SciELO - Scientific Electronic Library Online

vol.17 issue4-7Mathematical modeling of polystyrene particle size distribution produced by suspension polymerizationThe inclusion of surface characteristics in equations for turbulence closure author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol.17 n.4-7 São Paulo Dec. 2000 



F.P.Puget1, M.V.Melo1 and G.Massarani2
C. P. 68502, 21945-970, Rio de Janeiro - RJ, Brazil
Phone: +55 (21) 590-2241, Fax: +55 (21) 590-7135


(Received: September 20, 1999 ; Accepted: April 6, 2000)



Abstract - This work deals with the performance analysis of a separation set-up characterized by the ejector-hydrocyclone association, applied in the treatment of a synthetic dairy wastewater effluent. The results obtained were compared with the results from a flotation column (cylindrical body of a hydrocyclone) operated both batch and continuously. As far as the experimental set-up studied in this work and the operating conditions imposed to the process, it is possible to reach a 25% decrease in the total effluent chemical oxygen demand (COD). This corresponds approximately to 60% of the COD of the material in suspension. The best results are obtained for ratios air flow rate-feed flow rate (Qair/QL) greater then 0.15 and for ratios underflow rate-overflow rate (Qu/Qo) lower than 1.0.
Keywords: flotation, ejectors, effluent treatment, dairy wastewater effluent.




The presence of almost stagnant particles in solution or a very low difference between the phases densities are limiting factors for phases separation by sedimentation. Flotation, a unit operation applied in the separation of a solid- or dispersed liquid-phase from a liquid-phase, appears in this scenario as a potential alternative to sedimentation, for it can be used in a wider range of cases [Metcalf and Eddy, 1991].

The flotation process, when applied in wastewater treatment, use methods of air injection in such a way that small air bubbles (<0.1mm) should be formed. This is necessary due to the small size, low density and usually hydrophilic properties of the solids to be removed. These factors when combined make the removal of the solids in solution a difficult or even impossible task [Richter, 1976]. The generation of small air bubbles can be performed chemically, by electrolysis, by dissolving air in water at high pressures [Chambers et al., 1976] or by ejectors [Engenho Novo, 1998].

Ejectors are well-known and accepted devices in several industrial applications. Currently, ejectors are used as pumps, mixers, heaters, coolers, as devices to generate vacuum, and also as bubble generators. Ejectors show important advantages over other devices because of its simplicity of design and building. There are no mobile parts in ejectors and its installation and maintenance are simple and cheap. Ejectors can be built from any sort of moldable material which, depending on the requirements, can be chemically resistant [Perry, 1997].

An ejector is basically constituted by three components: a nozzle, a suction chamber, and a discharge pipe, also called diffuser or Venturi. The feed flow is brought into the ejector through the nozzle, which should have a outlet diameter (DN) much smaller then the feed pipe diameter. The maximum liquid-jet velocity takes place in the straight section of the diffuser, with diameter DT, where the pressure on the liquid is lower than at the inlet. Therefore, the formation of micro-bubbles of dissolved air takes place in the straight section of the diffuser. Furthermore, the vacuum created by the high-velocity flow rate of the liquid-jet promotes the suction of air through the suction chamber, implying in the formation of bubbles bigger than the bubbles formed from the dissolved air in the effluent.

This work deals with the analysis of the performance of a experimental separation apparatus for liquid wastewater flotation. The compact experimental set-up is characterized by the association ejector-hydrocyclone. The bubble generation is performed by a gas-liquid ejector and the foam and clarified flow-streams are conducted in the centrifugal field formed in the interior of the hydrocyclone.




Ejector. The equipment used in this work was assembled at the Particle Science Laboratory at the Federal University of Rio de Janeiro, Rio de Janeiro, Brazil. The ejector dimensions and configuration were defined based on the work developed by Otake et al. (1981) who investigated the usage of ejectors

for air bubble generation. These authors studied the effect of the ejector shape and operating conditions on the porosity and average diameter of the formed bubbles. Their results showed that the average bubble diameter is a function of the ratio DN/DT and the nozzle ejector outlet fluid velocity. This conclusion, as pointed out by the authors, seems to be valid in the ranges 0.21£ DN/DT£ 1.0 and 0.1£ Qair/QL£ 1.0.

Figure 1 shows the ejector dimensions and configuration. The equipment was built of acrylic in order to permit the visualization of its interior during operation.



Hydrocyclone. The hydrocyclone was built of glass in order to permit the visualization of the operating condition at which vortex formation is promoted. The feed flow is introduced tangentially in such a way to establish a centrifugal field in the interior of the equipment. The flocks formed from the precipitation of the protein and fat from milk are collected at the upper part of the equipment (overflow) while the clarified effluent is collected at the bottom of it (underflow). Figure 2 shows the dimensions and configuration of the hydrocyclone used in this work.



Experimental Set-Up

In Figure 3 the description of the experimental set-up used in this work is presented. The apparatus is constituted of a 150L feed tank, a Dancor 2.0HP centrifugal pump, a ejector and a sight gauge column built of acrylic, a hydrocyclone to promote phases separation, a vacuum meter and four Fischer & Porter rotameters for measurement of the air flow, feed flow, overflow and underflow rates.



Wastewater Characterization

The synthetic effluent used in this work was prepared by adding milk powder in water, and stirring vigorously to form a suspension with concentration of 0.1% w/v. Table 1 provides the specifications of the milk powder (brand La Serenissima, Mastellone Hnos S.A., Argentina) used in this work.



According to the specifications it is possible to observe that the dairy wastewater studied in this work is basically constituted by milk proteins and fat. The protein and fat precipitation was performed by the slow addition of a diluted aqueous solution of chloride acid (1:10 volumetric basis) until it is reached the isoelectric point’s pH.

Figure 4 shows the plot used for determining the isoelectric point of the milk. The measurements were carried out at the Interfacial Phenomena Laboratory at the Center of Mineral Technology, Rio de Janeiro, Brazil, using a ZetaMasterS/Malvern equipment. The isoelectric point, defined as the characteristic pH at which positive and negative charges of the protein cancel, is depicted in Figure 4 by the pH at which the zeta potential curve crosses the zero zeta potential abscissa (pH » 4.6). Below this point the protein is positively charged while above it the protein is negatively charged.



Experimental Procedure

System operation. The effluent is prepared by adding milk powder in water and then the fat and proteins are precipitated (by the slow addition of chloride acid) directly in the feed tank. Just after its preparation the effluent solution is pumped to the ejector and samples from the overflow and underflow of the hydrocyclone are taken for evaluation of the separation performance. Air flow, feed flow, overflow and underflow rates are obtained directly from the rotameters.

Experimental set-up performance. The performance of the separation apparatus can be evaluated by means of the following Equation


where h is the separation efficiency evaluated in terms of the decrease of the COD. CODu and CODL are the chemical oxygen demand in the underflow and feed flow respectively. The COD was determined by the potassium dichromate method [Greenberg et al., 1992], which is based on the oxidation of organic matter by a boiling mixture of chromic acid and sulfuric acid. The reduction of potassium dichromate is measured by calorimetry using a Hach – DR 200 direct reading spectrophotometer.

It is worthwhile to point that the efficiency was evaluated from the total effluent COD (COD of the soluble material plus the COD of the material in suspension). However, the experimental set-up is designed solely for the reduction of the COD of the material in suspension. As far as the dairy wastewater effluent is concerned, the suspension material COD corresponds to approximately 40% of the total COD.

Operating Conditions

The operating conditions for the experimental runs are reported in Table 2.




The influence of the operating variables (air flow, feed flow, overflow and underflow rates) on the separation apparatus performance was evaluated from the experiments.

Figure 5 shows the influence of the ratio Qair/QL on the separation apparatus performance for a ratio Qu/Qo equal to 1.0.



It can be observed that for ratios Qair/QL > 0.15 the total effluent COD is decreased in 25%, while for ratios Qair/QL < 0.15 a separation efficiency drop is noticed. The drop in efficiency is likely to be explained either by a insufficient air flow rate for flotation or by a high feed flow rate. A high feed flow rate would be responsible for small residence times of the particles in the equipment and would impose high turbulence in the flow. For the dairy wastewater studied only 40% of the total COD corresponds to a suspension material COD, then a decrease in 25% of the total COD corresponds to a 60% decrease in the COD of the material in suspension.

Figure 6 shows the influence of the ratio Qu/Qo on the separation apparatus performance for a ratio Qair/QL equal to 0.16.



It can be verified that the separation efficiency decreases with the increase of the ratio Qu/Qo. This is an expected behavior, for at high underflow rates particles in solution with potential chances to be recovered are taken away the system. These results were compared to the results obtained by Puget (1998) who used a flotation column operated both in batch and continuous mode for the treatment of a dairy wastewater effluent with same concentration as the one in study here. Puget (1998) showed that it is possible to diminish in about 70% the total COD of the effluent in 30s of batch operation, without need for addition of any other chemical. This result corresponds to a removal of almost 100% of the solid in suspension, indicating the good efficiency of the treatment of this effluent by flotation. In continuous mode of operation, a decrease of 50% in the total COD was possible with the hydrocyclone operating with tangential feed injection and an underflow rate of 120L/h which corresponds to average residence time of 50s. Although Puget’s (1998) results present better separation efficiencies, it should be observed that Puget’s (1998) experimental apparatus presented a much smaller operational capacity (about 10 times smaller). It’s worth mentioning that Puget’s (1998) experimental set-up was operated in the following flow rate ratios ranges 0.6 < Qu/Qo < 1.7 and 0.3 < Qair/QL < 7.0, which is significantly higher then the ranges found in the present work.



As far as the experimental apparatus presented in this work is concerned and the imposed operating conditions, it is possible to reach a decrease up to 25% of the total effluent COD. This corresponds approximately to 60% of the COD of the material in suspension. Moreover, it is shown that better results are obtained for the ratios Qair/QL > 0.15 and Qu/Qo < 1.0.

The experimental set-up presented in this work showed lower efficiencies than the ones obtained by Puget (1998). However, the high capacity of operation reached in this work compensates for the loss in efficiency as far as total effluent COD.



This work is part of the project "Modelagem, Simulação e Controle de Processos", FAPERJ’s grant # E-26/150.970/99. The authors also thank CNEN and CAPES for providing scholarships, and the Water Pollution Control Laboratory of PEQ/COPPE/UFRJ for providing the COD analysis chemicals and equipment.



CODL Chemical oxygen demand in the feed [M/L3]
CODu Chemical oxygen demand in the underflow [M/L3]
DN Ejector nozzle diameter [L]
DT Diffuses straight section diameter [L]
Qair Volumetric airflow rate [L3/q ]
QL Feed flow rate [L3/q ]
Qo Overflow rate [L3/q ]
Qu Underflow rate [L3/q ]
h Separation efficiency [-]



Chambers, D. B. and Cottrell, W. R. T., Flotation: Two Fresh Ways to Treat Effluents, Chemical Engineering, Vol. 83, No. 16, p. 95 (1976).        [ Links ]

Engenho Novo Tecnologia Ltda., "Processo de Flotação no Tratamento de Caldo de Cana e Xarope de Açúcar.", Industrial Catalogue, 15 p. (1998).        [ Links ]

Metcalf, A. and Eddy, M. S., Wastewater Engineering. Treatment, Disposal and Reuse, 3a ed., New York, McGraw-Hill (1991).        [ Links ]

Greenberg, A. E., Clesceri, L. S. and Eaton, A. D., Standard Methods - For the Examination of Water and Wastewater 18a ed., Part 5000, 6, APHA-AWWA-WPCF (1992).        [ Links ]

Otake, T., Tone, S., Kuboi, R. Takahashi, Y. and Nakao, K., Dispersion of a Gas by a Liquid-Jet Ejector, International Chemical Engineering, Vol. 21, No. 1, p.72 (1981).        [ Links ]

Perry, H. R. and Green, D. W., Perry’s Chemical Engineers Handbook, 7a ed., New York, McGraw-Hill (1997).        [ Links ]

Puget, F. P, Análise do Desempenho de Hidrociclone Flotante: Um Estudo de Caso, M.Sc. Thesis, Programa de Engenharia Química, COPPE/UFRJ, Rio de Janeiro, Brasil (1998).        [ Links ]

Richter H., Flotation: A Modern Process for Wastewater Treatment, International Chemical Engineering, Vol. 16, No. 4, p. 614 (1976).        [ Links ]

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License