Queries for waer

Transcription

1 Queries for waer This manuscript/text has been typeset from the submitted material. Please check this proof carefully to make sure there have been no font conversion errors or inadvertent formatting errors. Allen Press.

2 Quantification of the Inert Chemical Oxygen Demand of Raw Wastewater and Evaluation of Soluble Microbial Product Production in Demo-Scale Upflow Anaerobic Sludge Blanket Reactors under Different Operational Conditions Sergio F. Aquino 1 *, Roberto M. Gloria 2, Silvana Q. Silva 2, Carlos A. L. Chernicharo 2 ABSTRACT: This paper investigates the production of soluble microbial products (SMPs) in demonstration-scale upflow anaerobic sludge blanket reactors operated under different conditions and fed with raw wastewater. The results showed that % of the influent soluble chemical oxygen demand (COD) could be considered inert to anaerobic treatment and that the amount of COD produced by biomass varied from 30 to 70 mg.l 21, accounting for 45 to 63% of the soluble effluent COD. The accumulation of SMP appeared to be dependent on the hydraulic retention time (HRT) applied to the reactors, with a larger accumulation of SMP observed at the lowest HRT (5 hours); this may have been due to stress conditions caused by high upflow velocity (1.1 m.h 21 ). In terms of residual COD characterization, ultrafiltration results showed that higher amounts of high molecular weight compounds were found when HRT was the lowest (5 hours), and that the molecular weight distribution depended on the operational condition of the reactors. Biodegradability tests showed that the low and high molecular weight SMPs were only partially degraded anaerobically (10 to 60%) and that the high molecular weight SMPs were difficult to degrade aerobically. Water Environ. Res., 81 (2009). KEYWORDS: anaerobic digestion, inert chemical oxygen demand, soluble microbial products, effluent characterization, molecular weight distribution, biodegradability test. doi: / x Introduction The inherent advantages of anaerobic reactors coupled with aerobic post-treatment have already been acknowledged in the literature and contribute to the widespread application of such systems for treating domestic wastewater as well as some industrial wastewaters. In Brazil, the combination of upflow anaerobic sludge blanket (UASB) reactor with aerobic post-treatment is a real option for wastewater treatment because of the lower production of sludge, 1 * Federal University of Ouro Preto, Department of Chemistry, Campus Morro do Cruzeiro, , Ouro Preto, MG, Brazil; sergio@iceb.ufop.br (At the time this research was conducted, post-doctoral researcher in the Department of Sanitary and Environmental Engineering, Federal University of Minas Gerais, Belo Horizonte, Brazil). 2 Federal University of Minas Gerais, Department of Sanitary and Environmental Engineering, Belo Horizonte, Brazil. lower operational costs, and higher efficiencies observed in the combined treatment. Overall, the efficiency of biological treatment is evaluated in terms of removal of organic content, which is usually expressed as chemical oxygen demand (COD) and/or biochemical oxygen demand (BOD). Chemical oxygen demand and BOD are collective and indirect parameters that do not allow one to establish the origin and chemical nature of the organic contaminants. For instance, COD effluent from biological reactors may contain the following: recalcitrant organics originally present in the influent; byproducts of incomplete degradation (metabolites) such as volatile fatty acids (VFAs) that may accumulate during anaerobic treatment; and compounds derived from biomass decay (e.g., cell debris and lysis products) or that were deliberately excreted by biomass (e.g., extracelular polymeric substances, exoenzymes, siderophores, and quorum sensing organics) to play a role in the microbial survival strategy. The latter are called soluble microbial products (SMPs) and their generation during biological treatment contribute to the filtered COD of the effluent from both aerobic and anaerobic reactors. The importance of SMPs for treatment system evaluation is already acknowledged in the literature (Rittmann and McCarty, 2001); however, the chemical identity of the main SMP is still unknown. In fact, SMPs comprise a mixture of organics that originate from different sources. According to the literature, one main source of SMP is cell death and lysis, which occurs continuously in any treatment system (Aquino and Stuckey, 2004a). In addition, SMPs are also believed to be released due to the renewal of internal cellular structures and the detachment of extracellular polymers or extracellular polymeric substances (EPS) caused by shear and/or hydrolysis (Hejzlar and Chudoba, 1986; Laspidou and Rittmann, 2002). Other, more specific SMPs may also be produced as a mechanism of defense, acclimation, or any other cellular purpose (Andrews et al., 2003; Nies, 1999). There are many studies on the factors affecting production of SMPs in bench-scale systems and current knowledge suggests that the amount, origin, and nature of an SMP depend on the feed characteristics (type and strength), environmental conditions (ph, nutrient deficiency, and presence of toxic compounds), type of

3 Table 1 Operational phases of the research. Operational Phase Duration (days) Flowrate (m 3.h 21 ) Characteristics of the UASB reactors Area (m) Height (m) Volume (m 3 ) HRT (h) v 1 (m.h 21 ) v 2 (m.h 21 ) v 1 5 upflow velocity v 2 5 velocity through the settler apertures reactor, and operational parameters (hydraulic retention time [HRT] and shock loads) (Barker and Stuckey, 1999, 2001; Boero et al., 1996; Fang and Jia, 1998; Kuo et al., 1996; Noguera et al., 1994). Soluble microbial products are typically estimated by subtracting the COD due to intermediate VFAs and due to residual substrate from the soluble effluent COD fe.g., SMPs (as COD) 5 soluble effluent COD 2 [VFAs (as COD) 1 residual substrate (as COD)]g. Therefore, most studies on SMP production are conducted in benchscale reactors fed with simple substrates or biodegradable synthetic feed. Use of more complex feed, such as raw wastewater, brings up the difficulties in ascertaining the contribution of nondegraded influent organics to the effluent COD. One way of overcoming this hurdle may be by quantifying the inert COD in the raw wastewater, which could be done by methodology developed by Orhon and his colleagues (Babuna et al., 1998; Germirli et al., 1991; Orhon et al., 1999). The experimental procedure adopted by Orhon et al. (1999) involves running two batch reactors, one fed with unfiltered wastewater and the other with filtered wastewater. To get rid of the interference of the residual COD released due to biomass decay (production of SMPs during the test), seeding is performed with a minimum amount of biomass (10 to 50 mg.l 21 volatile suspended solids [VSS]). The soluble and total COD values of periodical samples from both reactors are then analyzed until all the biological activity is depleted. Then, the measurement of the initial and final threshold values of total and soluble CODs obtained from both reactors allows one to determine the initially inert fractions of the wastewater by means of mathematical manipulation of some equations presented in Babuna et al. (1998). Given that most results on SMP estimation are obtained from labscale reactors fed on rather simple substrates, the main objectives of this paper are to present results on the assessment of inert soluble COD of raw wastewater under anaerobic conditions and to evaluate SMP accumulation in UASB reactors operated in demonstrationscale under different operational conditions and fed with domestic wastewater. Materials and Methods Experimental Setup. The experimental apparatus consisted of two identical UASB reactors designed with flexibility for testing different configurations of gas-liquid-solid separators and digestion compartments. The reactors cross section and height were assembled according to the figures depicted in Table 1, resulting in total useful volumes varying from 10.1 to 16.8 m 3. One reactor (R1) was fed with raw wastewater that had received preliminary treatment by the Arrudas Wastewater Treatment Plant located in the metropolitan region of Belo Horizonte/MG, Brazil, and the other reactor (R2) was fed with sieved wastewater. The forced sieving unit that was operated upstream reactor R2 had 1 mm apertures, which hindered the passage of larger solids and reduced the average size of the particles entering the reactor (Figure 1). The reactors were seeded with sludge from two other UASB reactors (and were also fed raw wastewater after preliminary treatment) that had been operated for over 5 years in the Federal University of Minas Gerais/Copasa Centre for Research and Training on Sanitation. The reactors were kept in the same operational phase for periods varying from approximately 50 to 90 days (Table 1), therefore allowing time for reasonable sludge Figure 1 Flow sheet of the experimental setup (A) and picture of the geminated demo-scale UASB reactors (B).

4 Table 2 Characterization of wastewater fed to the UASB reactors and used for determining inert COD. Phase COD total (mg.l -1 ) COD filtered (mg. L -1 ) BOD total (mg.l -1 ) BOD filtered (mg. L -1 ) TSS (mg.l -1 ) VSS (mg.l -1 ) SS (ml.l -1 ) acclimatization and for the UASB reactors to operate under near steady-state conditions. Samples were collected twice a week for analysis of total and filtered COD, TSS, and VSS, which were all performed according to APHA et al. (1998) in the raw and screened wastewater as well as in the effluent from both reactors. All samples were collected during a 24-hour period by using four ISCO 3700 Ò automatic samplers (Teledyne Isco, Inc., Lincoln, Nebraska). Both reactors were monitored during five operational phases (Table 1). Determining Inert Chemical Oxygen Demand. The inert COD of domestic wastewater was determined according to the methodology detailed in Babuna et al. (1998). Basically, two erlemeyers (2 L) were incubated anaerobically with a low amount of biomass (VSS o ;50 mg.l 21 ) to prevent the production of SMP during the test. One erlemeyer was incubated with raw wastewater while the other received wastewater previously filtered through 1.2 lm membranes to remove particulate solids. The content of the reactor was completely stirred by means of magnetic bars, and both reactors received 3.5 g/l of NaHCO 3 as a source of alkalinity. The extent of the reaction in terms of the concentration of total and filtered COD in the liquid phase was followed for at least 50 days. Ultrafiltration Tests. The molecular weight distribution of samples of influent and effluent was determined using an Amicon ultrafiltration cell (8000 series, model 8200; Millipore, São Paulo, Brazil) in parallel mode, as described by Aquino and Stuckey (2004a) and Barker and Stuckey (1999). Ultrafiltration was performed using membranes with the nominal cutoff of 1, 10, and 100 kda. The fraction retained in each membrane was analyzed for COD and the molecular weight distribution determined in terms of mass percentage (w/w) by performing a material balance. The masses of carbohydrates, proteins, lipids, and COD retained in each fraction were determined by material balances, and the results were expressed in mass percentage. The ultrafiltration membranes were calibrated with compounds of known molecular weight, such as glucose (180 Da) and bovine serum albumine (66 kda), which were determined as carbohydrates and proteins using, respectively, the methodologies of Lowry and Dubois described by Aquino and Stuckey (2004a). Biodegradability Tests. The aerobic biodegradability of UASB effluents was determined based on the traditional BOD test (APHA et al., 1998). Fractions (x, 1 kda and x. 1 kda) resulting from ultrafiltration of the effluent were diluted with water rich in oxygen and nutrients and incubated for either 5 or 20 days at 20 8C using activated sludge as seed. The aerobic biodegradability, defined as the ratio BOD/COD, was then estimated by knowing the initial COD of the incubated sample and by measuring BOD after 5 or 20 days. Anaerobic biodegradability was obtained by determining the cumulative methane production, adapting the specific methanogenic activity (SMA) test. The test was performed using 250-ml antibiotic flasks, which received the sample, nutrient solution (Souza et al., 2006), and seed collected from the UASB reactor. According to the previous study, a food/microorganism ratio of 0.25 resulted in the highest rate of methane production and, therefore, was used. After the samples were added, the flasks were closed, sealed, fluxed with nitrogen and carbon dioxide (70:30 v/v), and incubated at 308 C under continuous stirring (210 rpm) for either 5 or 20 days. The biogas production was measured using wetted glass syringes and the quantification of methane in the biogas was made by gas chromatography, as described by Souza et al. (2006). Estimation of Soluble Microbial Product Production. Volatile fatty acids were quantified by gas chromatography after their acidic extraction from the aqueous phase with methyl tert-butyl ether, as described by Pontes et al. (2002). The SMP was then estimated as COD in the dissolved phase according to the following equation: COD p ¼½COD eff Š ½COD inf 3 f Š ½COD vfa Š ð1þ Where COD p 5 soluble COD produced by biomass or SMP, COD eff 5 soluble effluent COD, COD inf 5 soluble influent COD, f 5 fraction of the soluble influent COD inert to anaerobic metabolism, and COD vfa 5 soluble COD due to VFA. COD vfa ¼ 0:35 3 ðformiateþþ1:07 3 ðacetateþþ1:51 3 ðpropionateþþ1:82 3 ðbutirate þ isobutirateþ þ 2:04 3 ðvalerate þ isovalerateþ ð2þ Equation 1 shows that the COD produced by biomass was assumed to be all soluble effluent COD discounted for both the COD due to any VFA (the common intermediate of anaerobic digestion) and the soluble inert COD determined experimentally. Although the raw wastewater contains a myriad of organics with different biodegradability rates and the UASB reactor was operated in HRTs lower than the time allowed in the experiments to determine the inert COD, it was assumed that only the truly inert fraction would appear in the anaerobic effluent. The reasoning here is that most soluble organics are easily taken up by microbial cell and readily converted into VFAs by the fast growing acidogenic bacteria; hence, soluble influent biodegradable COD should not appear in the effluent in significant concentrations. Results and Discussion Overall Efficiency of Upflow Anaerobic Sludge Blanket Reactors. The characteristics of raw wastewater and effluents from the UASB reactors R1 and R2 are presented in Tables 2 and 3. From the data presented in these tables, it is possible to calculate the reactor efficiency in terms of COD and solids removal. The results

5 Table 3 Characterization of effluents from the UASB reactors fed with raw (reactor R1) and sieved (reactor R2) wastewater. COD total COD filtered TSS VSS Phase R1 R2 R1 R2 R1 R2 R1 R indicate that there is no statistically significant difference between reactors R1 and R2 regarding the removal of total and filtered COD along the different operational phases studied. Such results indicate that use of the pretreatment unit did not result in improvement of effluent quality. The average removal efficiency of total COD was ;60% for phases 1 and 2, ;40% for phase 3, ;50% for phase 4, and ;60% for phase 5; the removal efficiencies for filtered COD were somewhat higher, averaging from 70 to 80% during all five operational phases investigated. The lower removal efficiency of total COD compared to filtered COD is due to the presence of suspended solids (mostly anaerobic biomass) carried out by the effluent. The effluent VSS concentration was below 70 mg.l 21 during all phases except phase 3 when, due to the higher upflow velocity imputed by the low HRT (5 hours), the VSS concentration averaged 126 mg.l 21 and 101 mg.l 21 in the effluent of reactors R1 and R2, respectively. Inert Chemical Oxygen Demand and Soluble Microbial Product Production. The raw wastewater used for inert COD determination was the same fed to the UASB reactors; the characteristics of this wastewater are presented in Table 2. As Table 2 shows, wastewater entering the UASB reactor had typical characteristics of domestic wastewater, with average COD, BOD, and suspended solids concentrations varying from 323 to 467 mg.l 21, 187 to 214 mg.l 21, and 165 to 212 mg.l 21, respectively. Figure 2 shows a typical profile of the total and filtered COD during the batch test carried out to determine the inert COD of wastewater during anaerobic conditions. The final and initial values of total and filtered COD in each batch reactor were used to estimate inert COD according to procedures detailed in Babuna et al. (1998). The experiment to determine the inert COD fraction was repeated five times during the operational phases of research and the results are presented in Table 4. Aside from the second test, Table 4 shows that the measured values of inert soluble COD are within the same range. If the odd result of test 2 is not considered, the soluble fraction of the wastewater inert to anaerobic metabolism would average % in terms of COD. These results imply that from 30 to 70 mg.l 21 of the initial soluble COD entering the UASB reactors would not be degraded and would contribute to the soluble effluent COD. According to this estimation, the inert COD would correspond to 38 to 50% of the soluble COD effluent from the UASB reactors. Surfactants are among the organics that are recalcitrant toward anaerobic degradation and would, therefore, contribute to soluble inert COD. Monitoring of anionic surfactants in the UASB reactors investigated in this paper indicated that the concentration of methylene blue active substances in the raw wastewater is within the range of 7 to 15 mg.l 21, which should account for a COD of 6 to 12 mg.l 21. Besides surfactants, other cosmetic and pharmaceutical ingredients (e.g., ethylenediaminetetraacetic acid, antibiotics, and estradiol) are likely to be present in the raw wastewater and might account for the inert COD pool. The average value of inert COD ( % of soluble influent COD) was then used, along with VFA results (see eq 1), to estimate the amount of COD produced by biomass. Figure 3 shows the average composition of soluble COD effluent from both UASB reactors during the five operational phases studied. The results confirm that, for all operational phases, the majority of soluble COD from both reactors are not due to VFAs, rather, to influent organics not degraded in the reactor (COD i ) as well as compounds produced by biomass (COD p ). Figure 3 shows that, as expected, the presence of the forced sieving unit did not affect the composition of soluble COD effluent from the UASB reactors. It can be seen that the amount of soluble COD apparently produced by biomass (COD p ) varied from 30 to 70 mg.l 21 and remained fairly constant within the same operational phase, for reactors R1 and R2. In terms of the operational phases, results show that there was a marginal increase in the concentration of COD p during phase 3, which was characterized by Figure 2 Typical profile of COD concentration in batch anaerobic reactors inoculated with raw wastewater (A) and with filtered wastewater (B) during the tests for determining inert COD under anaerobic conditions.

6 the lowest HRT (5 hours) and highest upflow velocity (1.1 m.h 21 ). Figure 4 confirms that more SMP seemed to accumulate during phase 3; this might be related to stress conditions put upon the biomass, such as those observed by Aquino and Stuckey (2004a) during hydraulic shock loads. The ratio COD p /COD inf indirectly measures the amount of substrate diverted toward SMP production, and the average values shown in Figure 4(A) (from 6 to 16%) are within the range usually reported in the literature for high-rate anaerobic reactors. For instance, Aquino and Stuckey (2002) and Aquino et al. (2006) found that the average values of COD p /COD inf for submerged anaerobic membrane bioreactors fed on dilute synthetic wastewater (COD mg.l 21 ) were from 12 to 40% of the influent COD. Unfortunately, there is no information on SMP production in UASB reactors for the sake of comparison. The COD p /COD inf values reported here are somewhat higher than those reported by Aquino and Stuckey (2002) for anaerobic chemostats fed on simple substrates (e.g., glucose) run at higher retention times (HRT 5 15 days), reinforcing the fact that the more complex the influent and the higher the biomass accumulation within the system, the higher the SMP accumulation seems to be. The ratio COD p.q/vss is an estimate of the amount of SMP produced per biomass in the reactor and could be loosely compared to a biomass decay rate that leads to the production of COD p. Figure 4(B) shows that more COD p seemed to have accumulated in both reactors when the HRT was the lowest (5 hours) and biomass washout was the highest due to the high upflow velocity (1.1 m.h 21 ) in the digestion compartment. During phase 2, the amount of biomass in reactors R1 and R2 was estimated as 94.5 and 74.6 kg VSS, respectively. Changing the HRT to 5 hours resulted in a loss of approximately 18% of biomass in both reactors, so that the amount of biomass in reactors R1 and R2 during phase 3 averaged 77.7 and 61.1 kg VSS, respectively. These results show that although the mass of microorganisms in the UASB reactors during phase 3 was about 82% of that found during phase 2, the biomass that remained in the reactor seemed to have produced more SMP. It was also found that the biomass that coped with the washout was, in fact, more active, as evidenced by SMA tests. For instance, during phase 2 the average values of SMA varied from 0.10 to 0.15 gcod.gvss 21.d 21 in both reactors, while during phase 3 this figure was higher, from 0.25 to 0.35 gcod.gvss 21.d 21 (Teixeira et al., 2005). These results suggest that SMP production increased when cell metabolism was higher, and this agrees well with mathematical models that consider utilization associated products (UAPs) more important than biomass associated products (BAPs) for most situations (Aquino and Stuckey, 2008; Barker and Stuckey, 2001). In addition, molecular biology results based on denaturing gradient Table 4 Average values of inert soluble COD of raw wastewater under anaerobic conditions. Test mg COD.L 21 Inert soluble COD (COD i ) % soluble influent COD gel electrophoresis experiments revealed that, despite the loss of biomass and the stressful conditions imposed during phase 3, the diversity of methanogenic archaea did not change when compared to phase 2 (Silva et al., 2008). These results show that there was a microbial adaptation to the stressed conditions, rather than an ecological succession, and this seemed to have contributed to SMP production in both reactors. The higher SMP production observed during phase 3 might be due to a higher production of UAPs following the higher organic loads caused by the reduction in HRT, as shown by Aquino and Stuckey (2004a). However, another important source of SMP is the production of BAPs, which have lower degradation rates (Aquino and Stuckey, 2008) and might have accumulated during phase 3 due to a higher production of EPS following higher upflow velocity. In other words, one hypothesis is that microorganisms excreted more EPS as a survival strategy to attain a state of higher aggregation that helped in coping with the high dragging force caused by the low HRT. Figure 5 sums up the influence of HRT on the accumulation of COD p. It can be seen that the reduction in the average size of particles of the raw wastewater, inputted by the forced screening unit previous to the reactor R2, had little effect on SMP production. The results also indicate that the lowest production of SMP was associated with the most favorable conditions in terms of biomass retention, that is, when the HRT was 9 hours and the upflow velocity was 0.5 m.h 21. The lowest production of COD p observed at higher HRT may also be related to the better conditions for SMP degradation. Molecular Weight Distribution of Residual Chemical Oxygen Demand. The results presented in Table 5 show that all membranes retained the standard calibration compounds (mixture of glucose and bovine serum albumine [BSA]) according to their nominal cutoff. For instance, the BSA protein (66 kda) should pass through the 100 kda membrane and be retained by the 10 kda and Figure 3 Composition of the soluble effluent COD from reactors R1 (A) and R2 (B) (COD p 2 produced COD, COD i 2 inert COD, and COD vfa 2 COD due to VFAs).

7 Figure 4 Estimation of SMP production (COD p ) in terms of influent COD (A) and the amount of biomass present in the reactor (B). 1 kda membranes, while the glucose (180 Da) should be retained by none of the membranes used. The results show that the BSA was indeed retained by the 1 kda and 10 kda membranes and that glucose was not retained. The eventual retention of glucose by the 1 kda membrane and of BSA by the 100 kda membrane might be explained by polarization effects on the membrane. The experimental error associated with ultrafiltration was estimated by filtering one sample of wastewater (previously filtered through 1.2 lm) five times by each membrane used. The coefficient of variation (CV) was found to be 4.6, 8.7, and 5.2% for the 100 kda, 10 kda, and 1 kda membranes, respectively. Figure 6 shows the molecular weight distribution of the organics present in the influent and effluent from the UASB reactors during different operational phases. It can be seen that, in all operational phases, most organics dissolved in raw wastewater had low molecular weight (x, 1 kda), although there was a significant amount (20 to 50%) of high molecular weight compounds (x. 10 kda). Such high molecular weight compounds need to be hydrolyzed to gain access to the cell and this involves the release of exoenzymes into the bulk solution, which contributes to the production of SMP, more specifically, of UAPs. In terms of UASB effluents, Figure 6 shows that the majority of the soluble organics also had low molecular weight and that the operational phases of the reactors affected molecular weight distribution. It can be seen that the higher amounts of high molecular weight compounds were found during phase 3 when the HRT was the lowest (5 hours), and that changing the HRT to 9 hours (phase 4) and 7 hours (phase 5) resulted in the reduction of organics with molecular weight greater than 100 kda in the effluent. These results suggest that a better degree of hydrolysis is accomplished with HRTs higher than 5 hours, and that the higher accumulation of high molecular weight organics during phase 3 might be related to the higher production of SMP observed during that phase. Such high molecular weight organics might be cell-eps released in the bulk solution to play a role in aggregation of biomass, as previously hypothesized. Ultrafiltration results also showed that the molecular weight distribution of the effluent COD was, as expected, highly dependent on operational conditions of the reactors and independent of the forced sieving of wastewater. In addition, results showed that the operation of the reactor at an HRT of 5 hours resulted in the accumulation of colloidal organics (100 kda,3,1.2 lm) in the effluent. Such material causes COD and, as the following section shows, might not be promptly degraded in an aerobic post-treatment process. Biodegradability of Residual Chemical Oxygen Demand. Figure 7 presents data on aerobic and anaerobic biodegradability of fractions of UASB effluent during three operational phases. The results show that, for all operational phases investigated, only 10 to 50% of the residual COD was degraded anaerobically in 5 days of incubation, implying that SMPs accumulated during anaerobic treatment might be considered inert as far as conventional UASB reactors (operated with HRTs of less than 12 hours) are concerned. In its turn, the long-term anaerobic biodegradability of residual COD varied from 50 to 100%, and it can be seen that the lowest rates were obtained during phase 3 (lowest HRT) when the production of SMPs was the highest. The fact that some data on long-term anaerobic biodegradability were higher than 100% might be explained by the increased production of methane due to cell decay following the long incubation time. An overall analysis of Figure 7 shows that, as expected, aerobic biodegradability was somewhat higher when compared to the anaerobic biodegradability, especially when considering the shortterm biodegradability (measured in the fifth day of incubation) of Figure 5 Empiric relationship between SMP accumulation (COD p ) and HRT in the UASB reactors.

8 low molecular weight compounds (3,1kDa). The results confirm that aerobic treatment is indeed a good option to remove residual COD from anaerobic reactors because most organics present have low molecular weight. This is because the 5-day aerobic biodegradability varied only from 10 to 40% for high molecular weight SMPs, implying a limitation of aerobic treatment in hydrolyzing and degrading high molecular weight compounds in short incubation times. However, increasing the incubation time to 20 days resulted in complete aerobic degradation of the high molecular weight SMPs produced in UASB reactors. It can also be seen that there were no significant differences between the anaerobic degradability of low and high molecular weight compounds, and this agrees with the results of Barker et al. (1999). In terms of aerobic biodegradability, the results were somewhat different; Figure 7 shows that, at short-term, low molecular weight fractions were more easily degraded aerobically than high molecular weight fractions. Practical Implications of this Research. The composition of UASB effluents is summarized in Figure 8, which shows that the soluble COD corresponded from 46 to 66% of the total effluent COD for reactor R1 and from 44 to 62% for reactor R2 during all operational phases of the research. During nonstressed conditions (HRT 5 7 hours and 9 hours), the soluble COD effluent from the UASB reactors averaged 102 mg.l 21 for reactor R1 and 94 mg.l 21 for reactor R2, which corresponds from 18 to 22% of the total influent COD. This filtered COD is still high and, along Table 5 Calibration of ultrafiltration membranes using a mixture of glucose and BSA solution. Membrane cutoff 1 st calibration (1 st ultrafiltration) % of retention 2 nd calibration (7 th ultrafiltration) BSA Glucose BSA Glucose 100 kda 5.5, LD 1, LD, LD 10 kda 98.2, LD 85.2, LD 1 kda LD 5 Limit of detection (;10 mg.l 21 for glucose and ;5 mg.l 21 for BSA) 1 In this case, glucose and BSA were below LD; the % of retention is close to zero. with particulate COD, contributes to the noncompliance of UASB effluents in most environmental discharge limits. The improvement of UASB treatment efficiencies would contribute to the widespread use of anaerobic biotechnology, and this is closely related to the minimization of SMP and VFA accumulation in anaerobic systems. As shown in Figure 3, the soluble effluent COD was comprised of an inert fraction (40 to 50% of the soluble effluent COD) of Figure 6 Molecular weight distribution of organics (% COD mass) in the influent and effluent from reactors R1 and R2 during phases 3 (A, B), 4 (C, D), and 5 (A, B).

9 Figure 7 Anaerobic and aerobic biodegradability of ultrafiltration fractions of effluent (Eff) from reactors R1 and R2 during phases 3 (A, B), 4 (C, D), and 5 (E, F). VFAs (3 to 20% of soluble effluent COD) and SMPs (30 to 62% of the soluble effluent COD). Therefore, the reduction of soluble effluent COD might be sought by creating more favorable conditions in the reactor that allows for the removal of residual VFAs (which represented from 5 to 20 mg COD/L) or that allows for the minimization of SMP accumulation (which represented from 25 to 65 mg COD/L). The minimization of VFA accumulation can be done by ensuring the presence of nutrients, by eliminating any toxic compounds in the influent, and by observing kinetic and thermodynamic limitations of anaerobic microorganisms. In turn, the minimization of SMP production can be accomplished by optimizing the sludge age in the reactors and by minimizing stress conditions upon the biomass, which are related to the lack of nutrients and presence of toxic compounds (Aquino and Stuckey, 2003 and 2004b) as well as to high hydraulic and organic loads applied to the reactors. Along with reducing the production of filtered COD, it is also important to invest in improvement of UASB solid-liquid separators to reduce the amount of particulate COD which, as seen before, represented from 34 to 56% of the effluent COD from the UASB reactors. Figure 8 Composition of COD effluent from reactors R1 (A) and R2 (B) during the operational phases of the research.

10 Conclusions Results obtained in this study show that approximately 10% of the filtered COD of the raw wastewater could be considered inert to anaerobic metabolism. There was no significant accumulation of VFAs in the UASB reactors during the five operational phases studied, and filtered effluent COD was mainly comprised of inert COD (38 to 50%) as well as COD apparently produced by biomass (62 to 50%). The accumulation of SMP was not significantly influenced by the forced screening of raw wastewater, but it was affected by HRT. The largest ratios of COD p /COD inf (;0.15 for reactors R1 and R2) and COD p.q/vss (0.08 for reactor R1 and 0.1 for reactor R2) were observed when HRT was the lowest (5 hours). This might have been due to the higher organic loading rate and the stressful conditions caused by the higher upflow velocity in the digesting compartment. Although the molecular weight distribution of residual COD seemed to be dependent on the operational condition of the reactors, most organics in the effluent had low molecular weight (3,1 kda) and were amenable to both aerobic and anaerobic degradability during 20 days of incubation. The biodegradability tests showed that the high and low molecular weight organics that make up residual COD were difficult to degrade anaerobically at 5 days incubation, and that the low molecular weight compounds were promptly degraded at 5 days incubation while high molecular weight compounds could only be aerobically degraded extensively at 20 days incubation. Acknowledgments The authors are grateful for the financial support they received from the following Brazilian institutions: Minas Gerais State Research Foundation (FAPEMIG) (grant TEC-1051/2004), National Council for Scientific and Technological Development (CNPq) (for the post-doctoral scholarships), and Research and Projects Financing (FINEP) (PROSAB program). Submitted for publication March 28, 2008; revised manuscript submitted November 14, 2008; accepted for publication November 19, References Andrews, S. C.; Robinson, A. K.; Rodrigues-Quinones, F. (2003) Bacterial Iron Homeostasis. FEMS Microbiol. Rev., 27, American Public Health Association; American Water Works Association; Water Environment Federation (1998) Standard Methods for the Examination of Water and Wastewater, 19th ed.; Washington, D.C. Aquino, S. F.; Stuckey, D. C. (2002) Characterization of Soluble Microbial Products in Effluents from Anaerobic Reactors. Water Sci. Technol., 45 (10), Aquino, S. F.; Stuckey, D. C. (2003) Production of Soluble Microbial Products in Anaerobic Chemostats under Nutrient Deficiency. J. Environ. Eng. (ASCE), 129 (11), Aquino, S. F.; Stuckey, D. C. (2004a) The Effect of Organic and Hydraulic Shock Loads on the Production of Soluble Microbial Products (SMP) in Anaerobic Digesters. Water Environ. Res., 76 (7), Aquino, S. F.; Stuckey, D. C. (2004b) Soluble Microbial Products Formation in Anaerobic Chemostats in the Presence of Toxic Compounds. Water Res., 38 (2), Aquino, S. F.; Hu, A. Y.; Aurangzeb, A.; Stuckey, D. C. (2006) Characterization of Dissolved Compounds in Submerged Anaerobic Membrane Bioreactors (SAMBRs). J. Chem. Technol. Biotechnol., 81, Aquino, S. F.; Stuckey, D. C. (2008) Integrated Model of the Production of Soluble Microbial Products (SMP) and Extracellular Polymeric Substances (EPS) in Anaerobic Chemostats during Transient Conditions. Biochem. Eng. J., 38, Babuna, F. G.; Ince, O.; Orhon, D.; Simsek, A. (1998) Assessment of Inert COD in Pulp and Paper Mill Wastewater under Anaerobic Conditions. Water Res., 32 (11), Barker, D. J.; Stuckey, D. C. (1999) A Review of Soluble Microbial Products (SMP) in Wastewater Treatment Systems. Water Res., 33 (14), Barker, D. J.; Stuckey, D. C. (2001) Modelling of Soluble Microbial Products in Anaerobic Digestion: The Effect of Feed Strength and Composition. Water Environ. Res., 73 (2), Barker, D. J.; Mannucchi, G. A.; Salvi, S. M. L.; Stuckey, D. C. (1999) Characterisation of Soluble Residual Chemical Oxygen Demand (COD) in Anaerobic Wastewater Treatment Effluents. Water Res., 33 (11), Boero, V. J.; Bowers, A. R.; Eckenfelder, W. W. J. (1996) Molecular Weight Distribution of Soluble Microbial Products in Biological Systems. Water Sci. Technol., 34 (5-6), Fang, H. H. P.; Jia, X. S. (1998) Soluble Microbial Products (SMP) of Acetotrophic Methanogenesis. Bioresour. Technol., 66, Germirli, F.; Orhon, D.; Artan, N. (1991) Assessment of the Initial Inert COD in Industrial Wastewaters. Water Sci. Technol., 23 (Kyoto), Hejzlar, J.; Chudoba, J. (1986) Microbial Polymers in the Aquatic Environment II - Isolation from Biologically Non-Purified and Purified Municipal Waste Water and Analysis. Water Res., 20 (10), Kuo, W. C.; Sneve, M. A.; Parkin, G. P. (1996) Formation of Soluble Microbial Products during Anaerobic Treatment. Water Environ. Res., 68, Laspidou, C. S.; Rittmann, B. E. (2002) A Unified Theory for Extracellular Polymeric Substances, Soluble Microbial Products, and Active and Inert Biomass. Water Res., 36, Nies, D. H. (1999) Microbial Heavy-Metal Resistance. Appl. Microbiol. Biotechnol., 51, Noguera, D. R.; Araki, N.; Rittmann, B. E. (1994) Soluble Microbial Products (SMP) in Anaerobic Chemostats. Biotechnol. Bioeng., 44, Orhon, D.; Karahan, O.; Sozen, S. (1999) The Effect of Residual Microbial Products on the Experimental Assessment of the Particulate Inert COD in Wastewaters. Water Res., 33 (14), Pontes, P. P.; Chernicharo, C. A. L.; Porto, M. T. R. (2002) Volatile Fatty Acids Profile along the Height of an Upflow Anaerobic Sludge Bed (UASB) Reactor Treating Sewage under Different Hydraulic Retention Time. Proceedings of the XXVIII Interamerican Congress of Sanitary and Environmental Engineering (in Portuguese), Cancún, Mexico; pp 1 7. Rittmann, B. E.; McCarty, P. L. (2001) Environmental Biotechnology: Principles and Applications; McGraw-Hill: New York. Silva, S. Q.; Abreu, E. F.; Zerbini, A. M.; Chernicharo, C. A. L. (2008) Stability of Microbial Diversity in UASB Reactors under Different Operational Conditions. In Proceedings of IX Latin American Symposium on Anaerobic Digestion, Isla de Pascua, Chile. Souza, C. L.; Silva, S.; Aquino, S. F.; Chernicharo, C. A. L. (2006) Production and Characterization of Scum and its Role in Odour Control in UASB Reactors Treating Domestic Wastewater. In World Water Congress, Beijing [CD-ROM]; World Water Congress, Beijing, China, September 2006; pp Teixeira, A. R.; Aquino, S. F.; Chernicharo, C. A. L. (2005) Influence of the Forced Sieving of Raw Sewage on the Performance of UASB Treating Domestic Sewage. In Proceedings of the VIII Latin American Symposium on Anaerobic Digestion [CD-ROM], Punta del Est, Uruguai; pp