- Email: [email protected]
Anaerobic ponds treatment of starch wastewater: case study in Thailand
Bioresource Technology 95 (2004) 135–143
Anaerobic ponds treatment of starch wastewater: case study in Thailand B.K. Rajbhandari, A.P. Annachhatre
*
Environmental Engineering and Management, Asian Institute of Technology, P.O. Box 4, Klong Luang, Pathumthani 12120, Thailand Received in revised form 20 January 2004; accepted 26 January 2004 Available online 12 March 2004
Abstract Anaerobic ponds are particularly effective in treating high-strength wastewater containing biodegradable solids as they achieve the dual purpose of particulate settlement and organic removal. Performance of an anaerobic pond system for treatment of starch wastewater containing high organic carbon, biodegradable starch particulate matter and cyanide was assessed under tropical climate conditions. Approximately 5000 m3 /d of wastewater from starch industry was treated in a series of anaerobic ponds with a total area of 7.39 ha followed by facultative ponds with an area of 29.11 ha. Overall COD and TSS removal of over 90% and CN removal of 51% was observed. Active biomass obtained from the anaerobic ponds sediments and bulk liquid layer exhibited specific methanogenic activity of 20.7 and 11.3 ml CH4 /g VSS d, respectively. The cyanide degradability of sludge at initial cyanide concentration of 10 and 20 mg/l were determined to be 0.43 and 0.84 mg CN /g VSS d, respectively. A separate settling column experiment with starch wastewater revealed that a settling time of approximately 120 min is sufficient to remove 90–95% of the influent TSS. 2004 Elsevier Ltd. All rights reserved. Keywords: Anaerobic pond; Cyanide degradability; Organic carbon; Settling characteristics; Specific methanogenic activity; Starch factory wastewater
1. Introduction Anaerobic ponds (APs) are popularly employed for treatment of organic wastewater emanating from variety of industries such as food, pulp and paper, sugar and distillery. Anaerobic ponds are particularly effective in treating high-strength wastewaters containing biodegradable total suspended solids (TSS). In such cases the liquid layer in anaerobic ponds act as a settling basin for the suspended solids while the anaerobic biodegradation primarily takes place in pond sediments (Toprak, 1994). Anaerobic reactions taking place in the sediment include solubilization of biodegradable particulate matter followed by acidogenesis, acetogenesis and methanogenesis (Parker, 1979; Pescod, 1996). The reactions occurring in the bulk liquid are often negligible as compared to those in the pond sediments. Thus, anaerobic ponds achieve a dual purpose of sedimentation of particulate matter as well as anaerobic conversion of organics. However,
*
Corresponding author. Tel./fax: +66-2-524-5644. E-mail address: [email protected] (A.P. Annachhatre).
0960-8524/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2004.01.017
anaerobic pond operation also has many intrinsic problems such as high land requirements and emission of obnoxious and greenhouse gases such as hydrogen sulfide (H2 S), carbon dioxide (CO2 ) and methane (CH4 ) (Parker, 1979; Pescod, 1996; Toprak, 1997; Paing et al., 2003). In spite of these problems, anaerobic ponds are popular particularly wherever land is abundant (Arthur, 1983). Wastewater coming from starch factories is one such type of wastewater, which is treated extensively in anaerobic ponds. Starch is often produced in many parts of the world from tapioca. Tapioca roots contain 20– 25% starch. The starch extraction process essentially involves pre-processing of roots, followed by starch extraction, separation and drying. The process generates 20–60 m3 /ton of wastewater with a low pH in the range 3.8–5.2 (Economic and Social Commission for Asia and The Pacific, 1982). The wastewater is highly organic in nature with chemical oxygen demand (COD) up to 25,000 mg/l (Bengtsson and Treit, 1994). The wastewater consists of high TSS comprising starch granules in the range 3000–15,000 mg/l, which are highly biodegradable by nature. Tapioca starch wastewater also has high cyanide content up to 10–15 mg/l, which is highly toxic
136
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
to aquatic life at concentrations of cyanide as low as 0.3 mg/l have been reported as cause for a massive fish kill (Bengtsson and Treit, 1994). Problems related to water pollution are reported to be serious. The acidic nature of wastewater can harm aquatic organisms and reduce the self-purification capacity of the receiving stream. Suspended solids present in the wastewater can settle on the streambed and spoil fish breeding areas in the stream. Since these solids are primarily organic in nature, they decompose easily and thus deoxygenate the water. Similarly, high biochemical oxygen demand (BOD) of the wastewater also can cause rapid depletion of oxygen content in the receiving water body and promote the growth of nuisance organisms. Water pollution caused by tapioca starch production has been reported as a serious problem in many Asian countries, particularly in Thailand (Kiravanich, 1977) and in India (Padmaja et al., 1990). Tapioca also contains bound cyanide as a natural defense mechanism. During the starch manufacturing process, bound cyanide in the form of linamarin and lotaustralin from tapioca roots is hydrolyzed by the enzyme linamarase with decomposition to hydrogen cyanide (HCN), which finds its way into the wastewater. Cyanide containing starch wastewater can be effectively detoxified in anaerobic processes (Annachhatre and Amornkaew, 2000). Upflow anaerobic sludge blanket (UASB) processes are effective in treating starch wastewater (Annachhatre and Amatya, 2000), particularly, in removing cyanide (Annachhatre and Amornkaew, 2001). Adaptation by methanogens to cyanide concentrations of 5–30 mg/l has been reported in literature (Fedorak et al., 1986; Harper et al., 1983). Thus, in treating tapioca starch wastewater anaerobic ponds achieve a threefold objective namely: sedimentation of particulate matter, anaerobic conversion of organics and detoxification of cyanide. Accordingly, the work presented here assesses the performance of APs treating wastewater from tapioca starch industry, particularly related to COD, TSS and cyanide removal. Since APs serve as a settling basin for starch granules, the settling characteristics were also assessed by column experiments. Furthermore, the potential methane production rates of anaerobic biomass (sludge) obtained from the AP sediment as well as from bulk liquid layer were assessed from the specific methanogenic activity (SMA) test. The cyanide degradability of the anaerobic sludge obtained from the pond sediment layer was also assessed.
2. Methods Investigations on the existing wastewater anaerobic pond system were carried out in a tapioca starch and glucose factory situated in the Central province of
Thailand with a capacity of 250 tons starch/day. The factory uses groundwater as a source for process water, and generates combined wastewater of approximately 5000 m3 /d. The operating ambient temperature during the period of investigation was in the range of 30–35 C. 2.1. Treatment ponds A schematic of waste stabilization pond system (WSPS) of the starch factory is presented in Fig. 1 (Choi, 2001). The wastewater treatment system consists of 21 APs and facultative ponds (FPs) connected in series with total area of about 36.5 ha. Out of these, 6 are AP with area of 7.39 ha and 15 are FP with 29.11 ha. The study concentrated on anaerobic ponds system. During the study period only four anaerobic ponds were in operation. The typical size of an AP is approximately 250 m in length, 100 m in width and 4–5 m in depth. The pond parameters are presented in Table 1. The anaerobic ponds treat wastewater from a starch as well as from a glucose factory. The wastewater from the starch factory was first introduced to Pond #2 and subsequently flows to Pond #4 while the effluent from the glucose factory was introduced to Pond #3 and then flows to Pond #5 where wastewater from the starch and glucose factory were combined. The combined wastewater then flows into a series of FPs and treated effluent was finally discharged on to the groundwater recharge spreading basins. 2.2. Sludge activity tests The schematic of the SMA test set up is presented in Fig. 2. To determine SMA, a known amount of sludge obtained from the sediment layer of Pond #4 was transferred into serum bottles (115 ml) after washing three times with water to remove existing COD. While 100 ml of bulk liquid from the same pond was kept in a serum bottle to determine SMA of sludge in suspension in the bulk liquid layer. An appropriate amount of starch factory wastewater as substrate was added to the serum bottles so as to obtain an initial COD level in the range of 2000–2500 mg/l. Nutrients were added to maintain a carbon:nitrogen:phosphorus ratio of 300:5:1. pH was adjusted to between 7 and 7.8. Two grams per liter of NaHCO3 were also added along with substrate to buffer the serum bottle contents to near neutral pH conditions during the test. Subsequently the bottles were sealed with a rubber septum and an aluminum cap after purging oxygen with nitrogen (N2 ) gas and attached to the liquid displacement system. The liquid displacement bottle contained 3% NaOH solution. Methane gas production was measured at different time intervals up to 48 h. After every gas measurement, by swirling manually, the contents of the serum bottle were mixed. The tests were conducted in a 30 C temperature-
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
137
Fig. 1. Layout of WSPS of starch factory.
Table 1 Ponds parameters APs
Surface area (ha)
Volume (m3 · 104 )
Depth (m)
Inflowb (m3 /d)
Detention timeb (d)
1a 2 3 4 5 6a
1.07 1.34 2.06 1.76 0.49 0.67
4.41 5.57 8.62 7.34 1.78 2.46
4.5 4.5 4.5 4.5 4.5 4.5
4501 ± 731 498 ± 82 4501 ± 731 4999 ± 785
12.7 ± 1.9 177.8 ± 32.8 16.7 ± 2.5 3.6 ± 0.5
a b
Pond not in operation at the time of this study. Results are mean of nine values ± standard deviation.
from Pond #4 sediment layer was kept in serum bottles and filled with 70 ml of wastewater having substrate and nutrient similar to those mentioned for the SMA test. A stock cyanide solution was added into each serum bottle to achieve cyanide concentrations of 10 and 20 mg/l respectively. The bottle was then purged with N2 gas and immediately sealed with a rubber septum and an aluminum cap. The bottle was kept in a 30 C temperature-controlled room. Samples were taken with Hamilton syringe at every 8 h interval for 48 h and analyzed for cyanide content. Fig. 2. Sludge activity test setup.
2.3. Suspended solid settling experiment controlled room. Likewise, cyanide degradation activity of anaerobic pond sediment was carried out also in the serum bottles. A known amount of sludge obtained
Batch tests to investigate TSS settling characteristics of starch factory wastewater under quiescent condition
138
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
was carried out in a settling column of 10.0 cm diameter and 2.0 m height for different TSS concentrations. Wastewater of a desired TSS concentration for settling experiment was prepared by diluting the concentrated wastewater with tap water. The wastewater was poured into a settling column after stirring thoroughly. Samples from top of column were collected at different time intervals ranging from 2 to 60 min and analyzed for TSS concentration. 2.4. Analytical procedures Parameters including COD, BOD5 , TSS, volatile suspended solid (VSS) and dissolved solid (DS) were analyzed according to Standard Methods (APHA et al., 1998). The mass of the sludge used in the sludge activity test was measured in terms of VSS. All the samples were filtered through 0.45 lm glass fiber filters for the determination of soluble COD and BOD5 . Cyanide was measured spectrophotometrically (Spectroquant, E. Merck KGaA, Darmstadt, Germany) as per the procedure reported elsewhere (Annachhatre and Amornkaew, 2000). 2.5. Statistical analysis The anaerobic ponds process performance data were presented in terms of arithmetic averages of nine values ± standard deviation. The SMA tests were carried out with two replicates. Common linear regression curve was fitted to the data obtained from the two replicates test and a relation between volumes of methane production with respect to time was established. The SMA was calculated based on the slope of methane volume versus time curve and mass of sludge taken for the SMA test. Likewise a linear relation was established between the cumulative cyanide degradation with respect to time. Data from settling experiments was used to establish a non-linear relationship between the half-removal time and influent total suspended solid concentrations. All
statistical analyses (arithmetic average, standard deviation, linear and non-linear regression and correlation coefficient) were performed using Microsoft Excel 2000.
3. Results and discussion 3.1. Analysis of existing wastewater process Characteristics of raw wastewater: The pond system treats approximately 4500 m3 /d of wastewater from starch and approximately 500 m3 /d of wastewater from glucose factory. A scheme of anaerobic ponds and sampling points is shown in Fig. 3. The characterization of raw, influent and effluent wastewater of the pond systems is shown in Table 2. The wastewater characteristics at sampling point, ‘a’ and ‘d’, in Table 2 corresponds, respectively to raw wastewater from starch and glucose factory. Raw wastewater from starch factory was highly acidic in nature while from glucose factory was low acidic to neutral. As can be seen in Table 2, the major pollution load was due to wastewater from starch factory having BOD5 of 12,776 ± 499 mg/l as compared to BOD5 of 1046 ± 153 mg/l from glucose factory. The starch factory wastewater also had TSS of 9130 ± 3067 mg/l mainly as starch granules, which were highly biodegradable by nature. A cyanide concentration of 17.5 ± 1.5 mg/l was found in starch factory wastewater while no cyanide was detected in wastewater from glucose factory. Performance of anaerobic ponds: The details of the pond area and residence time are presented in Table 1. The overall residence time works out to be 33 ± 5 days for starch and 181 ± 33 days for glucose factory wastewater. The average pollution load for the total wastewater flows of 4999 ± 785 m3 /d calculated to be 63,258 ± 10,198 kg COD/d with 62,732 ± 10,152 kg COD/d from starch and 658 ± 138 kg COD/d from glucose factory. The average overall volumetric loading in the anaerobic ponds was 497 ± 82 kg BOD5 /m3 d (514 ± 82 kg COD/m3 d).
Starch wastewater 1 Glucose wastewater d
a
3
2
4 b
c
6
5 f
e Anaerobic ponds
Sampling point: a.
b. c. d. e. f.
Starch wastewater/influent to pond 2 effluent from pond 2/ influent to pond 4 effluent from pond 4/ influent to pond 5 Glucose wastewater/ influent to pond 3 effluent from pond 3/ influent to pond 5 effluent from pond 5/ influent to pond 6
Fig. 3. Scheme of anaerobic ponds and sampling points.
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
139
Table 2 Characterization of the influent and effluent of the pond system Parameters
Sampling points a
b
c
d
e
f
COD (mg/l) BOD5 (mg/l) TSS (mg/l) DS (mg/l) pH DO (mg/l) CN (mg/l)
13,941 ± 359 12,776 ± 499 9130 ± 3067 12,400 ± 133 4.2 ± 0.4 2.0 ± 0.4 17.5 ± 1.5
12,468 ± 930 11,700 ± 1249 7740 ± 3210 9900 ± 1150 4.2 ± 0.1 2.2 ± 0.3 17.0 ± 1.5
1414 ± 35 1102 ± 47 900 ± 314 4950 ± 235 5.7 ± 0.4 0.3 ± 0.1 10.5 ± 1.3
1314 ± 127 1046 ± 153 970 ± 285 5900 ± 763 6.8 ± 0.4 0.4 ± 0.2 Nil
952 ± 36 775 ± 42 400 ± 73 4625 ± 190 6.9 ± 0.1 0.8 ± 0.1 Nil
538 ± 94 230 ± 60 450 ± 83 3540 ± 125 7.6 ± 0.3 2.7 ± 1.5 8.5 ± 0.7
Results are mean of nine values ± standard deviation.
Out of six anaerobic ponds, Ponds #1 and #6 were not in operation during the study period. Pond #1 was filled up due to accumulation of starch granules from starch wastewater so wastewater from the starch factory was introduced into Pond #2. The average COD, BOD5 and TSS removal in Pond #2 is very small, about 10.5 ± 6.8%, 8.6 ± 6.2% and 18.0 ± 10.9%, respectively (Table 3). It was observed that Pond #2 was also partially filled up by starch granules and a channel was formed where wastewater flowed to Pond #4. This indicates that Ponds #1 and #2 operate mainly as settling basins for the suspended solids, and hence, they need to be desludged regularly. Accumulation of starch granules in the pond also reduces the residence time in the pond significantly. The pH in Pond #2 was acidic, in the range 4.1–4.3. Under this condition, methanogenesis cannot occur, as this condition is highly unfavorable for the growth of methanogenic bacteria (Duarte and Anderson, 1982). This is further brought out by the fact that BOD5 removal in Pond #2 was less than 10%. However, in Ponds #4 and #5, the pH was between 6 and 8 as these ponds were active anaerobically. In fact, intense biological activity was observed in these two ponds as evidenced by formation of excessive gas bubbles and the existence of floating sludge on the pond surface. According to Zehnder et al. (1982), the optimum pH range for all methanogenic bacteria is between 6.0 and 8.0, but the optimum value for the group as a whole is close to 7.0. Van Haandel and Lettinga (1994) reported the same observation.
Based on data in Table 3, it is apparent that the performance of Ponds #4 and #5 is satisfactory. Pond #4 was the most efficient one and provided average COD, BOD5 and TSS removal of 88.6 ± 0.6%, 90.5 ± 0.6% and 87.6 ± 2.8%, respectively. The average volumetric loading of 1031 ± 165 g BOD5 /m3 d in Pond #2 was very high and 6 ± 2 g BOD5 /m3 d in Pond #3 was very low, while 716 ± 128 and 300 ± 47 g BOD5 /m3 d in Ponds #4 and #5, respectively, were within the range found in much of the literature (Ellis, 1980; Arthur, 1983; Gomes de Sousa, 1987; Mara and Pearson, 1998). The starch wastewater also contained 17.5 ± 1.5 mg/l of cyanide. Since the ponds have been in operation for over 20 years, it was anticipated that the sludge would be well acclimatized to cyanide present in the wastewater. Average cyanide removal of 2.8 ± 2.5%, 38.4 ± 2.6% and 9.2 ± 5.0% was observed in anaerobic Ponds #2, #4 and #5, respectively. The overall removal rate for COD, BOD5 and TSS were 96.2 ± 0.6%, 98.2 ± 0.4% and 94.7 ± 1.3% respectively (Table 3), whereas the removal efficiencies for DS and CN were 71.4 ± 1.0% and 51.2 ± 1.1%, respectively. However, the quality of treated effluent from the series of anaerobic ponds (Table 3, corresponding to sampling point ‘f ’) still did not meet the effluent standard, therefore, further treatment of treated wastewater from the anaerobic pond system is required. The COD removal efficiency is in agreement with results reported elsewhere (Annachhatre and Amatya, 2000) for UASB reactor, treating the wastewater from the same starch factory. Pena et al. (2000) studied the performance of an
Table 3 Average loading and removal rate of anaerobic ponds Pond
Volumetric loading rate (g BOD5 /m3 d)
COD (%)
BOD (%)
TSS (%)
DS (%)
1 2 3 4 5 6
Not in operation 1031 ± 165 6±2 716 ± 128 299 ± 47 Not in operation
10.5 ± 6.8 26.9 ± 7.8 88.7 ± 0.8 60.7 ± 6.6
8.6 ± 6.9 25.0 ± 7.2 90.5 ± 0.6 78.3 ± 6.2
18.0 ± 10.9 57.2 ± 7.0 87.6 ± 2.8 44.0 ± 10.6
20.2 ± 9.3 20.8 ± 7.3 49.6 ± 3.8 28.0 ± 1.3
38.4 ± 2.6 9.2 ± 5.0
Overall
497 ± 82
96.1 ± 0.8
98.2 ± 0.5
94.7 ± 1.3
71.4 ± 1.0
51.2 ± 1.1
Removal rate
Results are mean of nine values ± standard deviation.
CN (%) 2.8 ± 2.5
140
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
AP and a UASB reactor treating the same domestic sewage under the same environmental conditions and reported similar performance of these two systems. 3.2. Sludge activity The SMA test results (Fig. 4) of sludge taken from the pond sediment layer exhibited a negligible level of methane production during the first 13 h and increased afterwards. This revealed that approximately 13 h was required for conversion of organic matter from starch factory wastewater to produce a sufficient amount of organic acid (substrate for methane producing bacteria) that is required for substantial methanogenic activity. However instantaneous methane production was observed in the case of sludge taken from pond bulk liquid layer because of the presence of residual organic acids in the liquid. The SMA test result obtained from this study is given in Table 4 along with the other values reported in literature (Valcke and Verstraete, 1983; James et al., 1990; Ince et al., 1995). As can be seen from Table 4, the
Cum. methane (ml)
60 y = 29.393x - 16.248 R2 = 0.9776
50
3.3. Suspended solid settling in ponds
40 30
Removal of organic matter in anaerobic ponds is brought about both by sedimentation and anaerobic digestion (Oswald, 1968; McGarry and Pescod, 1970). Anaerobic wastewater stabilization ponds are considered to be an important first-step treatment as they
20 10
VSS = 1.42 g
0 (a) 0.0
0.5
1.0
potential methane production rate of 20.7 and 11.3 ml CH4 /g VSS d for the sludge from pond sediment and bulk liquid layer, respectively, obtained from this study was lower than those reported values. This could explain the relatively long retention time required for AP compared to UASB when treating the same wastewater. Since APs are a low rate system they require retention times between 1 and 2 days at temperature around 25 C to achieve 70–80% of BOD5 removal efficiency, depending on the wastewater strength (Mara et al., 1992). UASB reactors also achieve the same level of treatment but at shorter retention times of around 6–8 h (Van Haandel and Lettinga, 1994). The cyanide degradability test of anaerobic sludge obtained from the Pond #4 sediment layer was conducted for initial cyanide concentration of 10 and 20 mg/l. The slope of the line for an initial cyanide concentration of 10 mg/l (Fig. 5) provides an average cyanide degradation rate of 4.02 mg CN /l d, corresponding to an average cyanide degradability of 0.43 mg CN /g VSS d. Likewise, an average cyanide degradation rate of 7.83 mg CN /l d was obtained for an initial cyanide concentration of 20 mg/l, yielding an average cyanide degradability of 0.84 mg CN /g VSS d.
1.5
2.0
2.5
6
y = 1.8153x + 1.6847 R 2 = 0.8188 Cum. CN - degration (mg /l)
Cum. methane (ml)
8
4 2
VSS = 0.16 g
0
(b)
0.0
0.5
1.0
1.5
2.0
2.5
Time (day) Replica 1
20
y = 7.828x R2 = 0.9581
VSS = 0.65 g
15 10 5 0 0.0
y = 4.0225x R 2 = 0.9728 0.5
1.0
1.5
2.0
2.5
Time (day)
Replica 2
Initial CN=10 mg/l
Fig. 4. SMA on sludge from (a) anaerobic pond sediment layer and (b) anaerobic pond bulk liquid layer.
Initial CN=20 mg/l
Fig. 5. Cumulative cyanide degradation against time.
Table 4 Comparison of SMA results Biomass from
Feed
Temperature (C)
SMA (ml CH4 /g VSS d)
Reference
Pond sediment layer Pond bulk liquid layer UASB UASB UASB
Starch wastewater Starch wastewater Brewery wastewater Acetate Acetate
30 30 36 ± 1 30 35
20.7 11.3 50 71–103 200–400
This study This study Ince et al. (1995) Valcke and Verstraete (1983) James et al. (1990)
1.0 S0=12528 mg/l
0.8
S0=9126 mg/l S0=7780 mg/l
0.6
S0=5112 mg/l S0=4588 mg/l
0.4
S0=3508 mg/l S0=3302 mg/l
0.2
S0=1632 mg/l S0=630 mg/l
Half removal time (min)
Fraction of TSS removal (S0-S)/S0
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143 80 60 y = 705.61x -0.4607 R2 = 0.8036
40 20 0 0
S0=490 mg/l Model(th=9 min)
0.0 0
50 100 150 200 250 300
Model(th=40 min)
141
3000 6000 9000 12000 15000 Influent TSS concentration (mg/l)
Fig. 7. Half-removal time against influent TSS.
Settling time (min)
Fig. 6. Settling curve for various influent suspended solid concentrations.
allow for settleable materials in the raw wastewater to separate out and fall to the bottom sludge zone (Saqqar and Pescod, 1995). Considering the importance of settling of suspended particles, the settling experiment was carried out to study the settling characteristics of suspended solid from starch wastewater. The relationship between settling time and suspended solids removal under various influent suspended solid concentrations found from experiment is shown in Fig. 6. The figure shows that approximately 90–95% of influent TSS removal occurs within 120 min due to sedimentation at and above 1600 mg/l influent TSS concentration. However nearly 70% and 60% removal occurs within 120 min in the case of 630 and 490 mg/l of influent TSS, respectively. The settling time of 120 min was very small compared to the retention time of the ponds, which revealed that the suspended solid settlement and accumulation mainly occurs near the inlet zone of the ponds. However the actual suspended solids removal in anaerobic ponds was observed to be less than the value expected from settling experiment. This was because of short-circuiting in case of Pond #2 due to the accumulation of starch granules and reducing the actual volume of ponds. In the case of APs #3 to 5, the re-suspension of settled solids occurred due to bubbling up of biogas as well as scouring of settled materials near the pond outlet zone along the outflow. The re-suspended solids carried out with pond effluent were the cause of reduced suspended solid removal efficiency of the ponds. Tay (1982) proposed a settling model given in Eq. (1) based on the settling characteristics of suspension and hydraulic characteristics of the tank for settling performance. The model considered detention time and halfsettling time as hydraulic characteristics and settling characteristics, respectively. Half-removal time is defined as the time at which 50% of influent suspended solid is removed.
S0 S T ¼ S0 T þ th
ð1Þ
where S0 ¼ influent TSS, mg/l; S ¼ effluent TSS, mg/l; T ¼ detention time, min; th ¼ half-removal time, min. The half-removal time with respect to influent suspended solid concentration obtained from the settling experiment with wastewater from starch factory is shown in Fig. 7. It appears that the half-removal time value at 1600 mg/l of influent TSS is on character data (Fig. 7). By ignoring this data a correlation line is drawn which follows a relationship of the type shown in Eq. (2) (R2 ¼ 0:80). th ¼ 705:61ðS0 Þ
0:4607
ð2Þ
For influent TSS of 500 and 12,500 mg/l, the half-removal time is calculated to be 9 and 40 min, respectively, from Eq. (2). The relation of TSS removal and settling time at half-removal time of 9 and 40 min obtained from model Eq. (1) is shown in Fig. 6. Most of the points obtained from settling experiment above influent TSS of 630 mg/l falls within the range of model values for 9 and 40 min half-removal times. It implies that the model is applicable to define the settling characteristics of starch wastewater.
4. Conclusions This study essentially focused on evaluating the efficiency of a series of anaerobic ponds treating high organic carbon and cyanide containing wastewater from a tapioca starch factory. SMA of sludge obtained from pond sediments as well from pond bulk liquid layer was assessed in order to determine the relative conversion rates in bulk liquid and sediment layer. Cyanide degradability of sludge taken from the pond sediment layer was also assessed. Furthermore, investigation on TSS settling characteristics of starch wastewater was also carried out. The existing anaerobic pond system effectively removed organic carbon and suspended solids. However,
142
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
the treated effluent required further treatment to meet the effluent standard before final discharge into any surface water. Overall COD and TSS removal of over 90% was achieved as influent COD of 13,941 ± 359 mg/l was reduced to less than 700 mg/l and influent TSS of about 9130 ± 3067 mg/l was reduced to less than 600 mg/l. Overall CN removal of 51% was observed from the system by reducing the concentration from 17.5 ± 1.5 mg/l in starch factory wastewater to 8.5 ± 0.7 mg/l in the final effluent of series of anaerobic pond system. Active biomass from the anaerobic pond sediments and bulk liquid had SMA of 20.7 and 11.3 ml CH4 /g VSS d, respectively. Likewise cyanide degradability of sludge obtained were 0.43–0.84 mg CN /g VSS d at initial cyanide concentration of 10–20 mg/l, respectively. Settling column test on wastewater from starch factory revealed that settling time of approximately 120 min were sufficient to remove 90–95% of the influent TSS.
Acknowledgements This research was carried out under ‘‘Modeling Tools for Environment and Resource Management (MTERM)’’ project funded by ‘‘Danish International Development Assistance (Danida)’’ and ‘‘Waste Water Treatment and Management’’ project under ‘‘Asian Regional Research Program on Environmental Technology (ARRPET)’’ funded by ‘‘Swedish International Development Agency (SIDA)’’. The authors are thankful to Prof. Jean-Luc VASEL, Foundation Universitaire Luxembourgeoise, Arlon, Belgium for his critical suggestions throughout this research.
References Annachhatre, A.P., Amatya, P.L., 2000. UASB treatment of tapioca starch wastewater. J. Environ. Eng. 126 (12), 1149–1152. Annachhatre, A.P., Amornkaew, A., 2000. Toxicity and degradation of cyanide in batch methanogenesis. Environ. Technol. 21, 135– 145. Annachhatre, A.P., Amornkaew, A., 2001. Upflow anaerobic sludge blanket treatment of starch wastewater containing cyanide. Water Environ. Res. 73 (5), 622–632. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. APHA, Washington, DC. Arthur, J.P., 1983. Notes on the design and operation of waste stabilization ponds in warm climates of developing countries. Technical paper no. 7. World Bank, Washington, DC. Bengtsson, B.E., Treit, T., 1994. Tapioca starch wastewater toxicity characterization by microtox and duckweed test. Ambio 23, 473– 477. Choi, Y., 2001. Generation and Reuse of Wastewater from Starch Industry. Master thesis, Asian Institute of Technology, Thailand. Duarte, A.C., Anderson, G.K., 1982. Inhibition modeling in anaerobic digestion. Water Sci. Technol. 14, 749–763.
Economic and Social Commission for Asia and The Pacific (ESCAP), 1982. Industrial pollution control guidelines. Vol. V: Tapioca Industry, Bangkok, Thailand, p. 17. Ellis, K.V., 1980. Stabilization ponds water quality, preliminary treatment and pre-treatment. In: Waste stabilization ponds, design and operation, WHO/EMRO. Technical publication no. 3, pp. 181–206. Fedorak, P.M., Roberts, D.J., Hrudey, S.E., 1986. The effects of cyanide on the methanogenic degradation of phenolic compounds. Water Res. 20, 1315–1320. Gomes de Sousa, J.M., 1987. Wastewater stabilization lagoon design criteria for Portugal. Water Sci. Technol. 19 (12), 7–16. Harper, S.R., Cross, W.H., Pohland, F.G., Chain, E.S.K., 1983. Adsorption enhanced biogasification of coal conversion wastewater. Biotechnol. Bioeng. Symp. 13, 401–420. Ince, O., Anderson, G.K., Kasapgil, B., 1995. Control of organic loading rate using the specific methanogenic activity test during start-up of an anaerobic digestion system. Water Res. 29 (1), 349– 355. James, A., Chernicharo, C.A.L., Campos, C.M.M., 1990. The development of a new methodology for the assessment of specific methanogenic activity. Water Res. 24 (7), 813–825. Kiravanich, P., 1977. Pollution control in the tapioca starch industry in Thailand. National Environment Board, Bangkok, Thailand. Mara, D., Pearson, H., 1998. Design Manual for Waste Stabilization Ponds in Mediterranean Countries Leeds. Lagoon Technology International, England. Mara, D.D., Alabaster, G.P., Pearson, H.W., Mills, S.W., 1992. Waste stabilization ponds: a design manual for Eastern Africa Leeds. Lagoon Technology International, England. McGarry, M.G., Pescod, M.B., 1970. Stabilization pond design criteria for tropical Asia. In: 2nd International Symposium for Waste Treatment Lagoons, Kansas city, pp. 114–132. Oswald, W.J., 1968. Advances in Anaerobic Pond Systems Design. Advances in Water Quality Improvement. University of Texas, Austin. Padmaja, G., Balgopalan, C., Kurup, G.T., Moorthy, S.N., Nanda, S.K., 1990. Cassava processing, marketing and utilization in India. In: ‘‘Cassava breading, agronomy and utilization research in Asia’’, Proceeding of the Third Regional Workshop, Malang, Indonesia, October 1990. Centro International de Agriculture Tropical (CIAT), pp. 327–338. Paing, J., Picot, B., Sambuco, J.P., 2003. Emission of H2 S and mass balance of sulfur in anaerobic ponds. Water Sci. Technol. 48 (2), 227–234. Parker, C.D., 1979. Biological mechanisms in lagoons. Prog. Water Technol. 11 (4/5), 71–85. Pena, M.R., Rodriguez, R., Mara, D.D., Sepulveda, M., 2000. UASBs or anaerobic ponds in warm climates? A preliminary answer from Colombia. Water Sci. Technol. 42 (10/11), 59–65. Pescod, M.B., 1996. The role and limitations of anaerobic pond systems. Water Sci. Technol. 33 (7), 11–21. Saqqar, M.M., Pescod, M.B., 1995. Modeling the performance of anaerobic wastewater stabilization ponds. Water Sci. Technol. 31 (12), 171–183. Tay, J.H., 1982. Development of settling model for primary settling tanks. Water Res. 16, 1413–1417. Toprak, H., 1994. Empirical modeling of sedimentation which occurs in anaerobic waste stabilization ponds using a lab-scale semicontinuous reactor. Environ. Technol. 15, 125–134. Toprak, H., 1997. Hydrogen sulfide emission rates originating from anaerobic waste stabilization ponds. Environ. Technol. 18, 795– 805. Valcke, D., Verstraete, W., 1983. A practical method to estimate the acetolastic methanogenic biomass in anaerobic reactors. J. WPCF 55, 1191–1195.
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143 Van Haandel, A.C., Lettinga, G., 1994. Anaerobic Sewage Treatment. A Practical Guide for Regions with a Hot Climate. John Wiley & Sons, Chichester, UK.
143
Zehnder, A.J., Ingvorsen, K., Marti, T., 1982. Microbiology of methanogenic bacteria. In: Anaerobic Digestion. Elsevier, Amsterdam, NL.
Anaerobic ponds treatment of starch wastewater: case study in Thailand B.K. Rajbhandari, A.P. Annachhatre
*
Environmental Engineering and Management, Asian Institute of Technology, P.O. Box 4, Klong Luang, Pathumthani 12120, Thailand Received in revised form 20 January 2004; accepted 26 January 2004 Available online 12 March 2004
Abstract Anaerobic ponds are particularly effective in treating high-strength wastewater containing biodegradable solids as they achieve the dual purpose of particulate settlement and organic removal. Performance of an anaerobic pond system for treatment of starch wastewater containing high organic carbon, biodegradable starch particulate matter and cyanide was assessed under tropical climate conditions. Approximately 5000 m3 /d of wastewater from starch industry was treated in a series of anaerobic ponds with a total area of 7.39 ha followed by facultative ponds with an area of 29.11 ha. Overall COD and TSS removal of over 90% and CN removal of 51% was observed. Active biomass obtained from the anaerobic ponds sediments and bulk liquid layer exhibited specific methanogenic activity of 20.7 and 11.3 ml CH4 /g VSS d, respectively. The cyanide degradability of sludge at initial cyanide concentration of 10 and 20 mg/l were determined to be 0.43 and 0.84 mg CN /g VSS d, respectively. A separate settling column experiment with starch wastewater revealed that a settling time of approximately 120 min is sufficient to remove 90–95% of the influent TSS. 2004 Elsevier Ltd. All rights reserved. Keywords: Anaerobic pond; Cyanide degradability; Organic carbon; Settling characteristics; Specific methanogenic activity; Starch factory wastewater
1. Introduction Anaerobic ponds (APs) are popularly employed for treatment of organic wastewater emanating from variety of industries such as food, pulp and paper, sugar and distillery. Anaerobic ponds are particularly effective in treating high-strength wastewaters containing biodegradable total suspended solids (TSS). In such cases the liquid layer in anaerobic ponds act as a settling basin for the suspended solids while the anaerobic biodegradation primarily takes place in pond sediments (Toprak, 1994). Anaerobic reactions taking place in the sediment include solubilization of biodegradable particulate matter followed by acidogenesis, acetogenesis and methanogenesis (Parker, 1979; Pescod, 1996). The reactions occurring in the bulk liquid are often negligible as compared to those in the pond sediments. Thus, anaerobic ponds achieve a dual purpose of sedimentation of particulate matter as well as anaerobic conversion of organics. However,
*
Corresponding author. Tel./fax: +66-2-524-5644. E-mail address: [email protected] (A.P. Annachhatre).
0960-8524/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2004.01.017
anaerobic pond operation also has many intrinsic problems such as high land requirements and emission of obnoxious and greenhouse gases such as hydrogen sulfide (H2 S), carbon dioxide (CO2 ) and methane (CH4 ) (Parker, 1979; Pescod, 1996; Toprak, 1997; Paing et al., 2003). In spite of these problems, anaerobic ponds are popular particularly wherever land is abundant (Arthur, 1983). Wastewater coming from starch factories is one such type of wastewater, which is treated extensively in anaerobic ponds. Starch is often produced in many parts of the world from tapioca. Tapioca roots contain 20– 25% starch. The starch extraction process essentially involves pre-processing of roots, followed by starch extraction, separation and drying. The process generates 20–60 m3 /ton of wastewater with a low pH in the range 3.8–5.2 (Economic and Social Commission for Asia and The Pacific, 1982). The wastewater is highly organic in nature with chemical oxygen demand (COD) up to 25,000 mg/l (Bengtsson and Treit, 1994). The wastewater consists of high TSS comprising starch granules in the range 3000–15,000 mg/l, which are highly biodegradable by nature. Tapioca starch wastewater also has high cyanide content up to 10–15 mg/l, which is highly toxic
136
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
to aquatic life at concentrations of cyanide as low as 0.3 mg/l have been reported as cause for a massive fish kill (Bengtsson and Treit, 1994). Problems related to water pollution are reported to be serious. The acidic nature of wastewater can harm aquatic organisms and reduce the self-purification capacity of the receiving stream. Suspended solids present in the wastewater can settle on the streambed and spoil fish breeding areas in the stream. Since these solids are primarily organic in nature, they decompose easily and thus deoxygenate the water. Similarly, high biochemical oxygen demand (BOD) of the wastewater also can cause rapid depletion of oxygen content in the receiving water body and promote the growth of nuisance organisms. Water pollution caused by tapioca starch production has been reported as a serious problem in many Asian countries, particularly in Thailand (Kiravanich, 1977) and in India (Padmaja et al., 1990). Tapioca also contains bound cyanide as a natural defense mechanism. During the starch manufacturing process, bound cyanide in the form of linamarin and lotaustralin from tapioca roots is hydrolyzed by the enzyme linamarase with decomposition to hydrogen cyanide (HCN), which finds its way into the wastewater. Cyanide containing starch wastewater can be effectively detoxified in anaerobic processes (Annachhatre and Amornkaew, 2000). Upflow anaerobic sludge blanket (UASB) processes are effective in treating starch wastewater (Annachhatre and Amatya, 2000), particularly, in removing cyanide (Annachhatre and Amornkaew, 2001). Adaptation by methanogens to cyanide concentrations of 5–30 mg/l has been reported in literature (Fedorak et al., 1986; Harper et al., 1983). Thus, in treating tapioca starch wastewater anaerobic ponds achieve a threefold objective namely: sedimentation of particulate matter, anaerobic conversion of organics and detoxification of cyanide. Accordingly, the work presented here assesses the performance of APs treating wastewater from tapioca starch industry, particularly related to COD, TSS and cyanide removal. Since APs serve as a settling basin for starch granules, the settling characteristics were also assessed by column experiments. Furthermore, the potential methane production rates of anaerobic biomass (sludge) obtained from the AP sediment as well as from bulk liquid layer were assessed from the specific methanogenic activity (SMA) test. The cyanide degradability of the anaerobic sludge obtained from the pond sediment layer was also assessed.
2. Methods Investigations on the existing wastewater anaerobic pond system were carried out in a tapioca starch and glucose factory situated in the Central province of
Thailand with a capacity of 250 tons starch/day. The factory uses groundwater as a source for process water, and generates combined wastewater of approximately 5000 m3 /d. The operating ambient temperature during the period of investigation was in the range of 30–35 C. 2.1. Treatment ponds A schematic of waste stabilization pond system (WSPS) of the starch factory is presented in Fig. 1 (Choi, 2001). The wastewater treatment system consists of 21 APs and facultative ponds (FPs) connected in series with total area of about 36.5 ha. Out of these, 6 are AP with area of 7.39 ha and 15 are FP with 29.11 ha. The study concentrated on anaerobic ponds system. During the study period only four anaerobic ponds were in operation. The typical size of an AP is approximately 250 m in length, 100 m in width and 4–5 m in depth. The pond parameters are presented in Table 1. The anaerobic ponds treat wastewater from a starch as well as from a glucose factory. The wastewater from the starch factory was first introduced to Pond #2 and subsequently flows to Pond #4 while the effluent from the glucose factory was introduced to Pond #3 and then flows to Pond #5 where wastewater from the starch and glucose factory were combined. The combined wastewater then flows into a series of FPs and treated effluent was finally discharged on to the groundwater recharge spreading basins. 2.2. Sludge activity tests The schematic of the SMA test set up is presented in Fig. 2. To determine SMA, a known amount of sludge obtained from the sediment layer of Pond #4 was transferred into serum bottles (115 ml) after washing three times with water to remove existing COD. While 100 ml of bulk liquid from the same pond was kept in a serum bottle to determine SMA of sludge in suspension in the bulk liquid layer. An appropriate amount of starch factory wastewater as substrate was added to the serum bottles so as to obtain an initial COD level in the range of 2000–2500 mg/l. Nutrients were added to maintain a carbon:nitrogen:phosphorus ratio of 300:5:1. pH was adjusted to between 7 and 7.8. Two grams per liter of NaHCO3 were also added along with substrate to buffer the serum bottle contents to near neutral pH conditions during the test. Subsequently the bottles were sealed with a rubber septum and an aluminum cap after purging oxygen with nitrogen (N2 ) gas and attached to the liquid displacement system. The liquid displacement bottle contained 3% NaOH solution. Methane gas production was measured at different time intervals up to 48 h. After every gas measurement, by swirling manually, the contents of the serum bottle were mixed. The tests were conducted in a 30 C temperature-
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
137
Fig. 1. Layout of WSPS of starch factory.
Table 1 Ponds parameters APs
Surface area (ha)
Volume (m3 · 104 )
Depth (m)
Inflowb (m3 /d)
Detention timeb (d)
1a 2 3 4 5 6a
1.07 1.34 2.06 1.76 0.49 0.67
4.41 5.57 8.62 7.34 1.78 2.46
4.5 4.5 4.5 4.5 4.5 4.5
4501 ± 731 498 ± 82 4501 ± 731 4999 ± 785
12.7 ± 1.9 177.8 ± 32.8 16.7 ± 2.5 3.6 ± 0.5
a b
Pond not in operation at the time of this study. Results are mean of nine values ± standard deviation.
from Pond #4 sediment layer was kept in serum bottles and filled with 70 ml of wastewater having substrate and nutrient similar to those mentioned for the SMA test. A stock cyanide solution was added into each serum bottle to achieve cyanide concentrations of 10 and 20 mg/l respectively. The bottle was then purged with N2 gas and immediately sealed with a rubber septum and an aluminum cap. The bottle was kept in a 30 C temperature-controlled room. Samples were taken with Hamilton syringe at every 8 h interval for 48 h and analyzed for cyanide content. Fig. 2. Sludge activity test setup.
2.3. Suspended solid settling experiment controlled room. Likewise, cyanide degradation activity of anaerobic pond sediment was carried out also in the serum bottles. A known amount of sludge obtained
Batch tests to investigate TSS settling characteristics of starch factory wastewater under quiescent condition
138
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
was carried out in a settling column of 10.0 cm diameter and 2.0 m height for different TSS concentrations. Wastewater of a desired TSS concentration for settling experiment was prepared by diluting the concentrated wastewater with tap water. The wastewater was poured into a settling column after stirring thoroughly. Samples from top of column were collected at different time intervals ranging from 2 to 60 min and analyzed for TSS concentration. 2.4. Analytical procedures Parameters including COD, BOD5 , TSS, volatile suspended solid (VSS) and dissolved solid (DS) were analyzed according to Standard Methods (APHA et al., 1998). The mass of the sludge used in the sludge activity test was measured in terms of VSS. All the samples were filtered through 0.45 lm glass fiber filters for the determination of soluble COD and BOD5 . Cyanide was measured spectrophotometrically (Spectroquant, E. Merck KGaA, Darmstadt, Germany) as per the procedure reported elsewhere (Annachhatre and Amornkaew, 2000). 2.5. Statistical analysis The anaerobic ponds process performance data were presented in terms of arithmetic averages of nine values ± standard deviation. The SMA tests were carried out with two replicates. Common linear regression curve was fitted to the data obtained from the two replicates test and a relation between volumes of methane production with respect to time was established. The SMA was calculated based on the slope of methane volume versus time curve and mass of sludge taken for the SMA test. Likewise a linear relation was established between the cumulative cyanide degradation with respect to time. Data from settling experiments was used to establish a non-linear relationship between the half-removal time and influent total suspended solid concentrations. All
statistical analyses (arithmetic average, standard deviation, linear and non-linear regression and correlation coefficient) were performed using Microsoft Excel 2000.
3. Results and discussion 3.1. Analysis of existing wastewater process Characteristics of raw wastewater: The pond system treats approximately 4500 m3 /d of wastewater from starch and approximately 500 m3 /d of wastewater from glucose factory. A scheme of anaerobic ponds and sampling points is shown in Fig. 3. The characterization of raw, influent and effluent wastewater of the pond systems is shown in Table 2. The wastewater characteristics at sampling point, ‘a’ and ‘d’, in Table 2 corresponds, respectively to raw wastewater from starch and glucose factory. Raw wastewater from starch factory was highly acidic in nature while from glucose factory was low acidic to neutral. As can be seen in Table 2, the major pollution load was due to wastewater from starch factory having BOD5 of 12,776 ± 499 mg/l as compared to BOD5 of 1046 ± 153 mg/l from glucose factory. The starch factory wastewater also had TSS of 9130 ± 3067 mg/l mainly as starch granules, which were highly biodegradable by nature. A cyanide concentration of 17.5 ± 1.5 mg/l was found in starch factory wastewater while no cyanide was detected in wastewater from glucose factory. Performance of anaerobic ponds: The details of the pond area and residence time are presented in Table 1. The overall residence time works out to be 33 ± 5 days for starch and 181 ± 33 days for glucose factory wastewater. The average pollution load for the total wastewater flows of 4999 ± 785 m3 /d calculated to be 63,258 ± 10,198 kg COD/d with 62,732 ± 10,152 kg COD/d from starch and 658 ± 138 kg COD/d from glucose factory. The average overall volumetric loading in the anaerobic ponds was 497 ± 82 kg BOD5 /m3 d (514 ± 82 kg COD/m3 d).
Starch wastewater 1 Glucose wastewater d
a
3
2
4 b
c
6
5 f
e Anaerobic ponds
Sampling point: a.
b. c. d. e. f.
Starch wastewater/influent to pond 2 effluent from pond 2/ influent to pond 4 effluent from pond 4/ influent to pond 5 Glucose wastewater/ influent to pond 3 effluent from pond 3/ influent to pond 5 effluent from pond 5/ influent to pond 6
Fig. 3. Scheme of anaerobic ponds and sampling points.
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
139
Table 2 Characterization of the influent and effluent of the pond system Parameters
Sampling points a
b
c
d
e
f
COD (mg/l) BOD5 (mg/l) TSS (mg/l) DS (mg/l) pH DO (mg/l) CN (mg/l)
13,941 ± 359 12,776 ± 499 9130 ± 3067 12,400 ± 133 4.2 ± 0.4 2.0 ± 0.4 17.5 ± 1.5
12,468 ± 930 11,700 ± 1249 7740 ± 3210 9900 ± 1150 4.2 ± 0.1 2.2 ± 0.3 17.0 ± 1.5
1414 ± 35 1102 ± 47 900 ± 314 4950 ± 235 5.7 ± 0.4 0.3 ± 0.1 10.5 ± 1.3
1314 ± 127 1046 ± 153 970 ± 285 5900 ± 763 6.8 ± 0.4 0.4 ± 0.2 Nil
952 ± 36 775 ± 42 400 ± 73 4625 ± 190 6.9 ± 0.1 0.8 ± 0.1 Nil
538 ± 94 230 ± 60 450 ± 83 3540 ± 125 7.6 ± 0.3 2.7 ± 1.5 8.5 ± 0.7
Results are mean of nine values ± standard deviation.
Out of six anaerobic ponds, Ponds #1 and #6 were not in operation during the study period. Pond #1 was filled up due to accumulation of starch granules from starch wastewater so wastewater from the starch factory was introduced into Pond #2. The average COD, BOD5 and TSS removal in Pond #2 is very small, about 10.5 ± 6.8%, 8.6 ± 6.2% and 18.0 ± 10.9%, respectively (Table 3). It was observed that Pond #2 was also partially filled up by starch granules and a channel was formed where wastewater flowed to Pond #4. This indicates that Ponds #1 and #2 operate mainly as settling basins for the suspended solids, and hence, they need to be desludged regularly. Accumulation of starch granules in the pond also reduces the residence time in the pond significantly. The pH in Pond #2 was acidic, in the range 4.1–4.3. Under this condition, methanogenesis cannot occur, as this condition is highly unfavorable for the growth of methanogenic bacteria (Duarte and Anderson, 1982). This is further brought out by the fact that BOD5 removal in Pond #2 was less than 10%. However, in Ponds #4 and #5, the pH was between 6 and 8 as these ponds were active anaerobically. In fact, intense biological activity was observed in these two ponds as evidenced by formation of excessive gas bubbles and the existence of floating sludge on the pond surface. According to Zehnder et al. (1982), the optimum pH range for all methanogenic bacteria is between 6.0 and 8.0, but the optimum value for the group as a whole is close to 7.0. Van Haandel and Lettinga (1994) reported the same observation.
Based on data in Table 3, it is apparent that the performance of Ponds #4 and #5 is satisfactory. Pond #4 was the most efficient one and provided average COD, BOD5 and TSS removal of 88.6 ± 0.6%, 90.5 ± 0.6% and 87.6 ± 2.8%, respectively. The average volumetric loading of 1031 ± 165 g BOD5 /m3 d in Pond #2 was very high and 6 ± 2 g BOD5 /m3 d in Pond #3 was very low, while 716 ± 128 and 300 ± 47 g BOD5 /m3 d in Ponds #4 and #5, respectively, were within the range found in much of the literature (Ellis, 1980; Arthur, 1983; Gomes de Sousa, 1987; Mara and Pearson, 1998). The starch wastewater also contained 17.5 ± 1.5 mg/l of cyanide. Since the ponds have been in operation for over 20 years, it was anticipated that the sludge would be well acclimatized to cyanide present in the wastewater. Average cyanide removal of 2.8 ± 2.5%, 38.4 ± 2.6% and 9.2 ± 5.0% was observed in anaerobic Ponds #2, #4 and #5, respectively. The overall removal rate for COD, BOD5 and TSS were 96.2 ± 0.6%, 98.2 ± 0.4% and 94.7 ± 1.3% respectively (Table 3), whereas the removal efficiencies for DS and CN were 71.4 ± 1.0% and 51.2 ± 1.1%, respectively. However, the quality of treated effluent from the series of anaerobic ponds (Table 3, corresponding to sampling point ‘f ’) still did not meet the effluent standard, therefore, further treatment of treated wastewater from the anaerobic pond system is required. The COD removal efficiency is in agreement with results reported elsewhere (Annachhatre and Amatya, 2000) for UASB reactor, treating the wastewater from the same starch factory. Pena et al. (2000) studied the performance of an
Table 3 Average loading and removal rate of anaerobic ponds Pond
Volumetric loading rate (g BOD5 /m3 d)
COD (%)
BOD (%)
TSS (%)
DS (%)
1 2 3 4 5 6
Not in operation 1031 ± 165 6±2 716 ± 128 299 ± 47 Not in operation
10.5 ± 6.8 26.9 ± 7.8 88.7 ± 0.8 60.7 ± 6.6
8.6 ± 6.9 25.0 ± 7.2 90.5 ± 0.6 78.3 ± 6.2
18.0 ± 10.9 57.2 ± 7.0 87.6 ± 2.8 44.0 ± 10.6
20.2 ± 9.3 20.8 ± 7.3 49.6 ± 3.8 28.0 ± 1.3
38.4 ± 2.6 9.2 ± 5.0
Overall
497 ± 82
96.1 ± 0.8
98.2 ± 0.5
94.7 ± 1.3
71.4 ± 1.0
51.2 ± 1.1
Removal rate
Results are mean of nine values ± standard deviation.
CN (%) 2.8 ± 2.5
140
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
AP and a UASB reactor treating the same domestic sewage under the same environmental conditions and reported similar performance of these two systems. 3.2. Sludge activity The SMA test results (Fig. 4) of sludge taken from the pond sediment layer exhibited a negligible level of methane production during the first 13 h and increased afterwards. This revealed that approximately 13 h was required for conversion of organic matter from starch factory wastewater to produce a sufficient amount of organic acid (substrate for methane producing bacteria) that is required for substantial methanogenic activity. However instantaneous methane production was observed in the case of sludge taken from pond bulk liquid layer because of the presence of residual organic acids in the liquid. The SMA test result obtained from this study is given in Table 4 along with the other values reported in literature (Valcke and Verstraete, 1983; James et al., 1990; Ince et al., 1995). As can be seen from Table 4, the
Cum. methane (ml)
60 y = 29.393x - 16.248 R2 = 0.9776
50
3.3. Suspended solid settling in ponds
40 30
Removal of organic matter in anaerobic ponds is brought about both by sedimentation and anaerobic digestion (Oswald, 1968; McGarry and Pescod, 1970). Anaerobic wastewater stabilization ponds are considered to be an important first-step treatment as they
20 10
VSS = 1.42 g
0 (a) 0.0
0.5
1.0
potential methane production rate of 20.7 and 11.3 ml CH4 /g VSS d for the sludge from pond sediment and bulk liquid layer, respectively, obtained from this study was lower than those reported values. This could explain the relatively long retention time required for AP compared to UASB when treating the same wastewater. Since APs are a low rate system they require retention times between 1 and 2 days at temperature around 25 C to achieve 70–80% of BOD5 removal efficiency, depending on the wastewater strength (Mara et al., 1992). UASB reactors also achieve the same level of treatment but at shorter retention times of around 6–8 h (Van Haandel and Lettinga, 1994). The cyanide degradability test of anaerobic sludge obtained from the Pond #4 sediment layer was conducted for initial cyanide concentration of 10 and 20 mg/l. The slope of the line for an initial cyanide concentration of 10 mg/l (Fig. 5) provides an average cyanide degradation rate of 4.02 mg CN /l d, corresponding to an average cyanide degradability of 0.43 mg CN /g VSS d. Likewise, an average cyanide degradation rate of 7.83 mg CN /l d was obtained for an initial cyanide concentration of 20 mg/l, yielding an average cyanide degradability of 0.84 mg CN /g VSS d.
1.5
2.0
2.5
6
y = 1.8153x + 1.6847 R 2 = 0.8188 Cum. CN - degration (mg /l)
Cum. methane (ml)
8
4 2
VSS = 0.16 g
0
(b)
0.0
0.5
1.0
1.5
2.0
2.5
Time (day) Replica 1
20
y = 7.828x R2 = 0.9581
VSS = 0.65 g
15 10 5 0 0.0
y = 4.0225x R 2 = 0.9728 0.5
1.0
1.5
2.0
2.5
Time (day)
Replica 2
Initial CN=10 mg/l
Fig. 4. SMA on sludge from (a) anaerobic pond sediment layer and (b) anaerobic pond bulk liquid layer.
Initial CN=20 mg/l
Fig. 5. Cumulative cyanide degradation against time.
Table 4 Comparison of SMA results Biomass from
Feed
Temperature (C)
SMA (ml CH4 /g VSS d)
Reference
Pond sediment layer Pond bulk liquid layer UASB UASB UASB
Starch wastewater Starch wastewater Brewery wastewater Acetate Acetate
30 30 36 ± 1 30 35
20.7 11.3 50 71–103 200–400
This study This study Ince et al. (1995) Valcke and Verstraete (1983) James et al. (1990)
1.0 S0=12528 mg/l
0.8
S0=9126 mg/l S0=7780 mg/l
0.6
S0=5112 mg/l S0=4588 mg/l
0.4
S0=3508 mg/l S0=3302 mg/l
0.2
S0=1632 mg/l S0=630 mg/l
Half removal time (min)
Fraction of TSS removal (S0-S)/S0
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143 80 60 y = 705.61x -0.4607 R2 = 0.8036
40 20 0 0
S0=490 mg/l Model(th=9 min)
0.0 0
50 100 150 200 250 300
Model(th=40 min)
141
3000 6000 9000 12000 15000 Influent TSS concentration (mg/l)
Fig. 7. Half-removal time against influent TSS.
Settling time (min)
Fig. 6. Settling curve for various influent suspended solid concentrations.
allow for settleable materials in the raw wastewater to separate out and fall to the bottom sludge zone (Saqqar and Pescod, 1995). Considering the importance of settling of suspended particles, the settling experiment was carried out to study the settling characteristics of suspended solid from starch wastewater. The relationship between settling time and suspended solids removal under various influent suspended solid concentrations found from experiment is shown in Fig. 6. The figure shows that approximately 90–95% of influent TSS removal occurs within 120 min due to sedimentation at and above 1600 mg/l influent TSS concentration. However nearly 70% and 60% removal occurs within 120 min in the case of 630 and 490 mg/l of influent TSS, respectively. The settling time of 120 min was very small compared to the retention time of the ponds, which revealed that the suspended solid settlement and accumulation mainly occurs near the inlet zone of the ponds. However the actual suspended solids removal in anaerobic ponds was observed to be less than the value expected from settling experiment. This was because of short-circuiting in case of Pond #2 due to the accumulation of starch granules and reducing the actual volume of ponds. In the case of APs #3 to 5, the re-suspension of settled solids occurred due to bubbling up of biogas as well as scouring of settled materials near the pond outlet zone along the outflow. The re-suspended solids carried out with pond effluent were the cause of reduced suspended solid removal efficiency of the ponds. Tay (1982) proposed a settling model given in Eq. (1) based on the settling characteristics of suspension and hydraulic characteristics of the tank for settling performance. The model considered detention time and halfsettling time as hydraulic characteristics and settling characteristics, respectively. Half-removal time is defined as the time at which 50% of influent suspended solid is removed.
S0 S T ¼ S0 T þ th
ð1Þ
where S0 ¼ influent TSS, mg/l; S ¼ effluent TSS, mg/l; T ¼ detention time, min; th ¼ half-removal time, min. The half-removal time with respect to influent suspended solid concentration obtained from the settling experiment with wastewater from starch factory is shown in Fig. 7. It appears that the half-removal time value at 1600 mg/l of influent TSS is on character data (Fig. 7). By ignoring this data a correlation line is drawn which follows a relationship of the type shown in Eq. (2) (R2 ¼ 0:80). th ¼ 705:61ðS0 Þ
0:4607
ð2Þ
For influent TSS of 500 and 12,500 mg/l, the half-removal time is calculated to be 9 and 40 min, respectively, from Eq. (2). The relation of TSS removal and settling time at half-removal time of 9 and 40 min obtained from model Eq. (1) is shown in Fig. 6. Most of the points obtained from settling experiment above influent TSS of 630 mg/l falls within the range of model values for 9 and 40 min half-removal times. It implies that the model is applicable to define the settling characteristics of starch wastewater.
4. Conclusions This study essentially focused on evaluating the efficiency of a series of anaerobic ponds treating high organic carbon and cyanide containing wastewater from a tapioca starch factory. SMA of sludge obtained from pond sediments as well from pond bulk liquid layer was assessed in order to determine the relative conversion rates in bulk liquid and sediment layer. Cyanide degradability of sludge taken from the pond sediment layer was also assessed. Furthermore, investigation on TSS settling characteristics of starch wastewater was also carried out. The existing anaerobic pond system effectively removed organic carbon and suspended solids. However,
142
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143
the treated effluent required further treatment to meet the effluent standard before final discharge into any surface water. Overall COD and TSS removal of over 90% was achieved as influent COD of 13,941 ± 359 mg/l was reduced to less than 700 mg/l and influent TSS of about 9130 ± 3067 mg/l was reduced to less than 600 mg/l. Overall CN removal of 51% was observed from the system by reducing the concentration from 17.5 ± 1.5 mg/l in starch factory wastewater to 8.5 ± 0.7 mg/l in the final effluent of series of anaerobic pond system. Active biomass from the anaerobic pond sediments and bulk liquid had SMA of 20.7 and 11.3 ml CH4 /g VSS d, respectively. Likewise cyanide degradability of sludge obtained were 0.43–0.84 mg CN /g VSS d at initial cyanide concentration of 10–20 mg/l, respectively. Settling column test on wastewater from starch factory revealed that settling time of approximately 120 min were sufficient to remove 90–95% of the influent TSS.
Acknowledgements This research was carried out under ‘‘Modeling Tools for Environment and Resource Management (MTERM)’’ project funded by ‘‘Danish International Development Assistance (Danida)’’ and ‘‘Waste Water Treatment and Management’’ project under ‘‘Asian Regional Research Program on Environmental Technology (ARRPET)’’ funded by ‘‘Swedish International Development Agency (SIDA)’’. The authors are thankful to Prof. Jean-Luc VASEL, Foundation Universitaire Luxembourgeoise, Arlon, Belgium for his critical suggestions throughout this research.
References Annachhatre, A.P., Amatya, P.L., 2000. UASB treatment of tapioca starch wastewater. J. Environ. Eng. 126 (12), 1149–1152. Annachhatre, A.P., Amornkaew, A., 2000. Toxicity and degradation of cyanide in batch methanogenesis. Environ. Technol. 21, 135– 145. Annachhatre, A.P., Amornkaew, A., 2001. Upflow anaerobic sludge blanket treatment of starch wastewater containing cyanide. Water Environ. Res. 73 (5), 622–632. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. APHA, Washington, DC. Arthur, J.P., 1983. Notes on the design and operation of waste stabilization ponds in warm climates of developing countries. Technical paper no. 7. World Bank, Washington, DC. Bengtsson, B.E., Treit, T., 1994. Tapioca starch wastewater toxicity characterization by microtox and duckweed test. Ambio 23, 473– 477. Choi, Y., 2001. Generation and Reuse of Wastewater from Starch Industry. Master thesis, Asian Institute of Technology, Thailand. Duarte, A.C., Anderson, G.K., 1982. Inhibition modeling in anaerobic digestion. Water Sci. Technol. 14, 749–763.
Economic and Social Commission for Asia and The Pacific (ESCAP), 1982. Industrial pollution control guidelines. Vol. V: Tapioca Industry, Bangkok, Thailand, p. 17. Ellis, K.V., 1980. Stabilization ponds water quality, preliminary treatment and pre-treatment. In: Waste stabilization ponds, design and operation, WHO/EMRO. Technical publication no. 3, pp. 181–206. Fedorak, P.M., Roberts, D.J., Hrudey, S.E., 1986. The effects of cyanide on the methanogenic degradation of phenolic compounds. Water Res. 20, 1315–1320. Gomes de Sousa, J.M., 1987. Wastewater stabilization lagoon design criteria for Portugal. Water Sci. Technol. 19 (12), 7–16. Harper, S.R., Cross, W.H., Pohland, F.G., Chain, E.S.K., 1983. Adsorption enhanced biogasification of coal conversion wastewater. Biotechnol. Bioeng. Symp. 13, 401–420. Ince, O., Anderson, G.K., Kasapgil, B., 1995. Control of organic loading rate using the specific methanogenic activity test during start-up of an anaerobic digestion system. Water Res. 29 (1), 349– 355. James, A., Chernicharo, C.A.L., Campos, C.M.M., 1990. The development of a new methodology for the assessment of specific methanogenic activity. Water Res. 24 (7), 813–825. Kiravanich, P., 1977. Pollution control in the tapioca starch industry in Thailand. National Environment Board, Bangkok, Thailand. Mara, D., Pearson, H., 1998. Design Manual for Waste Stabilization Ponds in Mediterranean Countries Leeds. Lagoon Technology International, England. Mara, D.D., Alabaster, G.P., Pearson, H.W., Mills, S.W., 1992. Waste stabilization ponds: a design manual for Eastern Africa Leeds. Lagoon Technology International, England. McGarry, M.G., Pescod, M.B., 1970. Stabilization pond design criteria for tropical Asia. In: 2nd International Symposium for Waste Treatment Lagoons, Kansas city, pp. 114–132. Oswald, W.J., 1968. Advances in Anaerobic Pond Systems Design. Advances in Water Quality Improvement. University of Texas, Austin. Padmaja, G., Balgopalan, C., Kurup, G.T., Moorthy, S.N., Nanda, S.K., 1990. Cassava processing, marketing and utilization in India. In: ‘‘Cassava breading, agronomy and utilization research in Asia’’, Proceeding of the Third Regional Workshop, Malang, Indonesia, October 1990. Centro International de Agriculture Tropical (CIAT), pp. 327–338. Paing, J., Picot, B., Sambuco, J.P., 2003. Emission of H2 S and mass balance of sulfur in anaerobic ponds. Water Sci. Technol. 48 (2), 227–234. Parker, C.D., 1979. Biological mechanisms in lagoons. Prog. Water Technol. 11 (4/5), 71–85. Pena, M.R., Rodriguez, R., Mara, D.D., Sepulveda, M., 2000. UASBs or anaerobic ponds in warm climates? A preliminary answer from Colombia. Water Sci. Technol. 42 (10/11), 59–65. Pescod, M.B., 1996. The role and limitations of anaerobic pond systems. Water Sci. Technol. 33 (7), 11–21. Saqqar, M.M., Pescod, M.B., 1995. Modeling the performance of anaerobic wastewater stabilization ponds. Water Sci. Technol. 31 (12), 171–183. Tay, J.H., 1982. Development of settling model for primary settling tanks. Water Res. 16, 1413–1417. Toprak, H., 1994. Empirical modeling of sedimentation which occurs in anaerobic waste stabilization ponds using a lab-scale semicontinuous reactor. Environ. Technol. 15, 125–134. Toprak, H., 1997. Hydrogen sulfide emission rates originating from anaerobic waste stabilization ponds. Environ. Technol. 18, 795– 805. Valcke, D., Verstraete, W., 1983. A practical method to estimate the acetolastic methanogenic biomass in anaerobic reactors. J. WPCF 55, 1191–1195.
B.K. Rajbhandari, A.P. Annachhatre / Bioresource Technology 95 (2004) 135–143 Van Haandel, A.C., Lettinga, G., 1994. Anaerobic Sewage Treatment. A Practical Guide for Regions with a Hot Climate. John Wiley & Sons, Chichester, UK.
143
Zehnder, A.J., Ingvorsen, K., Marti, T., 1982. Microbiology of methanogenic bacteria. In: Anaerobic Digestion. Elsevier, Amsterdam, NL.