PHOSPHATE REMOVAL IN AN UASB REACTOR TREATING SYNTHETIC SUBSTRATE SIMULATING DOMESTIC SEWAGE

PHOSPHATE REMOVAL IN AN UASB REACTOR TREATING SYNTHETIC SUBSTRATE SIMULATING DOMESTIC SEWAGE

J.T. de Sousa ¹ , R.F. Vazoller ² and E. Foresti ²

¹ Depto. de Química, Universidade Estadual da Paraíba, Av. Floriano Peixoto, 718,

Campina Grande, PB, Brazil CEP 58100-001

² Escola de Engenharia de São Carlos, Universidade de São Paulo,

Av. Dr. Carlos Botelho, 1465, CEP 13560-250

São Carlos, SP, Brazil. Phone [phone]. Fax [phone]

(Received: June 11, 1997; Accepted:October 30, 1997)

Abstract - The removal of phosphate in an up-flow anaerobic sludge blanket (UASB) reactor treating a synthetic substrate simulating domestic sewage and its relation with the presence of FeCl₃ in the influent was investigated. Phosphate removal efficiencies were 30%, 54% and 84% for influent FeCl₃ concentrations of 90 mg.L^-1 and 200 mg.L^-1, respectively. Scanning eletronic microscopy of granular sludge samples taken from the reactor showed the presence of Methanothrix- like microorganisms as the main predominant methanogenic bacteria. Iron precipitates such as vivianite [Fe₃(PO₄)₂ . 8 H₂O] were identified in granular samples by x-ray defraction and energy dispersion x-ray (EDX) analyses.

Keywords: Phosphate removal, synthetic substrate, reactor.

INTRODUCTION

Phosphate is a very common form of phosphorus present in different wastewaters. Municipal sewage contains organic phosphates, polyphosphates and orthophosphates that frequently need to be removed in order to permit wastewater treatment plant effluents to be discharged into superficial water bodies.

According to Maqueda et al. (1994), mixed liquors of digesters present orthophosphate ions which can generate precipitates such as struvite. Mamais et al. (1994) confirmed the formation of many different precipitates of anions (HCO₃^-, S^2- and PO₄^3-) and cations (Mg²⁺, Ca²⁺, NH₄⁺) associated with the anaerobic digestion of wastes. The authors also demonstrated that the addition of FeCl₃ to anaerobic digesters produces several precipitates.

According to Snoeyink and Jenkins (1980), iron salts such as FeCl₃ and Fe₂(SO₄)₃ can react with phosphates in the biodigesters’ reducing environments, forming vivianite [Fe₃(PO₄)₂ . 8 H₂O]. Therefore, ferrous salts can be used as precipitating agents to remove phosphorus from domestic sewage in anaerobic reactors. Due to the high costs normally associated with nutrient removal in conventional aerobic systems, such a feature could make anaerobic processes more attractive for domestic sewage treatment, since organic matter and phosphorus removal would be accomplished in a single treatment unit.

This paper reports and discusses the effect of FeCl₃ addition on phosphorus removal in a bench-scale up-flow anaerobic sludge blanket (UASB) reactor treating a synthetic substrate simulating domestic sewage.

MATERIALS AND METHODS

The bench-scale treatment system composed of a 4 liter UASB reactor followed by two 4-liter aerobic sequencial batch reactors (SBR) is schematically shown in Figure 1. Such an apparatus was used for studies on anaerobic-aerobic systems treating synthetic substrate simulating domestic sewage. The results referring to the removal of Chemical Oxygen Demand (COD) and Total Suspended Solids (TSS) were previously reported by Sousa and Foresti (1995).

The system was continuously operated at 30°C for 54 weeks, and the investigation of phosphorus removal was carried out during the last 25 weeks.

The UASB reactor was inoculated with anaerobic sludge (90 g of VSS) taken from another bench-scale UASB reactor. Thereafter, it was continuously fed with a synthetic substrate containing meat extract, sucrose, starch, cellulose and edible oil in COD-based percentages of 50%, 8%, 24%, 8% and 10%, respectively. The substrate also contained mineral constituents normally found in domestic sewage at the concentrations indicated in the literature. Such a composition was proposed by Torres (1992) for simulation of domestic sewage in lab-scale experiments.

Figure 1: Diagram of the combined UASB-SBRs system.

The influent COD concentration ranged from 340 mg.L^-1 to 540 mg.L^-1, and the average volumetric organic loading rate (OLR) applied was 2.53 kg COD . m^-3 . d^-1. A hydraulic detention time (HDT) of four hours was maintained constant during the experiment.

The effect of FeCl₃ on phosphorus removal was investigated by applying three different concentrations in distinct periods. During the first period, the influent FeCl₃ concentration was 90 mg.L^-1, corresponding to its original concentration in the substrate. During the second period, the FeCl₃ concentration was increased to 200 mg.L^-1 , and during the last period the sources of FeCl₃ were removed from the substrate composition.

Influent and effluent samples were analysed at least twice a week for reactor performance monitoring. Analyses of COD, VSS, nitrogen, alkalinity and pH were carried out according to the Standard Methods for the Examination of Water and Wastewater (1980), and the determination of volatile fatty acids (VFA) followed Dillalo and Albertson (1961).

Samples of granular sludge were analysed by x-ray diffraction (XRD - Rigaku Grigaflex Equipment) and scanning electronic microscopy (SEM - Digital Scanning Microscope Zeiss DSM-960). XRD samples had been previously fixed on glass slides, dried at room temperature and prepared according to the Nation (1983) and Araujo et al. (1995) techniques.

RESULTS AND DISCUSSION

Table 1 shows the average values of influent and effluent characteristics of the UASB reactor during six months of continuous operation.

From the results in Table 1, it can be concluded that the UASB reactor performed quite well in the removal of soluble and particulate organic matter. The values of pH and alkalinity are indicative of high process stability, as discussed in a previous paper (Sousa and Foresti, 1995).

Table 2 shows the data on influent and effluent phosphorus for each concentration of influent FeCl₃. The data are also shown in Figure 2a, 2b and 2c for each condition studied.

The average values of phosphorus removal efficiency were 54% and 84% for the influent FeCl₃ of 90 mg.L^-1 and 200 mg.L^-1, respectively. The data presented in Figure 2 (a and b) show that phosphorus concentrations (P_e) in the effluent were clearly dependent on the FeCl₃ availability. Moreover, the amplitude of P_e variations decreased with an increase in FeCl₃ concentration. For a 200 mg.L^-1 concentration, the P_e was very low and constant, independent of the influent phosphorus concentration (P_i). After the cessation of FeCl₃ addition (Figure 2c), phosphorus removal efficiency decreased progressively from 66% during the first week to 45% during the second and 30% during the last seven weeks. Moreover, P_e become directly proportional to P_i.

The photomicrograph in Figure 3 shows a longitudinal section of the granular anaerobic sludge of the UASB reactor taken during the experiment when the FeCl₃ influent concentration was 200 mg.L^-1. Scanning electronic microscopy of the same sample (Figure 4) shows the presence of precipitated matter-forming crystal with morphologies typical of phosphates. Analyses by energy dispersion x-ray (EDX) confirmed the predominance of iron and phosphate in the sample (Figure 4a). X-ray diffraction (RD) analysis (Figure 4b) of the precipitate identified vivianite [Fe₃(PO₄)₂ . 8 H₂O] as the predominant crystal.

Figure 5 shows aspects of the biological composition of the granular sludge. The main bacteria morphology was identified as belonging to the genus Methanotrrix (Figure 6), a methanogenic acetoclastic bacteria very often found in granular anaerobic sludge. Some coccus morphologies were also observed.

Characteristic Influent Effluent PH 7.0 ± 0.42 7.5 ± 0.16 Alkalinity (mg.L-1) 275 ± 90 326 ± 75 COD raw (mg.L-1) 435 ± 58 52 ± 15 COD filtered (mg.L-1) 170 ± 20 25 ± 10 TSS (mg.L-1) 241 ± 65 61 ± 20 TKN-Nitrogen (mg.L-1) 63 ± 12 47 ± 12

Characteristic	Influent	Effluent
PH	7.0 ± 0.42	7.5 ± 0.16
Alkalinity (mg.L-1)	275 ± 90	326 ± 75
COD raw (mg.L-1)	435 ± 58	52 ± 15
COD filtered (mg.L-1)	170 ± 20	25 ± 10
TSS (mg.L-1)	241 ± 65	61 ± 20
TKN-Nitrogen (mg.L-1)	63 ± 12	47 ± 12

Table 1: Average values of influent and effluent characteristics of the bench-scale UASB reactor

FeCl3 (mg.L-1) Phosphorus (mg.L-1) Added Influent Effluent 90 15.1 ± 2.3 7.8 ± 1.2 200 18.5 ± 4.0 2.9 ± 0.5 No addition of FeCl3 13.4 ± 2.6 9.2 ± 2.0

FeCl3 (mg.L-1)	Phosphorus (mg.L-1)
Added	Influent	Effluent
90	15.1 ± 2.3	7.8 ± 1.2
200	18.5 ± 4.0	2.9 ± 0.5
No addition of FeCl3	13.4 ± 2.6	9.2 ± 2.0

Table 2: Infuent and effluent phosphorus concentrations for each influent FeCl ₃ concentration applied

2a. FeCl₃ of 90 mg.L^-1

2b. FeCl₃ of 200 mg.L^-1

2c. No addition of FeCl₃

Figure 2: Values of influent () and effluent () phosphorus and efficiency removal () for different FeCl₃ concentrations.

Figure 3: Phosphate precipitates (SEM - 10.000x) in granular sludge (FeCl₃ of 200 mg. L^-1).

4a.

4b.

Figure 4: (a) EDX analysis showing Fe and P predominance (b) XRD analyses confirming the presence of vivianite.

Figure 5: Biological material associated with phosphorus precipitates (SEM - 5.000x).

Figure 6: Bacterial morphotypes of Methanothrix genus observed in the sludge (SEM - 20.000x).

CONCLUSIONS

The addition of FeCl₃ to the influent of anaerobic reactors treating a synthetic substrate simulating domestic sewage significantly improved the removal of phosphorus, which attained an efficiency of 84% at an FeCl₃ influent concentration of 200 mg.L^-1. The complete cessation of FeCl₃ addition caused a significant decrease in phosphorus removal efficiency, which dropped to 30%. The main mechanism for phosphorus removal seemed to be associated with the formation of precipitates of phosphorus and iron, such as vivianite, which appeared to have been incorporated into the granular sludge structure, as observed by scanning electronic microscopy.

AKNOWLEDGMENTS

The authors would like to aknowledge the funding received from FAPESP - Fundação de Amparo à Pesquisa do Estado de São Paulo, for the thematic project "Development of Anaerobic Processes for Wastewater Treatment. They would also like to acknowledge the grant received by J. Tavares de Souza from CAPES - Fundação para Capacitação e Aperfeiçoamento de Pessoal de Ensino Superior and the micrographic analyses conducted by IFSC - Instituto de Física de São Carlos.

REFERENCES

APHA; AWWA and WPCF, Standard Methods for the Examination of Water and Wastewater. 15^th ed. Amer. Public Health Association., Amer. Water Works Association, WPCF, Washington, D.C. (1980).

Araujo, J.C.; Cuba, F.J.T.; Campos, J.R. and Vazoller, R.F., Análise de Biofilmes Bacterianos Metanogênicos por Microscopia Eletrônica de Varredura Empregando-se o Método de Secagem por HMDS (1995) in press.

Dilallo, R. and Albertson, O.E., Volatile Acids by Direct Tritration. JWPCF, 33(4): 356-365 (1961).

Mamais, D.; Pitt, P.A.; Cheng, Y.W.; Loiaconoj, J. and Jenkins, D., Determination of Ferric Chlorid Dose to Control Struvite Precipitation in Anaerobic Sludge Digesters. Water Environment Reseach, 66(7), pp. 912-918 (1994).

Maqueda, C.; Pérez, J.C. and Labrato, J., Study of Struvite Precipitation in Anaerobic Digesters. Water Research, 28 (2), pp. 411-416 (1994).

Nation, J.L., A New Method Using Hexamethyldisilazone for Preparation of Soft Insect Tissues for Scanning Electron Microscopy. Stain Technology. 58(6), pp. 347-351 (1983).

Snoeyink, V.L. and Jenkins, D., Water Chemestry, 3^rd ed., New York, John Willey & Sons, pp.243-315 (1980).

Sousa, J.T. de and Foresti, E., Domestic Sewage Treatment in a UASB-SBR Reactor System. In: Internationl Symposium on Technology Transfer, Salvador, 1995, Pre-Prints, IAWQ, pp.129-140 (1995).

Torres, P., Desempenho de um Reator Anaeróbio de Manta de Lodo (UASB) de Bancada no Tratamento de Substrato Sintético Simulando Esgoto Sanitário. Master’s thesis, Escola de Engenharia de São Carlos, Universidade de São Paulo (1992).

Publication Dates

• Publication in this collection
  06 Oct 1998
• Date of issue
  Dec 1997

History

• Accepted
  30 Oct 1997
• Received
  11 June 1997