamento, Escola de Engenharia de Sa˜o Carlos, Universidade de Sa˜o Paulo (USP), Sa˜o Carlos-SP, Brazil Departamento de Engenharia Quı´mica e de Alimentos, Escola de Engenharia Maua´, Instituto Maua´ de Tecnologia, Caetano do Sul-SP, Brazil
b
Received 28 June 2005; received in revised form 23 October 2006; accepted 1 November 2006 Available online 20 December 2006
Abstract An investigation was carried out on the performance of an anaerobic sequencing batch biofilm reactor (ASBBR) treating diluted cheese whey when submitted to different feed strategies and volumetric organic loads (VOL). Polyurethane foam cubes were used as support for biomass immobilization and stirring was provided by helix impellers. The reactor with a working volume of 3 L treated 2 L of wastewater in 8-h cycles at 500 rpm and 30 C. The organic loads applied were 2, 4, 8 and 12 gCOD L1 d1 , obtained by increasing the feed concentration. Alkalinity was supplemented at a ratio of 50% NaHCO3 /COD. For each organic load applied three feed strategies were tested: (a) batch operation with 8-h cycle; (b) 2-h fed-batch operation followed by 6-h batch; and (c) 4-h fed-batch followed by 4-h batch. The 2-h fed-batch operation followed by 6-h batch presented the best results for the organic loads of 2 and 4 gCOD L1 d1 , whereas the 4-h fed-batch operation followed by 4-h batch presented results slightly inferior for the same organic loads and the best results at organic loads of 8 and 12 gCOD L1 d1 . The concentration of total volatile acids varied with fill time. For the higher fill times maximum concentrations were obtained at the end of the cycle. Moreover, no significant difference was detected in the maximum concentration of total volatile acids for any of the investigated conditions. However, the maximum values of propionic acid tended to decrease with increasing fill time considering the same organic load. Microbiological analyses revealed the presence of Methanosaeta-like structures and methanogenic hydrogenotrophic-like fluorescent bacilli. No Methanosarcina-like structures were observed in the samples. r 2006 Elsevier Ltd. All rights reserved. Keywords: ASBBR; Feeding time; Organic loading; Whey
1. Introduction Cheese whey possesses a high organic load with concentrations of approximately 60–80 gCOD L1 . This characteristic, combined with others such as low alkalinity and high biodegradability, make anaerobic treatment in high rate reactors difficult (Malaspina et al., 1996; Yan et al., 1988). Alkalinity present in the system is quickly consumed due to rapid conversion of lactose into shortchain volatile acids, making it necessary to constantly monitor alkalinity as well as pH (Backus et al., 1988). Supplemental addition of alkalinity as bicarbonate, carboCorresponding author. Tel.: +55 11 4239 3148; fax: +55 11 4239 3131.
E-mail address: [email protected] (J.A.D. Rodrigues). 0301-4797/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvman.2006.11.001
nate or hydroxide may also be necessary (Lo and Liao, 1986; Wildenauer and Winter, 1985). Factors that may significantly affect ASBR performance according to Zaiat et al. (2001) include feed strategy, agitation, reactor configuration and initial ratio between substrate and biomass concentrations (F/M). In batch or fed-batch operated reactors the influence of feed strategy is related to the F/M ratio. According to Angenent and Dague (1995), increase in feed time would cause a reduction in total volatile acids (TVA) and consequently increase reactor performance, since by supplying substrate within a longer time, availability to microorganisms will be less, occasioning less accumulation of these acids. Bagley and Brodkorb (1999) corroborated the assumptions of Angenent and Dague (1995) in easily
ARTICLE IN PRESS 928
L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
Nomenclature Symbols BA bicarbonate alkalinity, mgCaCO3 L1 CI influent substrate concentration, mgCOD L1 CFS filtered substrate concentration in the effluent, mgCOD L1 C TS non-filtered substrate concentration in the effluent, mgCOD L1 tF =tC ratios between feed time and cycle time, dimensionless TVA total volatile acids concentration determined by titrimetric method, mgHAc L1 TVAC total volatile acids concentration determined by chromatography method, mg L1 V Cycle feed volume per cycle, L VOL volumetric organic load, gCOD L1 d1
degraded wastewaters, suggesting, however, investigations in complex substrates. Kennedy et al. (1991) investigated the effect of feed time on the performance of an ASBR treating synthetic sucrose substrate with organic rates ranging from 2.5–18:5 gCOD L1 d1 and concentration of 7 gCOD L1 , varying filland-react periods (fill-to-react ratios of 0.2, 0.5 and 2). The process was not affected by volumetric organic loads (VOL) below 9 gCOD L1 d1 and for higher loads removal efficiency was reduced more than 25% for low ratios between feed time and cycle time ðtF =tC Þ. Suthaker et al. (1991) submitted an ASBR treating glucose based wastewater with a concentration of 35 gCOD L1 to different operation conditions (temperature from 25 to 34 C and tF =tC ratio from 0 to 0.75), with a VOL of 1:6 gCOD L1 d1 . The maximum conversion achieved was 73% for filtered samples for a 16-day cycle time and 4-day feed time. They concluded that feed time has a stronger influence than tF =tC ratio, being of fundamental importance in defining feed strategy. Shizas and Bagley (2002) also investigated an ASBR treating glucose based wastewater. The system operating with a VOL of 2:1 gCOD L1 d1 suffered overload at 3:2 gCOD L1 d1 . Reactor performance increased at high tF =tC and low initial substrate concentrations, indicating that operation can be optimized by altering these operation parameters. Rodrigues et al. (2003), operating an ASBR with granulated biomass treating low-strength wastewater ð0:5 gCOD L1 Þ, concluded that long feed times (tF =tC greater than 0.5) affects system performance related to organic matter removal efficiency, settleability characteristics and extra-cellular polymer synthesis. Ratusznei et al. (2003a), operating the ASBR with biomass immobilized on polyurethane foam and treating the same wastewater, observed a drop in efficiency at long feed times. This behavior was attributed to the time the immobilized mass
V Working working volume, L Greek characters eFS substrate removal efficiency considering filtered samples, % eTS substrate removal efficiency considering nonfiltered samples, % Abbreviations ASBR anaerobic sequencing batch reactor ASBBR anaerobic sequencing batch biofilm reactor COD chemical oxygen demand F/M ratio between substrate and biomass concentrations TSS total suspended solids VSS volatile suspended solids TS total solids TVS total volatile solids UASB upflow anaerobic sludge blanket
was exposed to air, since during feeding the bed was gradually immersed, occasioning periods at which the biomass had no contact with the wastewater; at short feed times the bed was rapidly immersed in the substrate, which did not occur at longer feed times. The main objective of this investigation was to assess the behavior of an ASBBR containing microorganisms immobilized on polyurethane foam treating diluted cheese whey when submitted to different VOL (2, 4, 8 and 12 gCOD L1 d1 Þ and feed strategies (fill times of 10 min, 2 h and 4 h by a cycle time of 8 h). 2. Materials and methods The reactor used was made of acrylic with internal diameter and height of 200 mm. A 100-mm high basket was used in the reactor to contain the biomass, as shown in the scheme in Fig. 1. This basket occupied half of the reactor volume to allow permanent immersion of the bed during fed-batch operation. The system operated in a chamber which maintained the temperature at 30 1 C. Stirring at 500 rpm was provided by two helix propellers. The helix propellers with diameter of 60 mm consisted of three blades; one positioned at 20 mm from the reactor bottom and the other above the bed at 120 mm from the reactor bottom. The inert support used consisted of 5-mm polyurethane foam cubes and was inoculated, according to methodology proposed by Zaiat et al. (1994), i.e., (i) the biomass was squeezed through a 1 mm mesh sieve as a way to crush the granules present and facilitate attachment of the biomass in the foam; (ii) the 5-mm foam cubes were added to the crushed inoculum and mixed in a way to saturate the foam completely with the sludge; and (iii) after 2 h the foam cubes containing the immobilized biomass were added to the reactor, taking care to avoid compression of the cubes. The sludge used was from a UASB reactor treating
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
929
Fig. 1. Scheme of the anaerobic sequencing batch biofilm reactor (ASBBR) containing immobilized biomass [(1) bioreactor, (2) stainless steel basket containing particles with immobilized cells, (3) impeller, (4) feed pump, (5) discharge pump, (6) substrate, (7) treated effluent, (8) timers].
wastewater from a poultry slaughterhouse that had total volatile solids (TVS) and total solids (TS) of 51 and 62 g L1 , respectively. The wastewater used consisted of reconstituted dehydrated cheese whey for which the ratio between dry whey and amount of COD was experimentally determined as 1:1. Alkalinity supplementation was performed by adding sodium bicarbonate, utilizing g-NaHCO3 /g-COD ratios proposed by Ratusznei et al. (2003b), i.e., to guarantee that bicarbonate alkalinity (BA) was not a limiting factor. This way, at the start of each phase a g-NaHCO3 /g-COD ratio of 100% was used and after experimental verification of stability this ratio was reduced to 50%. The following organic matter concentrations were used: 1 gCOD L1 (assay 1), 2 gCOD L1 (assay 2), 4 gCOD L1 (assay 3) and 6 gCOD L1 (assay 4). Bioreactor operation consisted of initially feeding 3 L wastewater to be treated in 8-h cycles. In each cycle 2 L wastewater were discharged; the remaining 1 L residual volume in the reactor was used to dilute the following 2 L wastewater fed in the next cycle. Feeding and discharge were performed by means of diaphragm pumps; discharge was done in 10 min. The study of feed influence was performed in the same bioreactor in three different periods, which defined the feeding strategies investigated. Each assay using a different feed strategy was performed with the same biomass without an acclimation period. (i) 8-h batch operation: Feeding of 2 L wastewater to be treated in 10 min and a g-NaHCO3 /g-COD ratio of 100% for initial alkalinity supplementation (phase I) and subsequently of 50% (phase II); (ii) 2-h fed-batch operation followed by 6-h batch: Feeding of 2 L wastewater to be treated in 2 h and a g-NaHCO3 / g-COD ratio of 50% for initial alkalinity supplementation (phase III);
(iii) 4-h fed-batch followed by 4-h batch: Feeding of 2 L wastewater to be treated in 4 h and a g-NaHCO3 / g-COD ratio of 50% for initial alkalinity supplementation (phase IV). This way, the VOL were approximately 2 gCOD L1 d1 (assay 1), 4 gCOD L1 d1 (assay 2), 8 gCOD L1 d1 (assay 3) and 12 gCOD L1 d1 (assay 4), determined by VOL ¼
V Cycle N Cycles=Day C I . V Useful
(1)
Monitoring was performed according to Standard Methods for the Examination of Water and Wastewater (1995). The following parameters were analyzed: organic matter concentration (COD) for filtered ðC FS Þ and non-filtered samples ðC TS Þ, pH, BA, TVA, TS, TVS, total suspended solids (TSS) and volatile suspended solids (VSS). Moreover, intermediate volatile acids and biogas composition were analyzed by gas phase chromatography. Composition of the biogas generated by anaerobic degradation was analyzed by gas chromatography using a Hewlett Packards 6890 gas chromatograph equipped with a thermal conductivity detector. Sample volume was 1 mL, drag gas was hydrogen at a flow rate of 50:0 mL h1 , and column, injector and detector temperatures were 35, 60 and 160 C, respectively. Volatile fatty acid (VFA) samples were analyzed by gas chromatography, using a gas chromatograph HP6890 with flame ionization detector at 300 C and an HP-INNOWAX column ð30 m 0:25 mm 0:25 mmÞ. The injector temperature was kept at 250 C; the oven was held at 100 C for 3 min, after which it was heated at a rate of 5 C min1 to 180 C, and held at that temperature for 5 min. After attainment of operation stability in each assay and in each implemented feed strategy, profiles along cycles were taken of COD concentrations for filtered samples, C S , pH, BA, TVA, intermediate volatile acids and biogas
ARTICLE IN PRESS 930
L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
composition. Microbiological identification was performed by means of common optical and phase contrast microscopy employing an Olympus BH2 microscope. 3. Results and discussion Average values of the monitored parameters are listed in Tables 1–4. Figs. 2 and 5 show the variation in organic matter concentration in the filtered and non-filtered samples, conversion efficiency, BA and TVA concentration in effluent samples for VOLs of 2–12 gCOD L1 d1 . For a VOL of 2 gCOD L1 d1 (assay 1—see Table 1 and Fig. 2) high efficiency was obtained in all phases, presenting low values of TVA concentration and increase in BA values of the effluent in relation to the influent. The operation with 2-h fed-batch followed by 6-h batch (phase III) presented improved operation stability when compared to the batch operation.
At the condition with VOL of 4 gCOD L1 d1 (assay 2—see Table 2 and Fig. 3) organic matter conversions in all phases were high, with improved performance for the 2-h fed-batch followed by 6-h batch (phase III). The system was stable presenting BA generation, reduced values of TVA and pH close to 7. Extra-cellular polymer formation, which affected the system performance, was also observed. This material was removed in the 42nd cycle (see Fig. 3). During the assay with VOL of 8 gCOD L1 d1 (assay 3—see Table 3 and Fig. 4) the system was less stable than in other conditions, but considering the ‘‘organic matter efficiency removal’’ profiles, the system may be considered slightly stable in all phases, despite a decrease in system efficiency and increase in TVA, both presenting a tendency to improve with longer fill times (phase IV). The high TVA concentrations did not occasion reactor instability, as no accumulation of these acids occurred. This fact was a consequence of the cycle time used, which did not allow
Table 1 Average values of the monitored variables in the assay with VOL of 2 gCOD L1 d1 (assay 1) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS (mgCOD L1 Þ eTS (%) eFS (%)
976 32ð17Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
44 6ð9Þ 280 21ð9Þ 8:3 0:2ð9Þ 1272 76ð9Þ 974 69ð9Þ 45 11ð9Þ 40 9ð9Þ –
177 65ð10Þ 78 29ð10Þ 82 7ð10Þ 92 3ð10Þ 27 8ð8Þ 334 14ð6Þ 7:2 0:4ð6Þ 761 54ð3Þ 430 70ð3Þ 77 29ð3Þ 65 36ð3Þ 71:2ð1Þ
150 28ð5Þ 41 8ð5Þ 85 3ð5Þ 96 1ð5Þ 30 6ð5Þ 341 42ð5Þ 6:8 0:2ð5Þ 693 31ð3Þ 372 55ð3Þ 95 36ð3Þ 91 34ð3Þ 72:9ð1Þ
192 5ð6Þ 85 15ð6Þ 80 1ð6Þ 91 1ð6Þ 59 21ð6Þ 295 14ð6Þ 6:8 0:2ð6Þ 622 103ð3Þ 313 69ð3Þ 83 8ð3Þ 75 17ð3Þ 71:2ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
Table 2 Average values of the monitored variables in the assay with VOL of 4 gCOD L1 d1 (assay 2) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS ðmgCOD L1 Þ eTS (%) eFS (%)
2005 64ð14Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
83 13ð10Þ 553 13ð10Þ 8:4 0:2ð10Þ 2345 369ð8Þ 1726 363ð8Þ 78 31ð8Þ 72 30ð8Þ –
265 67ð8Þ 115 45ð8Þ 87 3ð8Þ 94 2ð8Þ 56 26ð9Þ 653 23ð5Þ 7:6 0:1ð5Þ 1249 163ð4Þ 476 158ð4Þ 71 35ð4Þ 65 29ð4Þ 67:8ð1Þ
104 36ð3Þ 59 18ð3Þ 95 2ð3Þ 97 1ð3Þ 46 20ð4Þ 672 12ð4Þ 6:8 0:1ð4Þ 1117 61ð3Þ 502 48ð3Þ 98 11ð3Þ 93 10ð3Þ 71:8ð1Þ
266 68ð5Þ 147 36ð5Þ 87 3ð5Þ 93 2ð5Þ 90 13ð6Þ 626 12ð6Þ 6:8 0:1ð6Þ 1116 8ð2Þ 433 64ð2Þ 101 7ð2Þ 83 10ð2Þ 71ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
931
Table 3 Average values of the monitored variables in the assay with VOL of 8 gCOD L1 d1 (assay 3) for batch mode (phase II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ CFS ðmgCOD L1 Þ eTS (%) eFS (%)
4020 108ð26Þ – – –
TVA ðmgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
134 10ð12Þ 1103 40ð12Þ 8:4 0:3ð12Þ 4858 216ð11Þ 3430 182ð11Þ 175 11ð11Þ 153 15ð11Þ –
879 136ð12Þ 650 135ð12Þ 78 3ð12Þ 84 3ð12Þ 321 80ð14Þ 1103 121ð10Þ 6:8 0:1ð10Þ 2681 563ð5Þ 1018 112ð5Þ 187 49ð5Þ 166 36ð5Þ 63:7ð1Þ
963 144ð6Þ 615 142ð6Þ 76 4ð6Þ 85 4ð6Þ 301 67ð7Þ 1084 56ð7Þ 6:9 0:2ð7Þ 2402 226ð2Þ 928 209ð2Þ 226 71ð2Þ 195 75ð2Þ 65:5ð1Þ
904 278ð9Þ 566 285ð9Þ 78 7ð9Þ 86 7ð9Þ 285 97ð9Þ 1096 92ð9Þ 6:9 0:1ð9Þ 2282 334ð4Þ 978 85ð4Þ 252 8ð4Þ 226 7ð4Þ 67:8ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
Table 4 Average values of the monitored variables in the assay with VOL of 12 gCOD L1 d1 (assay 4) for batch mode (phase II), 2-h fed-batch (phase III) and 4h fed-batch (phase IV) Variable
Influent
Effluent batch
Effluent 2-h fed-batch
Effluent 4-h fed-batch
C TS (mgCOD L1 Þ C FS (mgCOD L1 Þ eTS (%) eFS (%)
5969 165ð19Þ – – –
TVA (mgHAc L1 Þ BA ðmgCaCO3 L1 Þ pH
204 28ð9Þ 1689 77ð9Þ 8:4 0:3ð9Þ 7132 154ð8Þ 4960 164ð8Þ 302 71ð8Þ 266 73ð8Þ –
1782 399ð8Þ 1215 444ð8Þ 70 7ð8Þ 78 8ð8Þ 527 181ð7Þ 1582 173ð7Þ 7:1 0:1ð7Þ 3537 352ð4Þ 1431 381ð4Þ 295 55ð4Þ 248 50ð4Þ 66:4ð1Þ
2055 113ð5Þ 1626 141ð5Þ 66 2ð5Þ 73 2ð5Þ 750 82ð6Þ 1429 72ð6Þ 7:1 0:2ð6Þ 3800 57ð2Þ 1602 34ð2Þ 318 37ð2Þ 276 8ð2Þ 63:6ð1Þ
1687 237ð5Þ 1118 181ð5Þ 72 4ð5Þ 81 3ð5Þ 602 93ð6Þ 1571 49ð6Þ 7:1 0:1ð6Þ 3502 238ð2Þ 1318 141ð2Þ 442 21ð2Þ 403 19ð2Þ 63:9ð1Þ
TS ðmg L1 Þ TVS ðmg L1 Þ TSS ðmg L1 Þ VSS ðmg L1 Þ CH4 (%)
Note: Values between brackets refer to the number of samples averaged.
III
IV
I
II
III
IV
700
100
400
80
560
80
300
60
420
60
200
40
280
40
100
20
140
20
0 0
10 20 30 40 50 60 70 80 90 Operation cycles
0
BA (mgCaCO3.l -1)
100
ε (%)
Cs (mgCOD.l -1)
II
0 0
10 20 30 40 50 60 70 80 90
TVA (mgHAc.l -1)
I
500
0
Operation cycles
Fig. 2. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 2 gCOD L1 d1 (assay 1) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV).
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935 I
II
III
IV
I
II
III
IV
1250
150
400
80
1000
125
300
60
200
40
100
20
0
BA (mgCaCO3.l -1)
100
ε (%)
Cs (mgCOD.l -1)
500
10
20 30 40 50 60 Operation cycles
70
75 500
50
250
25
0
0 0
100
750
80
TVA (mgHAc.l -1)
932
0
10 20 30 40 50 60 70 80 Operation cycles
0
Fig. 3. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 4 gCOD L1 d1 (assay 2) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV). III
IV
I 100
60 1000 40 500
20
0 0
20
BA (mgCaCO3.l -1)
80
1500
II
III
IV
3000
ε (%)
Cs (mgCOD.l -1)
II
500 375
2000
250 1000 125 0
0 40 60 80 100 120 140 Operation cycles
0
20
40
60
TVA (mgHAc.l -1)
I
2000
0 80 100 120 140
Operation cycles
Fig. 4. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’) of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 8 gCOD L1 d1 (assay 3) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fed-batch (phase IV). IV
4000
80
3000
60
2000
40
1000
20
0 0
10 20 30 40 50 60 70 80 90 Operation cycles
I
100
II
III
IV
4000
0
1000 800
3000
600 2000 400 1000 0
TVA (mgHAc.l -1)
III
BA (mgCaCO3.l -1)
Cs (mgCOD.l -1)
II
ε (%)
I
5000
200 0 0 10 20 30 40 50 60 70 80 90 Operation cycles
Fig. 5. Substrate concentrations in the effluent (—C FS , —C TS Þ and conversion efficiency (eFS —&, eTS —’), of bicarbonate alkalinity ðmÞ and of total volatile acids ðnÞ in the reactor operated at VOL of 12 gCOD L1 d1 (assay 4) for batch mode (phases I and II), 2-h fed-batch (phase III) and 4-h fedbatch (phase IV).
adequate consumption of the acids formed. Moreover, the system managed to generate a certain amount of BA beyond that consumed, presenting values similar to those of the influent.
An increase in influent organic matter concentration at a VOL of 12 gCOD L1 d1 (assay 4—see Table 4 and Fig. 5) resulted in a reduction in conversion efficiency. The TVA concentration in the effluent was high; however, no
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
tendency to accumulate was observed, i.e., values were erratic, being again a result of insufficient time for consumption of the generated acids. Analyzing the behavior of the process variables monitored in assays 1 and 2 (at VOLs of 2 and 4 gCOD L1 d1 Þ indicated that an increase in feed time did not significantly alter reactor performance. For the process variables monitored in assays 3 and 4 (at VOLs of 8 and 12 gCOD L1 d1 Þ system stability tends to increase with increasing feed time, as can be seen by the increase in both BA generation and organic matter in filtered and nonfiltered samples. High concentrations of acetic and propionic acid were found at VOLs of 2 and 4 gCOD L1 d1 (assays 1 and 2). At conditions of 8 and 12 gCOD L1 d1 (assays 3 and 4) additionally butyric and valeric acid were detected. Assuming TVA as the sum of all acids encountered along the cycle (see Figs. 6 and 7), no significant variation was seen in terms of their maximum values with increasing fill times. The difference occurred in the time at which the maximum value was attained. This time was longer as fill time increased. Moreover, acid accumulation increased with increasing organic load. The effect of fill time (i.e., fed-batch operation time) on the behavior of the volatile acids profile, considering the maximum value of volatile acids and the time when this
maximum occurred, shows that the maximum value remained approximately the same (independently of fill time) and the time at which this occurred was shifted (approximately proportional to the increase in fill time). This fact may be explained by a possible need for a minimum concentration of volatile acids for the consumption rate of these acids (for instance by methanogenesis) to attain values sufficient to promote their reduction. This behavior contradicts the expectation that a reduction in volatile acids formation rate brought about by longer fill time (i.e., by the reduced availability of organic matter) would promote a reduction in the maximum value attained as the methanogenesis step facilitates consumption of these acids. However, the propionic acid presented a behavior related to the feed strategy used (see Figs. 8 and 9). Analysis of the profile along the cycle shows that the longer the fill time, the lower the maximum concentration of this acid. Hence, the longer fill time does not seem to favor formation of propionic acid or favor its conversion to acetic acid. All assays showed a big difference between organic matter concentration (COD) in the filtered and non-filtered samples, affected by the formation of polymer-like material which was removed during discharge. This also explains the higher concentration of VSS and TSS determined in the
b
150
TVAC (mg.l -1)
TVAC (mg.l -1)
a
100
50
0 0
120
240 360 Time (minutes)
933
300
200
100
0
480
0
120
240
360
480
Time (minutes)
Fig. 6. Profiles of total volatile acids concentration by chromatography at VOL of 2 gCOD L1 d1 (a—assay 1) and of 4 gCOD L1 d1 (b—assay 2) for batch mode ð—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
b
1000 750
TVAC (mg.l -1)
TVAC (mg.l -1)
a
500 250 0
1200
800
400
0 0
120
240 360 Time (minutes)
480
0
120
240
360
480
Time (minutes)
Fig. 7. Profiles of total volatile acids concentration by chromatography at VOL of 8 gCOD L1 d1 (a—assay 3) and of 12 gCOD L1 d1 (b—assay 4) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
934
100 Concentration (mg.l -1)
Concentration (mg.l -1)
75
50
25
0
0
120
240 Time (minutes)
360
75 50 25 0
480
0
120
240 360 Time (minutes)
480
Fig. 8. Profiles of propionic acid concentration at VOL of 2 gCOD L1 d1 (a—assay 1) and of 4 gCOD L1 d1 (b—assay 2) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
b
250
Concentration (mg.l -1)
Concentration (mg.l -1)
a
200 150 100 50
500 400 300 200 100 0
0 0
120
240
360
480
Time (minutes)
0
120
240 360 Time (minutes)
480
Fig. 9. Profiles of propionic acid concentration at VOL of 8 gCOD L1 d1 (a—assay 3) and of 12 gCOD L1 d1 (b—assay 4) for batch mode (—phase II), 2-h fed-batch (’—phase III) and 4-h fed-batch (—phase IV).
effluent in all operation conditions. Another important fact is the significant reduction observed in polymer formation with increasing fill time. Microbiological analyses showed that the foam succeeded to immobilize the anaerobic microorganisms, evidenced by the large presence of Methanosaeta-like structures and methanogenic hydrogenotrophic-like fluorescent bacilli. No Methanosarcina-like structures were observed in the samples. Considering the average values obtained in all assays, after immobilization the bioparticles presented 1:4 0:1 gTVS =gfoam and 1:5 0:1 gTS =gfoam , with TVS in the reactor of approximately 69 g. 4. Conclusions At all investigated VOLs and feed times the ASBR containing immobilized biomass presented high organic matter conversions, indicating that this technology may be applied to the treatment of cheese whey at different operation conditions. It is worth to mention that the treatment of diluted whey may be the same as the treatment of raw whey considering as a design parameter the VOL and maintaining cycle length, volume treated per cycle and total volume of the medium in the bioreactor. However, since we deal with a biological process in which
intermediate metabolites may play an important role, an investigation should be undertaken with an appropriate experimental protocol to verify this behavior. Considering the organic matter removal efficiency behavior, for VOLs of 2 and 4 gCOD L1 d1 the 2-h feed time yielded improved conversion efficiency as well as better operation stability. For VOLs of 8 and 12 gCOD L1 d1 this behavior was observed at 4-h feed time. Furthermore, despite higher TVA concentrations found at the highest VOLs, there was no tendency toward acid accumulation, indicating that this behavior was due to cycle time which was not sufficient for complete volatile acids consumption. The profiles showed that in terms of TVA concentration, there was no significant difference between the maximum values at all conditions investigated. Yet, with increasing fill time these maximum values tended to occur at times near the end of the cycle. Moreover, a reduction was observed in the maximum values of propionic acid during a cycle for longer fill times, considering each applied organic load separately. The 4-h fed-batch operation followed by 4-h batch yielded lower maximum propionic acid concentrations, indicating that this technology may also be used for substrates that produce elevated concentrations of this acid and impede anaerobic treatment.
ARTICLE IN PRESS L.H.S. Damasceno et al. / Journal of Environmental Management 85 (2007) 927–935
Acknowledgments This study was supported by the Fundac- a˜o de Amparo a` Pesquisa do Estado de Sa˜o Paulo–FAPESP (Sa˜o Paulo, Brazil), process number 02/06.154-5. The authors gratefully acknowledge Dr. Baltus C. Bonse for the revision of this paper. References Angenent, L.T., Dague, R.R., 1995. A laboratory-scale comparison of the UASB and ASBR processes. In: 50th Purdue Industrial Wastewater Conference Proceedings. Ann Arbor Press, Chelsea. Backus, B.D., Clanton, C.J., Goodrich, P.R., Morris, H.A., 1988. Carbonto-nitrogen ratio and hydraulic retention time effect on the anaerobic digestion of cheese whey. Transactions of the ASAE 31, 1274–1282. Bagley, D.M., Brodkorb, T.S., 1999. Modeling microbial kinetics in an anaerobic sequencing batch reactor—model development and experimental validation. Water Environmental Research 71, 1320–1332. Kennedy, K.J., Sanchez, W.A., Hamoda, M.F., Droste, R.L., 1991. Performance of anaerobic sludge blanket sequencing batch reactors. Research Journal of Water Pollution Control Federation. 63, 75–83. Lo, K.V., Liao, P.H., 1986. Digestion of cheese whey with anaerobic rotating biological contact reactor. Biomass 10, 243–252. Malaspina, F., Cellamare, C.M., Stante, L., Tilche, A., 1996. Anaerobic treatment of cheese whey with a downflow–upflow hybrid reactor. Bioresource Technology 55, 131–139. Ratusznei, S.M., Rodrigues, J.A.D., Camargo, E.F.M., Ribeiro, R., Zaiat, M., 2003a. Effects of feeding strategy on a stirred anaerobic sequential fed-batch reactor containing immobilized biomass. Bioresource Technology 90, 199–205.
935
Ratusznei, S.M., Rodrigues, J.A.D., Zaiat, M., 2003b. Operating feasibility of anaerobic whey treatment in a stirred sequencing batch reactor containing immobilized biomass. Water Science and Technology 48, 179–186. Rodrigues, J.A.D., Ratusznei, S.M., Zaiat, M., 2003. Fed-Batch and batch operating mode analysis of a stirred anaerobic sequencing reactor with self-immobilized biomass treating lowstrength wastewater. Journal of Environmental Management 69, 193–200. Shizas, I., Bagley, D.M., 2002. Improving anaerobic sequencing batch reactor performance by modifying operational parameters. Water Research 36, 363–367. Standard Methods for the Examination of Water and Wastewater, 1995. 19th ed. American Public Health Association/American Water Works Association/Water Environment Federation, Washington, DC, USA. Suthaker, S., Polprasert, C., Droste, R.L., 1991. Sequencing batch anaerobic reactors for treatment of a high-strength organic wastewater. Water Science and Technology 23, 1249–1257. Wildenauer, F.X., Winter, J., 1985. Anaerobic digestion of high strength acidic whey in a ph-controlled up-flow fixed-film loop reactor. Applied Biochemistry and Biotechnology. 22, 367–372. Yan, J.Q., Liao, P.H., Lo, K.V., 1988. Methane production from cheese whey. Biomass 17, 185–202. Zaiat, M., Cabral, A.K.A., Foresti, E., 1994. Horizontal-flow anaerobic immobilized sludge reactor for wastewater treatment: conception and performance evaluation. Brazilian Journal of Chemical Engineering 11, 33–42. Zaiat, M., Rodrigues, J.A.D., Ratusznei, S.M., Camargo, E.F.M., Borzani, W., 2001. Anaerobic sequencing batch reactors for wastewater treatment: a developing technology. Applied Microbiology and Biotechnology 55, 29–35.