Development of a bench scale system for anaerobic acidogenic digestion of wastewater from isolated soy protein

A.S. Cassini, I.C. Tessaro and L.D.F. Marczak

Chemical Engng. Department, UFRGS, Porto Alegre, RS 90040-040, Brazil.
alinesc@enq.ufrgs.br; isabel@enq.ufrgs.br; Ligia@enq.ufrgs.br

Abstract - Wastewater from isolated soy protein (ISP) production is characterized by high organic load. A complex primary wastewater treatment system has been studied: an anaerobic acidogenic reactor, a tubular reactor and a sedimentation tank. Anaerobic digestion is a very complex process; the control of temperature, pH, alkalinity, TSS/VSS and COD is fundamental. The objective was to develop and compare the efficiency of a bench scale anaerobic acidogenic reactor, followed by the precipitation, coagulation and sedimentation steps, with an existing industrial ISP wastewater treatment system. The results obtained with bench system (reaction time of 6 h, 48 °C) were very close to the industrial ones: pH reduction of 4.5 to 3.7 (indicating the protein hydrolysis), high VFA accumulation (1,300 and 2,650 mg.L^-1 minimum and maximum concentration), no methane production and increase in TSS/VSS. This study gives an insight of the industrial primary wastewater treatment system and can be useful in future researches.

Keywords - Acidogenisis. Anaerobic Digestion. Isolated Soy Protein. Wastewater Treatment.

I. INTRODUCTION

Indispensable nowadays, a wastewater treatment system is able to reduce, drastically, the damages caused by industrial wastewaters. In some cases, the efficiency of this kind of treatment is so great that the wastewater, previously discharged, can be used as a water source in the industrial processes.

A conventional system treating a high organic load wastewater generally involves a primary and a secondary treatment system. In the primary system, a portion of the suspended solids and organic load is removed from the wastewater by, basically, physical and chemical operations. The wastewater from the primary system still contains considerable organic loads and a relatively high chemical oxygen demand (COD). As a consequence, this system cannot be used as the sole treatment method, except in particular situations. The main function of this system is to serve as a precursor to the secondary treatment system (Tchobanoglous et al., 1991).

The wastewater from ISP production is characterized by very high organic load, composed mainly of soluble proteins and carbohydrates, and has COD values greater than 16,000 mg·L^-1 with high nitrogen concentrations. Consequently, a very efficient wastewater treatment system is required to comply with the specific regulatory limits imposed by legislation.

Conventional biological treatment processes, such as aerobic and anaerobic digestions, have been used for a long time to treat this kind of effluent; these systems, however, demand significant physical space, due to the great volumes of the unit operations involved. Moreover, the effectiveness of these processes has become limited over the last two decades (Mohamammadi and Esmaeelifar, 2004).

The researches challenge is, therefore, to find more effective clean methods to remove the organic matter from the wastewater or, at least, to decrease the wastewater pollutant material concentration in order to pos-treat it, successfully, with conventional systems.

The bench scale systems are fundamental to understand and to get important data about the behavior of different wastewater treatment systems. Within this context, the objective of this work is to develop a bench scale primary treatment system, firstly, to treat the wastewater from an ISP industry and to observe each stage behavior and, second, to compare this bench scale system to the industrial one. The industrial system comprises a primary anaerobic acidogenic reactor (AAR), a tubular reactor and a primary sedimentation tank and this comparison will permit the study of different pre-treatment methods to optimize the industrial treatment system.

A great variety of studies simulating some unit operations of industrial wastewater treatment systems can be found in the literature: Bayard et al. (2008) developed laboratory-scale bioreactors (designed to simulate the anaerobic condition of sanitary landfill, with incubation time of 400 days to achieve the biodegradation) to study the behavior of untreated and pre-treated residues; Biswas et al. (2006) used a 10 dm³ anaerobic batch digester (equipped with a mechanical agitator) under controlled environment (pH 6.8, temperature = 40 °C) to study the biogas generation kinetics using municipal waste as feed stock; Borja et al. (2005) studied the effect of hydraulic retention time (HRT, 10 to 50 days) on the anaerobic acidogenic fermentation of two-phase olive pomace at laboratory-scale and mesophilic temperature (35 °C); Calli et al. (2005) studied the behavior of UASB reactors inoculated with different seed sludge, operated for 450 days under high ammonia concentrations, to investigate inhibition effects; Demirer and Chen (2005) investigated possible exploitation of the advantages of two-phase anaerobic digestion for unscreened dairy manure with hydraulic retention time of 2 (acidogenic phase) and 8 (methanogenic phase) days; Han et al. (2005) studied the performance of an upflow anaerobic sludge blanket (UASB) reactor treating wastewater with volatile fatty acids and alcohol generated during hydrogen fermentation of food; Shin et al. (2004) evaluated the hydrogen production from food waste by the mesophilic and thermophilic acidogenic culture acclimated with food waste at 5 days HRT for the effect of pH and volatile solid concentrations. Amongst these unit operations, aerobic and anaerobic (acidogenic and metanogenic) reactors are commonly found; these reactors, however, are stages of the aerobic and/or anaerobic secondary treatment systems being operated continuously and for long HRT. Studies simulating a primary AAR with a very short HRT had not been found in the available literature.

II. METHODS

A. The Industrial Primary System

As stated earlier, the studied industrial primary system involves three distinct stages: an anaerobic acidogenic reactor, a tubular reactor and a primary sedimentation tank. Figure 1 presents a schematic diagram of this system.

Figure 1. Schematic diagram of the industrial ISP primary wastewater treatment system.

As can be seen in Fig. 1, the first stage is composed by two anaerobic acidogenic reactors (AAR); these reactors operate in series and have the objective of turning insoluble the wastewater soluble proteins by the action of acidogenic bacteria which hydrolyses the wastewater complex biopolymers (such as carbohydrates and proteins). These microorganisms degrade carbohydrates and proteins into H₂, CO₂ and volatile fatty acids (VFA), mainly acetic, propionic and butyric acids. Since the apparent hydrolysis rate of the carbohydrates is higher than the rate of the proteins, these compounds are quickly converted into soluble sugars (Beal, 1995 and Campos, 1999), and the proteins become, practically, the unique insoluble compounds in the wastewater. The system pH, around 4.5, also favors the carbohydrates hydrolysis: as stated earlier by Hanaki and Matsuo (1986), pH values between 4.5 and 7.0 favors the carbohydrates hydrolysis, while the proteins hydrolysis is favored in pH values between 6.0 and 7.0.

The digestion in AAR occurs between 40 and 45 °C, under constant stirring, and during the process the pH value decreases until 3.7. The volume of each one of these reactors is 600 m³ and the HRT is 6 h.

After the AAR, the effluent goes through a tubular reactor (TR) that promotes the precipitation and coagulation of the proteins to facilitate their posterior removal by sedimentation. In this part of the process, three different chemical products are added: NaOH, to adjust the wastewater pH around 4.2 (the isoeletric point of the proteins) and FeCl₃ and an anionic polymer to coagulate the precipitated proteins. The next stage, a primary sedimentation tank (ST), removes the insoluble, precipitated and coagulated particles.

In Fig. 1 it is also presented the location where the three industrial wastewater samples are collected - raw wastewater (WWR), effluent that leaves AAR (WWIA) and the clarified wastewater from the primary system (WWIC); the place where is collected the sludge for bench experiments are also presented (from inside AAR, WWS).

B. Wastewater characterization

To determine the main characteristics of the studied industrial wastewater and to observe the industrial primary system behavior, analyses of pH, COD, protein content (PROT), total suspended solids (TSS) and volatile suspended solids (VSS) were carried out periodically with the three samples collected, WWR, WWIA and WWIC. This monitoring procedure lasted about three weeks, and all the analysis were in accordance with Standard Methods (Eaton et al. 1995).

C. Bench scale primary system

In the bench experiments, a bench scale AAR was developed; the photography of the system can be seen in Fig. 2.

Figure 2. Photography of the bench scale AAR system.

As presented in Fig. 2, the bench scale plant comprises a glass AAR (2 L) (I), immersed in a thermostatic bath (III), a magnetic stir plate (II), and a volumetric gas meter system (IV).

Firstly, two kinds of wastewater samples were collected: 1) the raw wastewater (WWR) and 2) the wastewater from the industrial AAR (WWS). The WWS was centrifuged in a bench centrifugal machine (10,000 rpm, during 6 min), in order to obtain the sludge (microbiologic biomass) used in the experiments. In the meantime, 1,800 mL of WWR was added in the AAR and both, the stirring and the temperature controller, were turned on. Once the desired temperature (48 °C), has been reached, 180 mL of centrifuged WWS were added to the AAR (inoculums). After collecting an initial sample, for pH and VFA control, the AAR was sealed and the experiment started.

In all experiments the reaction time was around 6 h, and the gas production was constantly monitored by volume reading directly from the gas meter apparatus. Moreover, at regular intervals (60 min), samples were collected from the AAR in order to measure the wastewater pH and VFA production.

After the anaerobic digestion, the same chemical products used in the industrial system were added to promote protein coagulation and flocculation. The tubular reactor mixing was simulated experimentally using two mixing velocities: fast at the beginning and slow at the end; a sedimentation time of 1 h was adopted.

After each experiment, analyses of COD, PROT, TSS and VSS were carried out with the raw wastewater (WWR), the effluent from the bench AAR (WWBA) and the bench clarified effluent (WWBC).

III. RESULTS

A. Wastewater characterization

The studied industrial wastewater characteristics (WWR, WWIA and WWIC) are presented in Table 1. This table shows the average values obtained during the monitored period.

Table 1. WWR, WWIA and WWIC characteristics.

As can be seen in Table 1, the COD, protein and TSS/VSS removal by the industrial primary system are around 25 %, 56 % and 85 %, respectively. These results are in accordance with the main objective of this primary system: the protein content reduction. This reduction, however, does not occur just by the insolubilization/precipitation/coagulation processes, but also due to protein degradation by the acidogenic bacteria (which consume nitrogen to their growth and development). Thus, the protein removal - and, consequently, the COD removal - occurs due to both, the sedimentation of the insoluble and precipitated protein, and the sedimentation and removal of the microbial biomass, which consumed part of the wastewater protein.

It can also be observed in Table 1 that the industrial wastewater COD and protein content increase during the anaerobic digestion. This behavior can be explained since a part of the activated sludge (acidogenic bacteria) flows together with the wastewater that leaves the AAR.

In the same way, the industrial wastewater TSS/VSS content also increases significantly during anaerobic digestion; since the objective of this stage is to turn insoluble the soluble protein, this result has already been expected.

The characteristics of the WWR samples presented wide range values and this behavior could be explained due to the parameters changes in ISP process which affect significantly the wastewater characteristics from the ISP production. On the other hand, the industrial primary AAR acts as an equalization tank, reducing the characteristics variation of both, the WWIA and the WWIC effluents.

B. Bench scale primary system

In these experiments, the efficiency of the bench scale primary system was compared with the industrial one relating to COD, TSS/VSS and protein removal. Consequently, analysis of these parameters were carried out with the industrial raw wastewater (WWR) - that feeds the bench system - with the bench AAR effluent (WWBA) and with the bench clarified effluent (WWBC). Moreover, as stated earlier, gas production, pH and VFA were constantly monitored.

Figure 3 presents the WWR and WWBA pH variation during different bench experiments. All the experiments were carried out under the same operating conditions.

Figure 3. pH variations of WWR and WWBA during different bench experiments.

It can be observed that the pH variation of the WWBA was lesser than that of the WWR, which varied within the range of 4.0 and 4.7, with an average value of 4.3 ± 0.18. The WWBA pH varied from 3.5 to 3.8, with an average value of 3.7 ± 0.14. The WWBC pH values are not presented in Fig. 3 because they remain almost constant and equal to 4.3 in all experiments; this result has already been expected, since in the sedimentation stage, the pH is adjusted to this value.

Comparing the results presented in Fig. 3 with the industrial pH values (WWR, WWIA), it can be observed that the pH behavior during bench AAR digestion was very similar to the industrial one, indicating, that it could be happen the hydrolysis of the biopolymers and, as a consequence, the formation of VFA by acidogenic bacteria (responsible for pH decrease).

Figure 4 presents the wastewater pH values as a function of time during bench experiments. All the experiments presented very similar behavior and, for the sake of clarity, just some of them are shown in the figure.

Figure 4. Wastewater pH behavior during the bench experiments, where different symbols represent different experiments.

As can be seen in Fig. 4, the raw wastewater enters in bench AAR with a pH value around 4.3 and this value decreases more significantly during the first 2 h of digestion, when wastewater pH reaches a value around 3.6. After this period, two different behaviors could be distinguished: in some cases, pH continued lowering (more slightly) until a pH value around 3.4 and then, with time, this value increased again until a pH value around 3.7 in the end of experiment; in others cases, after the first 2 h, the wastewater pH value remained steady until the end of experiment.

From these results, it can be speculated that the greatest VFA accumulation occurs during the first 2 h of experiment due to rapid hydrolysis of easily degradable compounds presented in the wastewater.

Fang and Yu (2000) studied the HRT effect on the acidogenesis of dairy waste at pH 5.5 in the temporal range of 4-24 h. The authors found that the acidification increased rapidly with HRT, mainly in the first 12 h of experiment; the biodegradability of the major constituents in dairy wastewater increased with HRT in the following order: carbohydrates > proteins > lipids.

Figure 5 presents a comparison between a wastewater pH curve, and its respective VFA accumulation curve, during a selected bench experiment. Pointed lines were inserted in order to facilitate visualization.

Figure 5. Relation between pH and VFA accumulation during a reaction time of 6 h.

It can be observed from Fig. 5 that the pH curve follows, in opposite direction, the VFA accumulation curve, confirming that the pH decreasing is caused by VFA accumulation due, mainly, to carbohydrate hydrolysis. It can also be observed that the VFA minimum concentration presented by the bench system was around 1,300 mg.L^-1. According to Neto (1992) and Beal (1995), the concentration of these kind of acids are related to the activity of acidogenic and metanogenic bacteria and system concentrations over 1,000 mg.L^-1 indicate the predominance of acidogenic bacteria. Consequently, it can be stated that the bench AAR system behavior was in accordance with expectation.

According to the study developed by Yu and Fang (2003), the VFA formation rate is hardly affected by pH. The operation at pH between 4.0 and 5.0 favors the production of propionate and hydrogen, whereas the operation at pH between 6.0 and 7.0 favors the production of acetate and butyrate.

Regarding to VFA production by acidogenic bacteria, it is possible to estimate the product formation rate (r_p). This kinetic parameter was estimated during the exponential period of the VFA accumulation curve and the average r_p obtained was around 0.23 g.(L.h)^-1. Guerreiro et al. (1999) reached the maximum VFA formation rate of 2 kg.m^-3.h^-1 during the acidogenesis of a fish meal factory effluent at 55 °C at HRTs ranging from 12-24 h.

The acidogenic prevalence in bench AAR system could also be confirmed through the control of gas production during experiments. As stated earlier, during acidogenisis the gaseous production is mainly constituted by CO₂ and H₂ and, during methanegenesis, the main gas produced is CH₄ (Tchobanoglous et al., 1991; Neto, 1992). Thus, during the first experiments, only the production of methane gas was measured, which was carried out with the gas meter system, as shown in Fig. 2; this system comprises a column filled with NaOH capsules, a bottle with an alkaline solution (KOH 50%) and an indicator which guarantees the dissolution of other gases and the measurement of methane production only.

During the first experiments, when methane production was measured, none volume of gas was produced during 6 h of experiments, confirming the predominance of acidogenisis in bench AAR. This result agrees with those obtained by Guerreiro et al. (1999) which did not perceive methane production in the first 6 h of an anaerobic reactor operation under continuous stirring.

After the confirmation that none methane was being produced, the gas cleaning system was removed and the production of total gas was measured in the following experiments. The total gas volume produced as a function of time is presented in Fig. 6.

Figure 6. Total gas generation curves during bench experiments (reaction time of 6 h).

It can be observed that this production was highly variable during the experiments and this variation could be related to different parameters, such as initial pH, COD, VSS and/or VFA values, as well as the variation of these same parameters during the experiments.

Demirer and Chen (2005) obtained a biogas production rate of 0.356 ± 0.072 L.day^-1 during the one-phase conventional anaerobic digestion of unscreened dairy manure with an organic loading rate (ORL) of 1.0 gVS.(L.day)^-1. According to these authors, an increase in the influent ORL causes an increase in biogas production.

The majority of the curves obtained in this study, however, presented a similar tendency: a more significant production in the first 2 h and a slightly one until the end of the experiment.

Figure 7 presents the COD values of WWR, WWBA and WWBC in different experiments. Pointed lines have been inserted in order to facilitate the visualization.

Figure 7. WWR, WWBA and WWBC COD values in bench experiments.

As can be seen in Fig. 7, the WWR CODs values were very variable and the variation of the WWBA and WWBC CODs values were proportional to them. The proportionality between the WWR and WWBA curves and the COD enhance between them, could be explained by the addition, in all the experiments, of an equal volume of sludge (microbiological biomass) to the AAR. The increase in wastewater COD was higher in bench system when compared with the industrial one. A reasonable explanation for this behavior is the higher microorganism concentration in the bench system reac- tor, in comparison with industrial system which operates in a continuous mode.

The proportionality between the WWR and WWBC curves could also indicates that the system stability is reached, since the COD removal rate remains roughly constant in all experiments. The bench primary system efficiency can be observed through the average COD removal obtained, which is about 24 %.

The COD represents the wastewater organic matter (sugars and protein, mainly) and this organic matter is consumed by acidogenic bacteria during anaerobic digestion (in their growth and development); thus, it is possible to estimate the AAR substratum consumption rate, r_s, through the comparison between the WWR and WWBC curves. These values were very variable in all bench experiments, with a r_s average value around 0.8 g.(L.h)^-1.

Figure 8 presents the WWR, WWBA and WWBC VSS average results in bench experiments. This figure also presents the total VSS content that fed AAR the WWR VSS content plus the added sludge (as inoculums) VSS content (WWR+S). Pointed lines have been, once more, inserted for the sake of better visualization.

Figure 8. Averages results of WWR, WWBA, WWBC and WWR+S VSS content.

As can be seen in Fig. 8, the bench primary system presented a high efficiency in SSV removal, (around 76 %, average value). This removal, however, were a little bit lower than the industrial one. Comparing the WWR and WWR+S curves, it is possible to note that the SSV increment was caused by the sludge addition. The difference between the WWR+S and the WWBA curves represents the microbial growth during anaerobic digestion. This growth can be estimated through the measurement of the biomass formation rate (r_x). The estimated values were once more, very variable; considering all bench experiments, it was possible to estimate an average r_x of 0.08 g.(L.h)^-1.

Finally, since the main objective of the bench (and industrial) primary system is to transform the soluble proteins into insoluble ones, and to remove them, the WWR, WWBA and WWBC protein content were also estimated. Figure 9 presents the protein content variation of WWR, WWBA and WWBC in bench experiments.

Figure 9. WWR, WWBA and WWBC protein content in bench experiments.

The average protein content removal was around 49 %, which represents a good result, although a little bit lower than the industrial one.

It can also be observed that the WWR protein contents in all bench experiments were quite variable, differing significantly from the WWBC protein contents. The difference between WWR and WWBA curves is caused by the acidogenic bacteria addition (as inoculums) and their growing. It would be interesting to find out how much of this protein is really from the soy and how much is from the bacteria. With these data, a more accurate conclusion would be possible to be taken regarding to protein content. This would be a subject of a further research.

From the kinetics parameters previously calculated - r_x, r_s and r_p - it is possible to estimate others of great experimental interest such as the growth specific velocity (μ_x), the substrate consumption specific velocity (μ_s) and the product formation specific velocity (μ_p).

The obtained values were rather variable, but they are useful in order to obtain data about the microorganism behavior inside the AAR. The estimated μ_x average value was around 9.5×10^-6 h^-1; the estimated μ_s average value was around 8.7×10^-5 g_s.(g_x.h)^-1, where g_s is the mass of consumed substratum and g_x is the mass of biomass. Finally, the estimated μ_p average value was around 3.8×10^-5 g_p.(g_x.h)^-1, where g_p is the mass of generated product.

Blonskaja et al. (2003) studied two anaerobic digestion systems of distillery waste and, with HRT ranging from 10 to 19 days, the authors obtained a value for μ_x equal to 0.159 h^-1. Since the study of these authors comprises a secondary anaerobic acidogenic reactor (and not a primary AAR), a μ_x value highly superior to that obtained in this study was already expected.

IV. CONCLUSIONS

The present work investigated the wastewater treatment system of an isolated soy protein industry. The analysis comprises a comparison between the results obtained in a bench scale and those from the industrial treatment.

The results showed that the proposed bench system was able to reproduce the industrial results: the average COD, protein and SSV/SST removal obtained by industrial primary system were 25 %, 56 % and 85 %, respectively, against 24 %, 49 % and 76 %, respectively, obtained by bench scale primary system.

The bench wastewater final pH was around 3.7 and the more significant pH decrease occurred in the first 2 h of the experiment, indicating a rapid hydrolysis of the wastewater carbohydrates to VFA, CO₂ and H₂. This behavior was also confirmed by the higher VFA accumulation during the first 2 h.

The production of methane was not perceived in the experiments in which the production of this gas was measured. The total gas production was rather variable during the experiments and it was observed a significant gas production during the first 2 h.

This bench scale study gives an insight of the industrial primary wastewater treatment system - since it operates under similar conditions - and can be very useful in future researches that aim the study of the application of a pre-treatment to the ISP wastewater, which would reduce the protein content and, consequently, the organic load of the wastewater that enters the industrial primary system and would allow the recovery of the protein lost as effluent during the ISP production process.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge Capes (Co-ordination for Formation of High Level Professionals) for providing financial support through a PhD. scholarship and the PRODOC Project.

REFERENCES
1. Bayard, R., J.A. Morais, M. Rouez, U. Fifi, F. Achour and G. Ducom, "Effect of biological pretreatment of coarse MSW on landfill behaviour: laboratory study," Water Sci. Tech., 58, 1361-1369 (2008).         [ Links ]
2. Beal, L.B. Avaliação de processo anaeróbio em duas fases tratando efluente de curtume, M.Sc thesis, Instituto de Pesquisas Hidráulicas, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brasil (1995).         [ Links ]
3. Biswas J., R. Chowdhury and P. Bhattacharya, "Kinetic studies of biogas generation using municipal waste as feed stock," Enz. Microb. Tech., 38, 493-503 (2006).         [ Links ]
4. Blonskaja, V., A. Menert and R. Vilu, "Use of two-stage anaerobic treatment for distillery waste," Adv. Environ. Res., 7, 671-678 (2003).         [ Links ]
5. Borja R., B. Sánchez, B. Rincón, F. Raposo, M.A. Martin and A. Martín, "Study and optimization of the anaerobic acidogenic fermentation of two-phase olive pomace," Proc. Biochem., 40, 281-291 (2005).         [ Links ]
6. Calli B., B. Mertoglu, B. Inanc and O. Yenigun, "Effects of high free ammonia concentrations on the performances of anaerobic bioreactors," Proc. Biochem., 40, 1285-1292 (2005).         [ Links ]
7. Campos, J.R., "Tratamento de esgotos sanitários por processo anaeróbio e disposição controlada no solo," ABES, Rio de Janeiro, RJ (1999).         [ Links ]
8. Demirer, G.N. and S. Chen, "Two-phase anaerobic digestion of unscreened dairy manure," Proc. Biochem., 40, 3542-3549 (2005).         [ Links ]
9. Eaton, A.D., L.S. Clesceri, A.E. Greenberg and M.A.H. Franson, Standards Methods for the examination of water and wastewater. American Public Health Association, Washington (1995).         [ Links ]
10. Fang, H.H.P. and H.Q. Yu, "Effect of HRT on mesophilic acidogenesis of dairy wastewater," J. Env. Eng., 126, 1145-1148 (2000).         [ Links ]
11. Guerreiro, L., F. Omil, R. Méndez and J.M. Lema, "Anaerobic hidrolysis and acidogenesis of wastewater from food industries with high content of organic solids and protein," Water Res., 33, 3281-3290 (1999).         [ Links ]
12. Han S., S. Kim and H. Shin, "UASB treatment of wastewater with VFA and alcohol generated during hydrogen fermentation of food waste," Proc. Biochem., 40, 2897-2905 (2005).         [ Links ]
13. Hanaki, K. and T. Matsuo, "Evaluation of effectiveness of two-phase anaerobic digestion process degrading complex substrate," Water Sci. Tech., 39, 958 (1986).         [ Links ]
14. Mohamammadi, T. and A. Esmaeelifar, "Wastewater treatment using ultrafiltration at a vegetable oil factory," Desal., 166, 329- 337 (2004).         [ Links ]
15. Neto, V.M. Processos Anaeróbios. Senai-RS, Estância Velha (1992).         [ Links ]
16. Shin, H., J. Youn and S. Kim, "Hydrogen production from food waste in anaerobic mesophilic and thermophilic acidogenesis," Int. J. Hyd. Ener., 29, 1355-1363 (2004).         [ Links ]
17. Tchobanoglous, G., F.L. Burton and H.D. Stensel, Wastewater engineering: treatment, disposal, and reuse, McGraw-Hill, Metcalf & Eddy Inc., New York (1991).         [ Links ]
18. Yu, H.Q. and H.H.P. Fang, "Acidogenic of gelatin reach wastewater in an upflow anaerobic reactor: influence of pH and temperature," Water Res., 37, 55-66 (2003).         [ Links ]

Received: June 9, 2009
Accepted: September 17, 2009
Recommended by Subject Editor: Ricardo Gómez