MASS TRANSFER IMPROVEMENT OF A FIXED-BED ANAEROBIC SEQUENCING BATCH REACTOR WITH LIQUID-PHASE CIRCULATION

Ana C.T. Ramos, Suzana M. Ratusznei, José A.D. Rodrigues and Marcelo Zaiat

Ana C.T. Ramos. Student of Chemical Engineering, Escola de Engenharia Mauá, Instituto Mauá de Tecnologia (EEM-IMT), Brazil. Departamento de Engenharia Química e de Alimentos (DEQ-EEM-IMT). Address: Praça Mauá 1, CEP 09.580-900, São Caetano do Sul-SP, Brazil.

Suzana M. Ratusznei. Doctor in Chemical Engineering, Universidade Federal de São Carlos (UFSCar), Brazil. Professor and Researcher, DQA-EEM-IMT. Address: Praça Mauá 1, CEP 09.580-900, São Caetano do Sul-SP, Brazil.

José A.D. Rodrigues. Doctor in Chemical Engineering, Universidade Estadual de Campinas (UNICAMP), Brazil. Professor and Laboratory Coordinator, Laboratory of Biochemical Engineering, DQA-EEM-IMT. Address: Praça Mauá 1, CEP 09.580-900, São Caetano do Sul-SP, Brazil. e-mail: rodrigues@maua.br

Marcelo Zaiat. Doctor in Hydraulics and Sanitation, Universidade de São Paulo (USP), Brazil. Professor and Vice-Coordinator of Graduate Studies, Department of Hydraulics and Sanitation, Escola de Engenharia de São Carlos, Universidade de São Paulo (SHS-EESC-USP). Address: Av. Trabalhador São-Carlense 400, CEP 13.566-590, São Carlos-SP, Brazil.

Resumen

Un reactor anaerobio de lotes secuenciales (ASBR) y rellenado con biomasa inmovilizada sobre espuma de poliuretano fue estudiado para tratamiento de agua residual de baja carga. El reactor fue operado con circulación de la fase líquida con la finalidad de minimizar la resistencia a la transferencia de materia, típica en sistemas con células inmovilizadas. El reactor fue operado con temperatura controlada en 30ºC y lotes de 8 horas fueron utilizados para tratar un agua residual con demanda química de oxígeno (DQO) de 500mg/L. El sistema alcanzó eficiencia de remoción de materia orgánica de 72% cuando la recirculación no fue utilizada y de 87% cuando la recirculación de la fase líquida fue empleada, o sea, el desempeño del reactor aumentó cuando la recirculación fue implementada. Además, el modelo cinético de primer orden con una concentración residual de materia orgánica fue ajustado a los datos experimentales para evaluar la influencia de la recirculación del líquido sobre el rendimiento del reactor. El parámetro cinético aparente de primer orden aumentó de 1,19 hasta 2,00h-1 cuando la velocidad superficial del líquido fue aumentada de 0,032 para 0,191cm/s, hasta alcanzar un valor estable (1,90h-1) para valores mayores de velocidad superficial (de 0,191 hasta 0,467cm/s).

Summary

An anaerobic sequencing batch reactor (ASBR) containing biomass immobilized in polyurethane foam was used to treat low strength wastewater. The reactor was operated with liquid-phase external circulation in order to minimize the mass transfer limitations, typical of systems containing immobilized biomass. The reactor was operated at 30ºC and an 8-hour cycle was used to treat a synthetic wastewater with chemical oxygen demand (COD) of 500mg/L. The system achieved overall substrate removal efficiency of 72% with no circulation and 87% when circulation was implemented, i.e., the reactor performance increased as mixing was implemented. Furthermore, a first-order kinetic model with residual substrate concentration was fitted to the experimental values to evaluate the influence of liquid circulation on the reactor performance during a cycle time. The apparent first-order parameter was seen to increase (from 1.19 to 2.00h-1) with increasing superficial velocity (from 0.032 to 0.191cm/s) up to a stable value (1.90h-1) for higher values of superficial velocity (from 0.191 to 0.467cm/s).

Resumo

Um reator anaeróbio operado em batelada seqüencial (ASBR) contendo biomassa imobilizada em espuma de poliuretano foi utilizado para o tratamento de água residuária de baixa carga. O reator foi operado com recirculação externa da fase líquida visando minimizar as limitações à transferência de massa, típicas dos sistemas contendo biomassa imobilizada. O reator foi operado a 30ºC utilizando ciclos de 8 horas para tratar água residuária sintética com demanda química de oxigênio (DQO) de 500mg/L. O sistema atingiu eficiência global de remoção de substrato de 72% na condição sem recirculação e 87% quando a circulação foi implementada, i.e., o desempenho do reator aumentou com a implementação da recirculação. Além disso, um modelo cinético de primeira ordem com uma concentração residual de substrato foi ajustado aos valores experimentais para avaliar a influência da recirculação do líquido no desempenho do reator durante o ciclo. O parâmetro aparente de primeira ordem aumentou (de 1,19 a 2,00h-1) com o aumento da velocidade superficial (de 0,032 a 0,191cm/s) até um valor estável (1,90h-1) para valores maiores de velocidade superficial (de 0,191 a 0,467cm/s).

KEYWORDS / Anaerobic Reactor / Immobilized Biomass / Mass Transfer / Wastewater /

Received: 10/25/2002. Modified: 03/19/2003. Accepted: 03/25/2003

Introduction

The Anaerobic Sequencing Batch Reactor (ASBR) appeared as an interesting alternative to similar continuous systems such as anaerobic filters, mainly due to simpler and low-cost operation, better retention of solids, better process control and the possibility to be used in processes producing intermittent effluents and to control effluent quality since withdrawal is only performed when the desired effluent levels are attained (Sung and Dague, 1995). Some works focused on treating low-strength wastewater and on operating aspects such as the effects of temperature and hydraulic batch time (Ndon and Dague, 1997; Dague et al., 1998).

A typical cycle of an ASBR consists of four steps: feeding, reaction, settling and withdrawal (Dague et al., 1992; Sung and Dague, 1995; Zhang et al., 1996). During the reaction step mixing is typically accomplished through biogas circulation (Dague et al., 1998), liquid-phase circulation (Brito et al., 1997; Camargo et al., 2001) or mechanical agitation (Ratusznei et al., 2001). Upon termination of the reaction, mixing ceases, leading to biomass settling.

Duration of the settling step may be the rate-limiting step of the cycle time as it is directly related to the formation of biomass granules. Immobilizing biomass on inert supports in ASBRs is useful for providing biomass immobilization under adverse operational conditions. Utilization of inert supports results in higher solids retention, avoiding washout of solids and enabling elimination of the settling step (Ratusznei et al., 2000). In this way utilization of inert supports make the ASBR technology even more attractive, as both the uncertainty regarding granulation and the sedimentation step are eliminated. Processes in which biomass immobilization is used are known to present limitations due to mass transfer resistances, especially when treating low strength wastewater in which biogas production is not vigorous enough (Brito et al., 1997; Zaiat et al., 2001). However, implementing liquid-phase circulation may improve the mass transfer fluxes, thus resulting in high substrate removal (Camargo et al., 2001).

This paper reports on the influence of the liquid superficial velocity on the mass transfer improvement in a fixed bed anaerobic sequencing batch reactor treating low strength wastewater with external liquid-phase circulation.

Materials and Methods

The anaerobic sequencing batch reactor with external circulation of the liquid phase is shown in Figure 1. It consisted of a cylindrical flask with an external diameter of 60mm, height of 460mm and wall thickness of 3.5mm, thus resulting in a total volume of approximately 1.0L. The circulation system, which was not included in the non-mixing assays, was composed of a 200-mm-high reservoir with external diameter of 40mm and wall thickness of 3.3mm, with a volume of about 0.2L, and an adjustable flow pump. Thus, the total reactor volume was 1.2L.

The biomass immobilization support was placed into the cylindrical flask between punctured stainless steel disks, two at the extremities of the flask and three others dividing the reactor cross-sectionally in four parts (stages), supported by stainless steel tripods. These disks were used to avoid bed compacting, mainly caused by high flow rates (Camargo et al., 2001).

Eight-hour cycles were carried out at a temperature of 30 ±1ºC. An approximate volume of 0.5L of fresh synthetic wastewater was fed for 4 min at the beginning of the cycle and the same volume of the treated medium was discharged in 7 min at the end of the cycle. The time required for these steps was estimated so as to avoid high liquid superficial velocities during the feeding and especially during withdrawal and hence to prevent solid washout from the support.

The fixed-bed consisted of polyurethane foam cubes with 5mm sides onto which biomass was attached leading to total solids concentrations in the inoculated foam (C_TS) of 1.356 ±0.001mg ts / g foam, and total volatile solids concentrations in the inoculated foam (C_TVS) of 0.801 ±0.051mg tvs / g foam, where the unit foam stands for the foam without biomass, i.e., totally clean. The immobilization procedure was the same as that performed by Zaiat et al. (1994).

The stability and performance assays using the configuration of the reactor proposed in this work were initially evaluated without external liquid phase circulation and then with external liquid phase circulation, in order to compare the results obtained in the first condition with the other conditions. The liquid superficial velocities during circulation ranged from 0.032 to 0.467cm/s. The tests took 35 days (105 cycles) under no liquid circulation condition, 36 days (108 cycles) with superficial velocity of 0.032cm/s, and the tests with superficial velocities of 0.095, 0.191, 0.312 and 0.467cm/s lasted 20 days (60 cycles) each.

The superficial velocity was calculated by

where Q is the liquid circulation flow rate fixed at 1.0, 3.0, 6.1, 9.9 and 14.9L/h, j is the fixed-bed porosity for inoculated foam evaluated as 40% by Zaiat et al. (1997) and A is the reactor cross-sectional area calculated as 22cm².

The synthetic wastewater concentration of approximately 500mgCOD/L, considering non-filtered sample, was prepared with sucrose (35mg/L), starch (114mg/L), cellulose (34mg/L), meat extract (208mg/L), soy oil (51mg/L), NaCl (250mg/L), MgCl₂6H₂O (4.5mg/L), NaHCO₃ (200mg/L), and commercial detergent (3 drops/L). The medium was sterilized (121ºC, 15 min) in order to maintain its characteristics throughout the experimental period.

Organic matter concentration (Chemical Oxygen Demand-COD), total volatile acids concentration (TVA), bicarbonate alkalinity concentration (BA), total solids concentration (TS), total volatile solids concentration (TVS), total suspended solids concentration (TSS) and volatile suspended solids concentration (VSS) were measured by the Standard Methods for the Examination of Water and Wastewater (SMEWW, 1995) in both the influent and effluent in various cycles for all conditions. The discharged volume (V) and pH were also measured. The concentrations of methane and carbon dioxide in the biogas were evaluated through gas chromatography (Hewlett Packard model 6890) using a thermal conductivity detector, a Porapack Q column, a temperature of 35^oC, H₂ as carrier gas and a sample volume of 1mL.

The overall substrate removal efficiencies based on non-filtered (e_T) and filtered (e_S) samples of the effluent were calculated as

where CI is the non-filtered concentration of substrate in the influent, C_ET and C_ES are the non-filtered and filtered concentration of substrate in the effluent, respectively.

The amount of biomass attached on the inert support inside the reactor (X) was calculated as

where g is the ratio between the clean and the inoculated foam mass, i.e., the foam without and with biomass attached during the immobilization procedure, respectively, resulting in 0.025mg clean foam / mg inoculated foam, M is the total inoculated foam mass inside the reactor during the stability and performance assays, measured as 600g of inoculated foam, and C_TVS is the total volatile solids concentrations in the inoculated foam, measured as 0.801mg tvs / g clean foam. The parameter g is a correction factor, since the total volatile solids were measured based on the clean foam (mg tvs / g of clean foam) and the quantity of foam in the bioreactor was measured as inoculated foam (g-inoculated-foam). It is worth pointing out that the X value was assumed to be constant and equal to an average value for all the experiments, since a low variation (less than 10%) was observed.

The organic volumetric loading (Bv), the specific organic loading (Bx) and the specific organic removal loading (SOR) were calculated as

where V_reactor is the value of total reactor volume (L) and V_dfed is the average daily treated volume (L/d).

When filtered substrate, bicarbonate alkalinity and total volatile acids concentrations in the effluent presented no significant variation from one cycle to the next, measurements were made of the filtered substrate concentration profile in the reactor during a cycle time. This procedure was performed in order to estimate the influence of the liquid-phase circulation on dynamic organic matter uptake rate during a cycle and, in this way, quantify the effects of this flow rate (or liquid superficial velocity) on the external mass transfer phenomena. Thus, it is possible to evaluate the enhanced mass transfer rate caused by high hydraulic mixing that improves the wastewater-sludge contact.

The material balance of this reactor is presented in Equation (8), where C_S is the filtered substrate concentration in the reactor during a cycle time (t) and R_S is the substrate uptake rate considering filtered samples. The kinetic model is

where C_SR is the final value of C_S when the value of R_S is zero and k₁ is the apparent first order kinetic parameter embodying the intrinsic kinetic constant as well as the internal and external mass transfer resistances.

Analysis of the influence of superficial velocity on the reactor performance was done by a non-linear fitting of the integrated first-order kinetic model (Eqs. 8 and 9) to the temporal profiles of filtered substrate concentration in the reactor (C_S) through the Levenberg-Marquardt numerical procedure (Microcal Origin 6.0^®)

where C_S0 is the initial value of C_S. Non-linear fitting was performed by the Integral method, a very reliable method where fitting is done without any data manipulation.

Results and Discussion

The mean values of the monitored parameters in the influent and effluent for all the liquid superficial velocities are shown in Tables I and II. The treated wastewater volume for the system without liquid circulation was lower than that treated in the other systems, since the circulation reservoir was not used.

The proposed configuration of the reactor and the immobilization method performed were both feasible, since no solids washout was observed and high organic matter removal efficiencies were attained. This fact is especially important when low strength wastewater is being treated. Thus, the reactor design guaranteed bioconversion with good solids retention.

The results shown in Table I indicate a significant improvement of the non-filtered organic matter removal efficiency in the reactor from 72% to 87% when the liquid-phase circulation was implemented. This improved performance is probably due to the increase in mass transfer resulting from the enhanced mixing caused by the recirculation flow rate. Moreover, similarity of the non-filtered and filtered substrate concentration values clearly reveal that there was no solid washout during the experimental period. This fact should be attributed to the excellent conditions offered by the polyurethane foam matrices, providing high solids retention in the system since the biomass can strongly adhere to the support matrices (Varesche et al., 1997).

Stability of the system for both conditions without and with liquid circulation can be confirmed through analysis of Tables I and II, showing low standard error of the organic matter removal efficiencies, low volatile acids concentration values and generation of bicarbonate alkalinity considering the affluent and influent samples.

Some design parameters obtained in this study were the organic volumetric loading (Bv) applied in the reactor of 513mgCOD/L·d and the specific organic loading (Bx) of 51.3mgCOD/g tvs·d. Moreover, the specific organic removals (SOR) were 36.5mgCOD/g tvs·d and 44.3mgCOD/g tvs·d, for the conditions without and with circulation, respectively. The values of total reactor volume (V_reactor), average daily treated volume (V_dfed), amount of biomass content in the reactor (X), average non-filtered influent substrate concentration (C_I) and average non-filtered effluent substrate concentration (C_ES) without and with circulation were 1.2L, 1.28L, 12g tvs, 481mgCOD/L, 139mgCOD/L and 66mgCOD/L, respectively.

Although implementation of liquid circulation had improved the final organic matter removal efficiency, the effective influence of the liquid superficial velocity on the reactor performance cannot be seen from the data presented in Table I. Such an influence can only be adequately investigated through the dynamic filtered substrate profile during a cycle and its kinetic analysis. This way, Table III shows the adjusted parameters of the first order kinetic model with residual organic matter concentration (Eq. 10) fitted by the integral non-linear method.

Statistical analysis of adjusted k₁ values can be taken into account through the standard error shown in Table III and through Figure 2, in which the first order model and the experimental points are plotted for the conditions without circulation and with a liquid superficial velocity of 0.191cm/s.

The influence of the liquid superficial velocity on the mass transfer improvement could be evaluated through the values of the apparent first-order kinetic parameter obtained for experiments, in which the liquid circulation was implemented. Two distinct patterns could be observed: an increase in k₁ value from 1.19 to 2.00h^-1 when the superficial velocity was increased from 0.032 to 0.191cm/s and the attainment of a stable k₁ value (1.90h^-1) for higher superficial velocities (from 0.191 to 0.467cm/s). This result demonstrates that enhancement of the reactor performance could be accomplished through mass transfer improvement obtained by liquid circulation. Moreover, there is a limiting liquid superficial velocity above which the kinetic parameter remains practically constant. However, the choice of an operating superficial velocity must take into account the energy costs related to liquid recirculation (Noyola et al., 1988).

Figure 3 shows that the hyperbolic Equation (12) represented well the influence of v_S on k₁ and the empirical coefficients a and b were estimated as 2.0 ±0.1h^-1 and 0.021 ±0.006cm/s, respectively (r²= 0.896). It is worth mentioning that the inert support bed had not been subjected to compaction, as a result of the division of the inert support bed in four parts by the punctured stainless steel disks. Compaction was observed by Camargo et al. (2001) in this type of reactor operating without bed division.

At this point, it is worth pointing out that the condition without circulation resulted in a first order parameter (1.98h^-1) approximately equal to those obtained in experiments with liquid superficial velocities higher than 0.191cm/s. However, although both systems are expected to have similar reaction rates, the filtered substrate residual concentration for the system without liquid circulation (116mgCOD/L) was higher than that observed in systems with liquid circulation (52-60mgCOD/L). Figure 2 shows a comparison between profiles obtained in systems without and with liquid circulation where the kinetic parameter was similar but the residual substrate concentration was higher for the first condition.

It is also important to mention that the kinetic parameters estimated for systems with liquid circulation are related to the liquid-phase mass transfer coefficients while the parameter estimated for the static condition can be related to liquid medium diffusion processes.

Another important result obtained by the kinetic analysis was the determination of the actual batch-cycle time required to guarantee good final effluent quality in terms of organic matter concentration. This information is assessed through analysis of the overall conversion rates variation throughout a batch cycle and it is an optimization procedure with economical benefits for the overall organic matter conversion process. Such an analysis indicates that three hours are sufficient to achieve the final removal efficiency for liquid superficial velocity of 0.191cm/s (Figure 4), since the filtered substrate concentration removal rate is nearly zero for higher batch cycle times.

This result is slightly superior to the value obtained in tests carried out in a 2 L continuous horizontal-flow anaerobic immobilized biomass (HAIB) reactor using the same polyurethane foam as support and treating the same synthetic wastewater at 25^oC. A maximum COD removal efficiency of 84% was reached when a residence time of 3.3h was applied (Sarti et al., 2001).

Conclusions

Implementing external liquid-phase circulation in the sequencing batch reactor improved organic matter removal rates, since the substrate mass transfer fluxes from the bulk phase to the immobilized microorganisms were improved.

Kinetic analysis indicated that only three hours were required to attain maximum removal efficiency. Values of the apparent first-order coefficient (k₁) increased for superficial velocities from 0.032 to 0.191cm/s, while this parameter remained approximately constant for superficial velocities ranging from 0.191 to 0.467cm/s, showing that the liquid-phase mass transfer resistance was probably overcome when liquid superficial velocities higher than 0.191cm/s were applied.

ACKNOWLEDGMENTS

The authors acknowledge Walter Borzani for helpful discussions and suggestions, and Baltus C. Bonse for the revision of this paper. This study was supported by the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), projects 97/05.987-3 and 97/13.270-1 (J.A.D. Rodrigues), 98/10.303-9 (S.M. Ratusznei) and 00/02.681-5 (A.C.T. Ramos).

REFERENCES

1. Brito AG, Rodrigues AC, Melo FL (1997) Feasibility of a pulsed sequencing batch reactor with anaerobic aggregated biomass for the treatment of low strength wastewater. Water Sci. Technol 35: 193-198.        [ Links ]

2. Camargo EFM, Zaiat M, Rodrigues JAD, Ratusznei SM, Borzani W (2001) Influence of liquid superficial velocity on the performance of an anaerobic sequencing batch reactor containing immobilized biomass. In Proceedings of the 9th World Congress on Anaerobic Digestion. Part 2. Antwerpen, Belgium. pp. 89-91.        [ Links ]

3. Dague RR, Habben CE, Pidaparti SR (1992) Initial studies on the anaerobic sequencing batch reactor. Water Sci. Technol. 26: 2429-2432.        [ Links ]

4. Dague RR, Banick GC, Ellis TG (1998) Anaerobic sequencing batch reactor treatment of dilute wastewater at psychrophilic temperatures. Water Environ. Res. 70: 155-160.        [ Links ]

5. Ndon UJ, Dague RR (1997) Effects of temperature and hydraulic retention time on anaerobic sequencing batch reactor treatment of low-strength wastewater. Water Res. 31: 2455-2466.        [ Links ]

6. Noyola A, Capdeville B, Roques H (1988) Anaerobic treatment of domestic sewage with a rotating stationary fixed-film reactor. Water Res. 22: 1585-1592.        [ Links ]

7. Ratusznei SM, Rodrigues JAD, Camargo EFM, Zaiat M, Borzani W (2000) Feasibility of a stirred anaerobic sequencing batch reactor containing immobilized biomass for wastewater treatment. Bioresource Technol. 75: 127-132.        [ Links ]

8. Ratusznei SM, Rodrigues JAD, Camargo EFM, Zaiat M, Borzani W, (2001) Influence of agitation rate on the performance of a stirred anaerobic sequencing batch reactor containing immobilized biomass. Water Sci. Technol. 44: 305-312.        [ Links ]

9. Sarti A, Vieira LGT, Zaiat M, Foresti E (2001) Influence of the liquid-phase mass transfer on the performance of a packed-bed bioreactor for wastewater treatment. Bioresource Technol. 78: 231-238.        [ Links ]

10. SMEWW (1995) Standard Methods for the Examination of Water and Wastewater. 19th edition. American Public Health Association. APHA/AWWA/WEF. Washington DC. 1325 pp.        [ Links ]

11. Sung S, Dague RR (1995) Laboratory studies on the anaerobic sequencing batch reactor. Water Environ. Res. 67: 294-301.        [ Links ]

12. Varesche MB, Zaiat M, Vieira LGT, Vazoller RF, Foresti E (1997) Microbial colonization of polyurethane foam matrices in horizontal flow anaerobic immobilized sludge (HAIS) reactor. Applied Microbiology and Biotechnology 48 (4), 543-538.        [ Links ]

13. Zaiat M, Vieira LGT, Foresti E (1997) Spatial and temporal variations of monitoring performance parameters in horizontal-flow anaerobic immobilized sludge (HAIS) reactor treating synthetic substrate. Water Res. 31: 1760-1766.        [ Links ]

14. Zaiat M, Cabral AKA, Foresti E (1994) Horizontal-flow anaerobic immobilized sludge reactor for wastewater treatment: conception and performance evaluation. Revista Brasileira de Engenharia-Caderno de Engenharia Química 11: 33-42. (In Portuguese).        [ Links ]

15. Zaiat M, Rodrigues JAD, Ratusznei SM, Camargo EFM, Borzani W (2001) Anaerobic sequencing batch reactors for wastewater treatment: a developing technology. Appl. Microbiol. Biotechnol. 55: 29-35.        [ Links ]

16. Zhang R, Yin Y, Sung S, Dague RR (1996) Anaerobic treatment of swine waste by the anaerobic sequencing batch reactor. In 51^st Purdue Industrial Waste Conference Proceedings. Chelsea, USA. pp. 315-320.        [ Links ]