Performance evaluation of anaerobic baffled reactor (ABR) treating pulp and paper wastewater in start-up period

The role of the anaerobic process has changed dramatically in wastewater treatment since the mid-1980s (Woodard & Curran Inc 2006). Its application for industrial wastewater treatment was classically preferred because of its low operating and maintenance costs. Yet, some major advantages could highlight its applicability. These are excess sludge reduction, energy recovery via biogas production, low requirements with respect to nutrients, high organic loading rates (OLR), and the possibility of toxic component biodegradation (Rittman & McCarty 2001).

The anaerobic baffled reactor (ABR) is known as a high rate system and was initially developed in 1985. It was immediately attractive to researchers and operators, and was introduced as a promising approach for municipal and industrial wastewater treatment (Barber & Stuckey 1999). The configuration of ABR is similar to a series of upflow anaerobic sludge blanket (UASB) reactors. However, it does not rely on granulated sludge forming, although granulation can occur over time (Baloch 2011). Primarily, ABR has a variety of advantages in its configuration and maintenance, including its simplicity of design and mechanical equipment, relatively low hydraulic retention time (HRT) requirements, and high stability with respect to hydraulic, toxic and organic shock loads. Furthermore, the unique configuration of ABR may provide partial separation of acidogenic and methanogenic bacteria (Barber & Stuckey 1999).

Recently, ABR has been successfully used for treatment of various wastewaters. These have come from a variety of sources, including domestic (Krishna et al. 2009), distillery (Akunna & Clark 2000), soybean protein processing (Zhu et al. 2008), swine wastes (Boopathy 1998), textile dyes (Goel 2010), and heavy oil production (Ji et al. 2009). It has also been studied as a pretreatment for different biological units like waste stabilization (Jamshidi & Gholikandi 2014), and duckweed ponds (Nasr et al. 2009). In addition its configuration has been upgraded by electrolytic enhancement methods (Gholikandi et al. 2013), and by bamboo carriers (Feng et al. 2008), or modified for low strength municipal (Bodkhe 2009) and industrial wastewaters, like those from pulp and paper making (Hassan et al. 2014).

Paper mill industries use huge amounts of the lignocellulose components of plants and can produce large volumes of wastewater. The influent characteristics can be highly variable depending on the manufacturing processes, and their mechanical and chemical refinements (Kurita 1999). Typically, the plants, chemical recovery, and preparation units are all major producers of alkaline and fiber effluents. These are complicated mixtures consisting of several compounds, most of which are not readily biodegradable (Wang et al. 2006).

The combination of anaerobic and aerobic treatment processes was found to be efficient for the removal of soluble organics in pulp and paper mill wastewater (Pokhrel & Viraraghavan 2004). It is thought that the large and complex molecules can be broken down into more readily biodegradable substances in an anaerobic process. These can then be further digested in the aerobic units. The performance of ABR in treating the black liquors of the pulp and paper industry has been studied by a few researchers. Grover et al. (1999) verified that the optimum efficiency can be achieved in ABR inoculated with black liquor having chemical oxygen demand (COD) of 4,000 mg/L while the pH, OLR and temperature are 7.5, 2 kg-COD/m³.day and 35 °C, respectively. The maximum COD removal and OLR were reported respectively as 66% and 7 kg-COD/m³.day by Kennedy et al. (2006). The start-up performance of a modified anaerobic baffled reactor (MABR) treating recycled paper mill (RPM) wastewater as batch and continuous phases was investigated by Zwain et al. (2013). They concluded that high COD removal efficiency (up to 71%) can be attained within a month. Yet, the start-up period of ABRs for paper mill wastewater treatment has been studied on only a few occasions.

This essay focuses on the performance evaluation of ABR treating wastewater from the Mazandaran Wood and Paper Industry (MWPI). For this purpose, the pilot study was carried out using a real substrate that had been inoculated. The concentrations of organics were traced within the ABR through its chambers. These show how it performs during the start-up period and in organics removal. The results are also discussed in comparison with the previous experimental outcomes from an identical ABR pilot-plant treating municipal wastewater. The latter work was intended to highlight the impact of paper mill wastewater on the operation of the reactor.

The MWPI is located in the north of Iran, near Sari. At present, two conventional activated sludge units with an average HRT of 40 hours and SCOD removal efficiency of 68% are used for wastewater treatment. It is obvious that the effluent would be unlikely to meet the national standard limits. This may be due to the low-rate biodegradability characteristics of the influent, which has meant that the aeration tanks had to be operated with high HRTs. The layout of the wastewater treatment plant and the average HRTs of units are shown in Figure 1.

The pilot plant was installed and inoculated directly from the equalization tank through the pumps, as shown in Figure 1. The overall characteristics of the influent are shown in Table 1. The ABR was made up of four chambers with an overall net volume of 45 liters. The total length, width, and height were 60, 15, and 50 cm, respectively. The HRT was reduced gradually in the start-up period from 64 to 24 hours, in two-week time steps over four months. As a consequence, the nominal upflow velocity was increased from 14 to 35 cm/h, a similar time to that used in the previous studies (Nasr et al. 2009; Krishna et al. 2009). This system was operated at a controlled temperature of 34 to 39 °C using a hot water bath and heaters (Figure 2). The digested sewage sludge (DSS) was obtained from the ABR of Babol Industrial Town and used as the seed culture. It was screened through a 2 mm sieve before application. The DSS had 13.4 g/L total suspended solids (TSS) and 3.9 g/L volatile suspended solids.

Biological and physico-chemical parameters like temperature, pH, alkalinity, total nitrogen (TN) and phosphorus (TP), volatile fatty acids (VFAs), including acetic and propionic acids, 5-day biochemical oxygen demand (BOD₅), soluble and total COD, and TSS were analyzed during the 4-month run time in all samples obtained from the influent, chambers and effluent of the ABR. These analyses were performed according to the standard methods for the examination of water and wastewater (APHA 2005). The filtrate SCOD and TSS were analyzed respectively by filter paper 602H and 597 Schleicher & Schuell^®. A HACH^® DR 2000 spectrophotometer was used for COD, TN and TP measurement. VFAs were analyzed with a gas chromatograph flame ionization detector, and the SCOD and biochemical oxygen demand (BOD) laboratory samples were prepared by Centrifuge Z200A HERMLE^® and AQUALYTIC^® OxiDirect, respectively.

The physico-chemical characteristics of the substrate confirm that the biodegradability of the contaminants in the wastewater was low. The average ratio of BOD to COD (BOD: COD) was about 0.22 and so it might not be possible to treat the substrate biologically (Tchobanoglous et al. 2003). This could be due to the inhibitors present in the influent, which are typically sulfides, tannins, resin acids, long chain fatty acids (Palatsi et al. 2009), some heavy metals (Colussi et al. 2009), and halogenated compounds (Ali & Sreekrishnan 2001). Identical lignin-derivative compounds are also highly toxic to methanogens (Chen et al. 2008). The complete conversion of VFAs to biogas may be influenced by the presence of lignin and related compounds. Slow rate reactions, and the partial hydrolysis and consequent acid accumulation in the start-up period can possibly be explained in this way as well. Similar results have been reported for black liquor treatment by Grover et al. (1999).

Regarding the experimental results, it was observed that after about 90 days, the COD removal efficiency of the ABR was much more stable (Figure 3). This period was also identified as the proper start-up period for low strength municipal wastewater in the pilot scale (Bodkhe 2009; Krishna et al. 2009). As illustrated in Figure 4, the SCOD removal efficiency increased gradually to its maximum value of 60%, for which the best HRT would be about 24 to 30 hours. This is much less than the 48 hours that was recommended previously by Grover et al. (1999). However, the removal efficiency was almost the same as that obtained and reported by Zwain et al. (2013) and Kennedy et al. (2006).

The trend of SCOD removal in start up (Figure 4) was recognized as a two order function (R² = 0.88) which is almost logarithmic. This implies that the gradual decrease in HRT would not inhibit the overall growth of microbial activity in start-up. Conversely, it might lead to an increase in the reaction rate. If SCOD removal is plotted via OLR, the influence of HRT at start-up can be recognized. Since the substrate is pumped through the equalization tank, and in accordance with the experimental results, the influent SCOD concentrations have a certain range of variations. In Figure 5, the points attributed to high OLRs (0.9 to 1.6) are calculated in low HRTs. It seems that the HRT has a considerable effect on ABR performance and operation during the start-up period, when treating pulp and paper wastewater. Since the efficiency of COD removal did not vary significantly at the end of the start-up period, it was recommended that the HRT be maintained at 24 hours at least. It is noteworthy that this pilot-plant could remove 85% of the organic load in municipal wastewater treatment with an HRT of 24 hours. The influent contained average SCOD and BOD concentrations of 450 and 335 mg/L, respectively.

The total concentrations of VFAs (i.e. the acetic and propionic acids) in the influent and effluent did not vary significantly. They were only reduced by an average of 3% in the final days of the start-up period. This implies that they were digested at the same rates as the rates at which they were produced. In addition to this, the pH increased from the range 6.8 to 7.1 in the influent to 7.2 to 7.4 in the effluent. This is the optimum range for the methanogenic phase. The total alkalinity did not change significantly. Although the pH and alkalinity values were firmly lower than those observed in similar studies. Dangcong & Qiting (1993) observed that the black liquor could be effectively treated over the pH range 9.5 to 10.6. Some researchers have suggested that the low organic load may have led to small amounts of acid being produced and that, consequently, there was little requirement for neutralization. The results were however, supported by the work of Prasad (1992) and Gholikandi et al. (2014) on an effluent from a bagasse based paper mill and domestic wastewater, respectively.

As shown in Figure 6, the ratio of BOD to COD increased gradually to more than 0.55 through the chambers of the ABR. It seems that some part of the COD was degraded and converted into BOD. This is due to the fact that BOD is a 5-day test. The slowly biodegradable compounds are not usually measured by BOD. However, they are detected in COD. Therefore, hydrolysis and degradation of slowly biodegradable compounds by anaerobic bacteria in the initial chambers may lead into a reduction in COD (about 40%) but an increase in BOD. The increase in BOD concentration continued to the third chamber (to about 400 mg/L) while it slightly reduced to 350 mg/L in the final part. In addition, the VFA concentration increased in the first chamber and then fell slowly as the liquor passed through intermediate chambers 2 and 3 to the last (total removal of 3%). In an identical pilot-plant and HRT configuration treating municipal wastewater, it was observed that VFA production and removal followed the same trend (Gholikandi et al. 2014). The only significant difference is that, for municipal wastewater, because of the prevalence of readily biodegradable compounds, the rapid digestion of BOD led to the BOD:COD ratio falling more sharply (Figure 7). The first chamber of both pilot-plants was the site of the principal COD removal and VFA production. Also, it can be inferred that acidogenesis (or hydrolysis) and methanogenesis were partially separated in passage through the ABR, because, in the first two chambers, the BOD and VFA concentrations increased while in the latter two, they were reduced. This cannot lead to the conclusion, however, that methanogenesis had started, because, in comparison with the municipal wastewater, the COD values were not reduced significantly.

From the above it seems that the use of an ABR to treat paper mill wastewater was largely similar to its application for treating municipal wastewater. This may be due to the strength of the substrates which have, as a result, controlled the whole process. However, it may be necessary for the ABR to comprise four chambers if it is to be used to treat industrial wastewaters, to ensure stable start-up and operation. It is noted that the long chain, polymer compounds were hydrolyzed and broken down into readily biodegradable compounds that could be digested further in the biological units, like activated sludge. Consequently, the application of ABR as a pretreatment prior to the current aeration tanks at MWPI can not only improve the efficiency of COD removal but may also reduce the costs of operation and maintenance.

This study is aimed at assessing the feasibility of successful start up of an ABR system to treat pulp and paper wastewater. For this purpose, its performance was evaluated at pilot scale. The experimental results show that the ABR was successfully started up in three months and achieved considerable organic removal efficiency. It also verifies that ABR can facilitate the biological digestion of slowly biodegradable compounds of a substrate. In addition, the simple and sustainable mode of operation of the ABR can help in its application as a pretreatment process for activated sludge units. All of these show that the use of an ABR can be recommended strongly, with an optimum HRT of about 24 hours and four chambers.

We would like to thank MWPI, Caspian Sea Ecology Research Institute and Mrs. Salimi, Wastewater Treatment Plant of Babol Industrial Town, for their technical support and assistance in this project.