Studies on laboratory-scale submerged aerobic fixed film reactor (SAFF) packed with synthetic media having specific surface area of 165 m2/m3 with a void volume of 89% were carried out to assess its performance under various organic loading rates (OLR) and hydraulic retention times (HRT). Synthetic wastewater having chemical oxygen demand (COD) and biochemical oxygen demand (BOD) of 400 ± 10% and 210 ± 10% mg/L respectively was fed and the reactor was subjected to OLRs ranging from 0.37 to 1.26 kg COD/m3.d. It was observed that steady sloughing of biofilm occurs within the SAFF reactor all the times and average concentration of sloughed biomass in the effluent was 26 mg/L. The COD and BOD removal efficiencies varied between 85 and 89% and 86 to 94%, respectively. The kinetic studies demonstrated that SAFF reactor followed Stover–Kincannon and Grau models, with high correlation coefficients (R2) of 0.9977 and 0.9916, respectively. Thus, the values of kinetic coefficients such as maximum substrate utilization rate, Umax = 64.1 g/(L.d); saturation value constant, KB = 72.31 g/(L.d) and Grau second-order substrate removal rate constant, Ks = 2.44 day−1 can be useful to develop and design large scale SAFF reactors. Finally, the study reveals that the optimum range for OLR can vary within 0.68–0.94 kg COD/m3.d.
Since the inception of fixed-film growth processes in aerobic biological wastewater treatment, trickling filters (TF) and rotating biological contactors (RBC) were widely implemented and studied. However, submerged aerobic fixed film (SAFF) reactors are a relatively recent development in the fixed attached growth biological process (Schlegel & Koeser 2007). The SAFF is a hybrid technology that is gaining attention due to its high degree of treatment efficiency with the added advantage of low footprint area and less operation and maintenance costs. It comprises of a column or tower packed with media for biofilm growth and diffused aerator to supply oxygen. It can be operated in both up-flow and down-flow mode. Additionally, it allows accumulation of high concentration of active biomass with high sludge retention time, thereby resulting in low sludge production, and resistance to hydraulic or organic shock loads along with viable aesthetics (Goncalves et al. 1998). Moreover, Huang & Bates (1980) observed that biofilms in aerobic processes were effective in organic matter removal and had a better settleability than anaerobically developed biofilm. A support structure where biofilm develops is commonly known as media, and it is an integral part of the SAFF reactor. Selection of media has a profound influence on the rate of biofilm attachment and growth, and it should be chemically inert and economically viable. Provision of media having high specific surface area is beneficial as it provides more area for microbial attachment without occupying working volume in the reactor (Mendoza-Espinosa & Stephenson 1999; Azizi et al. 2013). It is worthy to note here that specific surface area of the media shall be in tandem with the voidage. A media having high specific surface area with less voidage, e.g. sponge (Zhang et al. 2016) may create problems of frequent clogging within the reactor, but improves the solids retention abilities. In contrast, media with high specific surface area and more voidage, e.g. plastic media reduces the clogging problems but warrants complimentary secondary sedimentation unit (Dermou et al. 2007; Goncalves 2007). Thus, it is up to a designer to choose amongst two combinations, as each of them has its own set of positives and negatives.
There are 816 sewage treatment plants (STPs) operating in India, having a treatment capacity of 23,277 MLD based on various technologies such as activated sludge process (ASP), sequential batch reactor (SBR), up-flow anaerobic sludge blanket reactor (UASB), etc. However, there is only one STP based on SAFF technology, having a treatment capacity of 0.21 MLD (CPCB 2015). Many full-scale plants have been developed to treat high strength and toxic industrial wastewater, and there is insistent need for studies for municipal sewage which has low organic strength (CPHEEO 2013). Several studies with low strength wastewater have been reported recently. Xu et al. (2019) evaluated SBR for enhanced nutrient removal having influent chemical oxygen demand (COD) of 350 mg/L; Wang et al. (2017) developed a novel approach for nutrient removal with influent COD of only 150 mg/L in ASP. However, fixed film aerobic reactors are less frequently studied for low strength wastewater. This study therefore aims to evaluate the performance of SAFF reactor to treat low strength wastewater and determine kinetic coefficients. The kinetic studies enable the identification of operational parameters that affect biological performance, thereby contribute to the establishment of optimal operational conditions for the treatment process. Generally, simplified kinetic models are used to reduce the numbers of variables, which facilitate the study, design, and scale-up of biological reactors (Hamoda 1989). The Stover–Kincannon model and Grau (second-order substrate removal model) have been used widely for the submerged aerobic filter (Optaken 1982; Yu et al. 1998; Borghei & Hosseiny 2002).
MATERIAL AND METHODS
The schematic of a laboratory-scale SAFF reactor is shown in Figure 1. The reactor was cylindrical having a total volume of 5.5 L (effective volume 4.5 L) and made from transparent acrylic material. Other components of experimental setup included influent and effluent containers, peristaltic pump for pumping synthetic wastewater, air pump for supplying oxygen in the reactor, rotameter to control and measure air-flow rate, circular diffuser disc for oxygen delivery/distribution inside the reactor.
The reactor was packed with media having a specific surface area of 165 m2/m3. A perforated acrylic disc at the bottom was provided for supporting the filter media, and a similar disc was kept at the top to prevent the floating tendency of media while keeping them under submerged condition. Once the biofilm is developed, media weight will increase, which will further help in keeping the media submerged and immobilized. The media bed occupied a height of 13 cm (i.e. 80% of total volume) within the two perforated discs. The flow rate was varied by peristaltic feed pump, and the air-flow rate was regulated by rotameter.
The efficient performance of any fixed-film process is a function of viability of biofilm development (Kornegay & Andrews 1968; La Motta 1976). For commissioning, the reactor was operated for 30 days and was under batch mode for the first 20 days. During this period, it was fed with domestic sewage acquired from NEERI campus and an airflow rate of 1 L per minute (LPM) was provided in order to develop biofilm on the media. The commissioning started in batch mode in order to allow microorganisms to develop on the surface of the media without being washed due to hydraulic erosion. In order to promote initial biofilm growth, quiescent conditions were maintained under batch mode. This resulted in enhanced biofilm growth on media because of the presence of readily available bacterial population in sewage. It was thought that relatively quiescent conditions will favour microbial attachment to media quickly under batch mode. Once the biofilm was developed, the reactor was operated under continuous mode for the next 10 days. Since the commissioning started in winter (i.e. December and January), sewage temperature was maintained at 35 °C with the help of a water bath. During this period, sewage was mixed with laboratory prepared synthetic wastewater having a composition as per Wei et al. (2014). Subsequently, the quantity of sewage was slowly reduced such that at the end of the 10th day, only synthetic wastewater was fed.
After successful commissioning, the reactor was fed with synthetic wastewater at various hydraulic loadings. The hydraulic loadings were altered only when the reactor achieved steady removal efficiency with respect to COD and biochemical oxygen demand (BOD). The flow rate at initial loading was 4.32, 5.27, 7.54, 10.64, 13.98 L/d and subsequent organic loadings were 0.37, 0.47, 0.68, 0.94, 1.26 OLR (kg/m3.d), respectively, as shown in Table 1. The reactor was operated in a laboratory at room temperature (25 ± 4 °C) which was devoid of any direct sunlight. As a result, algae formation was not observed in the reactor throughout its operational period. The rotameter, air, and peristaltic pump were kept at higher levels to prevent backflow from the reactor in case of electrical or mechanical failure. Only when the dissolved oxygen (DO) concentration in the reactor dropped below 3 mg/L, the airflow rate was increased.
|Flow rate (L/d) .||Organic loading rate (kg/m3.d) .||Hydraulic retention time (h) .||Days of operation .|
|Flow rate (L/d) .||Organic loading rate (kg/m3.d) .||Hydraulic retention time (h) .||Days of operation .|
|Grau second order|
|Umax .||KB .||ks .||m .||n .|
|Grau second order|
|Umax .||KB .||ks .||m .||n .|
The system performance was monitored by analyzing Physico-chemical parameters such as pH, temperature, DO, COD, BOD, total Kjeldahl nitrogen (TKN), total phosphate (TP) and total suspended solids (TSS) for both influent and effluent throughout the operational period. Parameters such as COD, DO, TP, and TSS were performed daily, whereas TKN and BOD were analyzed four times a week. The samples were analyzed in compliance with the procedure mentioned in Standard MethodsAPHA (2005). Biomass concentration in the reactor at the end of the study was determined by taking five filter media unit representative of the whole reactor and weighing biofilm attached to them individually. The biofilm was scraped and washed from the media with deionized water and was collected in a beaker; the content of the beaker was passed through filter paper for TSS measurement in accordance with APHA (2005).
= Max. utilization rate g/(L.d)
= Discharge (L/d)
= Volume of reactor (L)
= Influent and effluent substrate conc. (mg/L)
= Saturation value constant g/(L.d).
Second-order substrate removal model (Grau model)
= Discharge (L/d)
= Hydraulic retention time (HRT) (days)
= Influent and effluent substrate conc. (mg/L)
= Grau second-order substrate removal rate constant (d−1)
= Biomass concentration in the reactor (g VSS L−1)
= So/(ks. X) (d−1)
= Constant for Grau second-order model (dimensionless).
RESULTS AND DISCUSSION
Performance of SAFF in terms of COD and BOD removal
Variation in COD removal efficiency along with its concentration in influent and effluent are presented in Figure 2. It was perceived that COD removal efficiency followed a distinct pattern, as soon as the OLR was enhanced, the efficiency went on to deteriorate (to 35–45%) for 4–7 days, before improving and subsequently reaching its highest value for each loading. Similar such patterns in removal efficiencies with the change in OLR was observed by Patel & Madamwar (2002) while operating anaerobic fixed film reactor and Ravi et al. (2013) while operating RBC. The duration between SAFF was subjected to new organic loading and started showing improvement, termed as acclimatization period or lag phase of microbes. In this period, the biomass get accustomed to new organic loading by multiplying themselves, which is proportional to the substrate rate provided. This period can extend further only when the reactor is under heavy sloughing (i.e. losing effective biomass), which was observed only in the third OLR (maximum of 10 days), before achieving stability. Maximum COD removal efficiency obtained at 0.37, 0.47, 0.68, 0.94 and 1.26 kg COD/m3.d organic loading were 93.00%, 92.16%, 90.43%, 90.63%, and 92.59%, respectively. As soon as the reactor stabilized, it was subjected to advance organic loading.
The BOD concentration for influent and effluent along with removal efficiencies are demonstrated in Figure 2. It was apparent that there was a similarity between the BOD and COD graphs. Maximum BOD efficiency was obtained at 0.37, 0.47, 0.68, 0.94 and 1.26 kg COD/m3.d organic loadings were 94.07%, 91.18%, 92.08%, 90.82%, and 94.55%, respectively. It can be deduced that the effluent BOD concentration was higher than the recently revised discharge standards, i.e. 10 mg/L for BOD (Nitin Shankar Deshpande vs Union of India & Others 2019). However, slightly higher values compared to standards can be reduced if provided with secondary clarifier. However, provision of settling tank ahead of SAFF reactor may further boost up SS removal which will subsequently result in higher COD and BOD removal efficiencies.
Effect of OLR on DO concentration
The DO concentration measured in SAFF reactor during the study period is presented in Figure 3. In order to maintain the minimum operating level of DO to 3 mg/L within the reactor (Borghei et al. 2008), airflow rate was increased with the increase in OLR. At OLR of 0.68 kg COD/m3.day, DO levels dropped below 3 mg/L on the 40th day. Thus, airflow rate was increased in a stepwise manner from 100 to 200 mL/min in two successive days. Similarly, at OLR of 1.26 kg COD/m3.d, DO concentration dropped to 2.5 mg/L on the 77th day, thereby resulting in an increase in airflow rate from 300 to 400 mL/min from the 77th to the 79th day. However, on the 80th day, it again dropped down to 2.7 mg/L. Hence, the airflow rate was increased to 400 mL/min. The sudden drop in DO concentration with the increase in OLR can be attributed to enhanced substrate utilization, which ultimately results in increased microbial population within the reactor.
Microorganisms are subjected to normal growth cycle (growth curves), which involves four distinct phases, i.e. lag phase, growth or exponential phase, stationary phase, and decline or death phase (Pelzar et al. 1993). However, in biological unit, it is important to maintain a stationary phase, i.e. where the microbes which are lost must be replenished in order to maintain constant microbial concentration essential for proper functioning of a biological treatment unit. In attached or combined growth process, the biomass is lost by sloughing. As a result, biomass gets carried away in the effluent. It is imperative to mention that steady sloughing keeps on happening all the time, mainly because of the action of hydraulic erosion on the exposed outer surface of biofilm (Lessard & Bihan 2003). Additionally, it can also be caused by famine and bacterial cells supersaturation (Henze et al. 2010). The SAFF follows attached growth process, and sloughing was measured in terms of outlet TSS concentration on every operational day and is presented in Figure 4. It is observed that TSS concentration at outlet for organic loading of 0.37, 0.47, 0.68, 0.98, and 1.26 kg COD/m3.d varied from 2–50, 6–100, 14–102, 2–86, and 4–48 mg/L, respectively. For OLR of 0.47 and 0.68 kg COD/m3.d, a trend of heavy biomass sloughing was observed. A maximum sloughing period of 14 days (Days 36–50) was observed under OLR 0.68 kg COD/m3.d. The surge in the BOD and COD values of effluent is credited to enhanced biomass sloughing during OLR of 0.47 and 0.68 kg COD/m3 as evident in Figure 2. The reduced biomass loss through sloughing at OLR 0.94 and 1.26 kg COD/m3 resulted in quick and easy acclimatization of microbes under enhanced OLR with a minimum period of 4 days required to attain stability. In this study, only once the reactor was subjected to a high airflow rate of 1 LPM (before the 4th OLR on days 50 to 52), because of the emergence of dead biomass layer a few times.
All forms of life, from microorganisms to human beings, share certain nutritional requirement for growth and normal functioning (Pelzar et al. 1993). However, the major objective of this study was not high nutrient removal, and they were merely provided as a building block for the microbes. The TKN and TP concentration at influent and effluent were analyzed, and an average was taken for each loading for calculating nutrient uptake. The average removal efficiencies obtained at 0.37, 0.47, 0.68, 0.94 and 1.26 kg COD/m3.d organic loading were 48.67, 60.91, 59.54, 48.71 and 49.45% for TKN and 15.52, 18.30, 27.14, 33.83, and 42.77% for TP, respectively.
From Figures 5 and 6, it is concluded that no high removal of nutrients took place during the study, which implies that it was only consumed by the biomass to carry out cell activity, cell development, and metabolites along with particular microbial end products (Pelzar et al. 1993). Also, it shows that the SAFF never operated under an anoxic condition. Since the removal of BOD and TKN co-occurs as the microbes consume these carbon and nitrogen sources to maintain their life cycle. Nevertheless, for complete nitrogen removal or treatment of nitrogen-rich wastewater, the SAFF can be coupled with the anoxic phase wherein denitrification will take place (Schlegel and Teichgraeber 2000). The phosphate removal, if required, can be achieved by using a biological process as a pretreatment step followed by the addition of ferric chloride (FeCl3) (Kim et al. 2016).
Biomass assessment studies were carried out at the end of the reactor's operational run. The concentration of attached biomass was 2,896 mg/L. It was found that some biomass was attached to sidewalls and discs with a concentration of 580 mg/L. Therefore, the total concentration of attached biomass in the reactor worked out to be 3,476 mg/L. In addition to attached biomass, some biomass was also found in a suspended condition, and its concentration was 372 mg/L. It indicates that most of the biomass was attached to the media. Thus, total biomass concentration in the reactor summed up to be 3,848 mg/L. This value of biomass concentration in SAFF reactor is comparable with the mixed liquor suspended solids (MLSS) in several suspended aerobic biological systems such as ASP (MLSS-3,400–4,000 mg/L), extended aeration (MLSS-3,000–5,000 mg/L) and SBR (MLSS–4,300 mg/L). It is important to mention here that, in aerobic fixed film reactors, there is no mechanism to waste the sludge, which is the case with suspended aerobic biological processes (CPHEEO 2013). Consequently, biomass tends to increase to the levels that it may cause clogging in the reactor. A similar phenomenon was observed on days 50 and 52, as mentioned in the section ‘Biofilm Sloughing’. The possible remedy for this is to increase the airflow rate at least once a week to get rid of clogging or else backwashing with wash water can also be done (Goncalves 2007). It was also observed that the clear water zone at the top of the reactor was covered completely by biomass. Therefore, an additional depth (except freeboard) in the unit must be provided. At the end of the final OLR of 1.26 kg COD/m3.d, F/M ratio was 0.32, and the corresponding solid retention time (SRT) was 53 days. Higher SRT is an indication of the fact that sludge production will be less. Thus, it can be seen as savings in sludge recirculation (in case of suspended aerobic systems) and management cost (Leu et al. 2012; Amanatidou et al. 2016). The media before the start of the study and biofilm growth on the media after the completion of the study is shown in Appendix A (available with the online version of this paper).
The performance data obtained from SAFF under various organic and hydraulic loading indicate that high COD removal efficiencies were obtained. In this study, Stover–Kincannon and Grau second-order substrate removal models were applied for COD removal in SAFF reactor. It was assumed that the functioning of the reactor and experiments were consistent.
In this model, the graph was plotted between V/Q.(So-Se) (L.d/g COD) i.e., reciprocal of total organic removal rate against V/(Q. So) (L.d/g COD), i.e. reciprocal of total organic loading rate (Işik & Sponza 2005), the graph is shown in Figure 7. From the graph, intercept of straight line is represented as 1/Umax and slope represent KB/Umax. The values of intercept and slope are 0.0156 and 1.128, respectively, with high correlation coefficients of (R2) 0.9977. The saturation value constant (KB) and maximum utilization rate (Umax) obtained are 72.31 g/(L.d) and 64.10 g/(L.d), respectively.
Grau model (second-order substrate removal model)
In the Grau model graph was plotted between HRT* So/(So-Se) and HRT (in days), and the same is shown in Figure 8. The values of m and n are obtained from the intercept and slope of the line fitted on the graph. The values of m and n are 0.0427 and 1.0895 with high correlation coefficients (R2) of 0.9916. The substrate removal rate constant (ks) attained from the equation m = S0/(ks. X) is 2.44 (per day).
In this study, for the Stover–Kincannon model the saturation constant (KB) and maximum utilization rate (Umax) values obtained are lower compared to the studies conducted by Yu et al. (1998) and Borghei et al. (2008) and higher than the values obtained by Borghei & Hosseiny (2002) for the same model. This difference in values for the constants may be attributed to higher rates of substrate utilization, which depends on the source of carbon in wastewater. Yu et al. (1998) used soybean wastewater, and simulated sugar manufacturing wastewater was used by Borghei et al. (2008), which is readily degradable in comparison with the primary source of carbon, i.e. starch in this study, thus a decrease in values of KB and Umax was observed. The kinetic coefficients obtained are presented in Table 2.
However, for Grau second-order kinetic model, the multi-component substrate removal rate constant (kS) value obtained under this study fall in the range determined in other studies by Borghei et al. 2008. The kS value will increase as the substrate removal rate increases, which depends on initial substrate concentration (S0) and its nature along with biomass concentrations (X) in the reactor.
Results obtained from this study showed that SAFF reactor is efficient in treating low strength wastewater. The COD and BOD removal efficiencies during the entire study period were well above 80% at OLRs ranging from 0.37 to 1.26 kg COD/m3.d and HRT of 25 to 8 h. It seems quite reasonable to conclude that the optimum range for HRT and OLR can be within 10–14 h. and 0.68–0.94 kg COD/m3.d, respectively. COD, BOD and SS removal efficiencies can be further improved by provision of sedimentation tank after SAFF. The SAFF reactor also exhibited good biomass retention resulting in long SRT with the average concentration of sloughed biomass varying between 10 and 45 mg/L at various OLR during the entire study period. At the same time, it is also important to mention that no odor and clogging problems were encountered and thus SAFF reactor exhibits its aesthetic acceptability. Finally, the data obtained from this study fitted well to both Stover–Kincannon and Grau second-order models with high regression coefficient (R2) of 0.9977 and 0.9916, respectively. Hence, the kinetic coefficients can be used to develop full scale SAFF based wastewater treatment plant.