Application of alternative carriers without protected surface in moving bed bio ﬁ lm reactor for domestic wastewater treatment

Biological reactors with immobilized biomass on free carriers have provided new perspectives for wastewater treatment, once they reduce the system size and increase the treatment capacity. In this study, the performance of three Moving Bed Bio ﬁ lm Reactors (MBBR) using different carriers (with and without protected surface area) were evaluated for domestic wastewater treatment in continuous ﬂ ow. Each MBBRs ( i.e., R1, R2, and R3) was ﬁ lled at a ratio of 50% with high-density polyethylene carriers with different characteristics: both R1-K1 and R2-Corrugated tube with protected surface and R3-HDPE ﬂ akes without protected surface. Chemical oxygen demand (COD) removal of 80 + 5.0, 80 + 3.5, and 78 + 2.4% was achieved by R1, R2, and R3, respectively. The oxygen uptake by bio ﬁ lm attached on the carriers was 0.0079 + 0.0013, 0.0033 + 0.0015, and 0.0031 + 0.0026 μ g DO·mm (cid:1) 2 for the K1, corrugated tube, and HDPE ﬂ akes, respectively. No signi ﬁ cant differences were observed between the performance of the three MBBRs in terms of physico-chemical parameters (alkalinity, pH, and dissolved inorganic carbon) and COD removal. Results showed that the carrier type and its characteristics (total area and with/without protected area) did not affect the organic matter removal. Thus, the carrier without a protected surface in MBBR could be a promising low-cost option for domestic wastewater treatment.


Carrier
Three different carriers (K1, corrugated tube, and HDPE flakes) were used as support material for attached biofilm bioreactors. The characteristics of carriers used in this research are given in Table 1.
The reactors R1, R2, and R3 were filled with 50% of carriers in the reaction zone according to Zinatizadeh & Ghaytooli (2015).

Experimental setup
The study was carried out in three lab-scale reactors operating in parallel (see Figure 1), which were fed with raw domestic wastewater from a sanitary sewer located at the University of São Paulo (São Carlos, São Paulo state, Brazil). The physico-chemical characterization of the raw wastewater is summarized in Table 2. Each acrylic reactor had a volume of 7.5 and 1.7 L for aeration tank and decanter, respectively. A sieve was placed in the decanter inlet to keep the carriers in the reaction zone of the MBBR. The mixed liquor recirculation was not performed because the reactors were operated as MBBR (Gong et al. 2012, Wei et al. 2016. The MBBRs were continuously fed using metering pumps (ProMinent Concept Plus, Germany) at a flow rate of 3.75 L·h À1 , which resulted in a hydraulic retention time (HRT) of 2 h. The low HRT is usually used in MBBR (0.5-3 h) and this value was adopted based on previous studies (Ødegaard 1999;Piculell et al. 2014;Metcalf & Eddy 2016;Bassin et al. 2016). The reactors were aerated by an air blower (Boyu, ACQ-003, China) at an  (1) storage tank of raw wastewater, (2) feed pump, (3) air blower, (4) rotameter, (5) MBBR, (6) carriers, (7) aerobic tank, (8) sieve (opening size of 5 mm), (9) decanter, and (10) MBBR effluent.
Water Practice & Technology Vol 00 No 0, 3 Uncorrected Proof air flow rate of 5 L·min À1 . A coarse bubble diffuser was used to provide aeration and mixing inside the bioreactors, maintaining the dissolved oxygen (DO) concentration higher than 3 mg·L À1 (Collivignarelli et al. 2019). The sludge excess from the decanter was discharged once a week. A sludge volume ranging from 1 to 2 L was removed each time, with a mean concentration of 1,039 + 343 mg TS·L À1 and 620 + 266 mg VTS·L À1 . The MBBRs were operated at room temperature (24 + 5°C). The reactors were operated for 80 days and samples (influent and effluent) were taken once a week to be analyzed in terms of physico-chemical parameters and organic matter.
The MBBRs were started up without a previous inoculum to evaluate the potential of each carrier to the biofilm formation using the microorganisms present in domestic wastewater.

Physico-chemical analyses
The MBBR influent and effluent were characterized by pH, alkalinity, dissolved inorganic carbon (DIC), total chemical oxygen demand (tCOD), and soluble chemical oxygen demand (sCOD). For DIC and sCOD analyses, the wastewater samples were previously filtered into 0.45 μm fiberglass membrane. Total Solids (TS) and Volatile Total Solids (VTS) were monitored only in the mixed liquor and raw wastewater. The dissolved oxygen (DO) measurements were done using a portable oximeter (YSI, EcoSence DO 200A, USA). The ammonia nitrogen and orthophosphate were monitored only in the raw wastewater. All analyses were performed in duplicate and according to APHA (2012).

Microsensor application for oxygen depletion test
To measure the DO consumption due to the organic matter oxidation by the biofilm, experiments were carried out using a mini respirometric cell with a capacity of 15 mL. The microsensor Clark-type DO was developed and built according to Gonzalez et al. (2011), and it consisted of two glass plates, a rubber o-ring, and an extravasation point of wastewater excess, and access to the DO microsensor.
The respirometric analyses were performed following the steps: (1) the raw wastewater samples were filtered into 1.2 μm membrane to remove the biological flocs which can interfere in the tests, and the filtered sample was fully saturated with DO before the tests using an aquarium air blower and porous stone; (2) 15 mL of saturated wastewater sample was put into the respirometric cell to measure the DO concentration without the carriers (blank); and then, (3) the DO concentration was measure in samples containing 15 mL saturated wastewater and the carriers. In all tests, the DO concentration was monitored for 10 min and the data were collected every one second.

Calculations
The MBBR performances were evaluated in terms of organic matter removal (or COD removal) and were expressed as CODr (%) according to Equation (1).
where tCODi is the tCOD of the wastewater influent (mg·L À1 ) and sCODe is sCOD of the effluent (mg·L À1 ). The Volumetric Organic Loading rates (VOL) (gCOD·m À3 ·d À1 ) of the influent wastewater were calculated according to Equation (2).
where OL (organic loading) is the influent organic loading (kgCOD·d À1 ) and Vr is the working volume of MBBRs (m 3 ).
The Surface Organic Loading rates (SOL) (gCOD·m À2 ·d À1 ) of the influent wastewater were calculated according to Equation (3).
where CS (carrier surface) is the total surface area (m²) available for biofilm growth in each MBBR (see Table 1).

Statistical analyses
Statistical analyses were performed using GraphPad Prism software (version 6.01, USA). The ANOVA and Tukey test were used to compare the influent and effluent wastewater quality with a significance level of 0.05.

Nitrification
The concentration of alkalinity, pH, and DIC (bicarbonate) is usually affected by the biological oxidation of the ammonia nitrogen. The hydrogen ion production by nitrogen oxidation can decay the values of pH, alkalinity, and DIC during the reactor operation (Metcalf & Eddy 2016;Guerrero & Zaiat 2018). Souza et al. (2018) checked that without the alkalinity addition to the raw wastewater, the concentration of nitrogen ammonia increased in the effluent of a nitrification system. Based on this finding, it is expected that the nitrification process is not effective in matrices with low buffering capacity (e.g., domestic wastewater) and the occurrence will decay the alkalinity significantly. However, the nitrification process was not observed in the three MBBRs by the results found for alkalinity, pH, and DIC ( Figure 2). The alkalinity values observed were 187 + 55 mg CaCO 3 ·L À1 for the influent and 191 + 53, 193 + 53, 186 + 44 mg CaCO 3 ·L À1 for R1, R2, and R3 effluent, respectively. The pH values found were 7.5 + 0.2 for the influent and 7.4 + 0.2, 7.5 + 0.2, and 7.4 + 0.2 for R1, R2, and R3 effluent, respectively. The DIC concentration observed was 50 + 4 mg·L À1 for the influent and 52 + 6, 56 + 11, and 53 + 4 mg·L À1 for R1, R2, and R3 effluent, respectively. No significant differences were found for alkalinity, pH, and DIC concentration observed in the influent and effluent from the three MBBRs (Tukey test, p . 0.05). These Uncorrected Proof findings are not in accordance with observed previously at nitrification bioreactors for domestic wastewater treatment that decreased pH value, alkalinity, and inorganic carbon (Yuan et al. 2020;Bressani-Ribeiro et al. 2021;Freitas et al. 2021). For efficient nitrification in attached biofilm, the organic matter concentration must be removed before the establishment of nitrifying organisms, once heterotrophic organisms have a higher biomass yield than nitrifying bacteria and it dominates the carrier's surface (Metcalf & Eddy 2016). Due to that, the organic matter concentration must be maintained low for ammonia nitrogen oxidation in reactors treating both organic and nitrogen compounds (Campos et al. 1999). In this study, the reactors were operated with a high COD/N ratio (7.15) once no pre-treatment was applied, which explained the absence of the nitrification process during the operation. A high concentration of organic compounds in the oxidation-nitrification zone leads to a competition for oxygen uptake in the biofilm between heterotrophic and autotrophic organisms for organic matter and ammonia oxidation, respectively (Azimi et al. 2007), with an advantage to first one over the nitrification process.
The absence of the nitrification process may also be caused due to the low HRT used in the tests. The MBBRs were operated at HRT of 2 h, which is lower than 7-8 h and 10-13 h required for the duplication of ammoniaoxidizing-bacteria (AOB) and nitrite-oxidizing-bacteria (NOB), respectively (Peng & Zhu 2006). Bassin et al. (2016) observed that the HRT reduction from 6 to 3 h had a negative impact on nitrogen oxidation, even though both organic and nitrogen loading rates were kept constant during the operation. Based on these results, a high concentration of COD and low HRT may impact negatively on the growth and attachment of nitrifying organisms on the carriers.

Organic matter removal
Organic matter is still the main pollutant of sewage in developing countries like Brazil (Brasil 2019). Hence, the main goal of domestic wastewater treatment is to reduce the organic loading before it is released into the environment. The influent and effluent COD concentration and the MBBRs performance over the operation time are shown in Figure 3.
A good performance was achieved since the beginning of MBBRs operation due to the fast growth of the biofilm on the carriers' surface, which is typically observed in aerobic reactors (Yang et al. 2018). The results found are in agreement with other studies that achieved stable biofilm growth and high COD removal (75-82%) at the beginning of operation (6-17 days) (Gonzalez et al. 2011;Dias et al. 2018). However, the overall COD removals observed in the three MBBRs (78-80%) were lower than ones reported for aerobic MBBR (85-88%) (Azizi et al. 2013;Zinatizadeh & Ghaytooli 2015). The slight difference in the performance may happen due to the limited hydrolysis of the particulate organic matter present in the influent, which depends on the amount of biomass present in the MBBR and the HRT applied in the bioreactor (Ødegaard et al. 2000). Then, the use of low HRT (2 h) difficult to achieve a better organic matter removal due to the negative effect on the hydrolysis process, as Uncorrected Proof previously observed by Wang et al. (2018). The increase of HRT could be a strategy to improve the COD removal. For instance, Dias et al. (2018) found the average COD removal of 88% at HRT of 19 h.
Nevertheless, the results found in this study are in agreement with other studies which evaluated the MBBRs performance by organic matter removal (Table 3). Wei et al. (2016) investigated three lab-scale MBBR for real wastewater treatment at different carrier filling ratios of 40, 50, and 60%. The COD removal of 70% was achieved by the three reactors and this performance was lower than obtained in this study (78-80%) with a carrier filling ratio of 50%.
It is important to note that all studies (Table 3) were carried out in MBBR with different volumes, wastewater, HRT, carrier filling ratio, type, and surface area of carriers. Besides these operational differences, similar performances were obtained. These results indicate that MBBR filling with alternative carriers without protected surface (HDPE flakes) can remove organic matter as well as an alternative (corrugated tube) or commercial carrier (K1) widely employed for wastewater treatment.

Volumetric and surface organic loading rates
The reactors were fed with the same raw domestic wastewater and influent flow rate, which results in the same VOL influent of 3.60 + 0.50 kgCOD·m À3 ·d À1 . The removal of 2.92 + 0.30, 2.93 + 0.35, and 2.86 + 0.40 kgCOD·m À3 ·d À1 were achieved for R1, R2, and R3, respectively. The VOL influent and VOL removal are shown in Figure 4. The better VOL removal occurred in the R2 filled with the corrugated tube with a close performance to those achieved by R1 or R3.

Uncorrected Proof
The VOL influent applied in this study (3.60 + 0.5 kgCOD·m À3 ·d À1 ) agrees with previous studies using MBBR for wastewater treatment. VOL values ranging from 0.8 to 5.2 kgCOD·m À3 ·d À1 were applied to MBBRs with COD removals ranging from 90 to 95% (Marques et al. 2008;Bassin et al. 2016). Despite the lower performance reached by MBBRs (78-80%) here, the mentioned studies used synthetic wastewater which is more easily degradable than real domestic wastewater used in this study.
It is usual to correlate the organic matter loading applied to MBBR with the total surface area of the carriers, due to its surface availability for the growth of attached biofilm (Ødegaard et al. 2000). The SOL influent and SOL removal are shown in Figure 5. Surface organic loading of 11 + 1.6, 19 + 2.7, and 7 + 1.1 gCOD·m À2 ·d À1 were applied to R1, R2, and R3, respectively. In this study, the SOL influent differences were due to the total surface area available by carriers in each MBBR (see Table 1). The surface organic loading removal of 8.7 + 0.9, 15 + 1.8, and 5.7 + 0.8 gCOD·m À2 ·d À1 were achieved by R1, R2, and R3, respectively. Vieira et al. (2014) evaluated a pilot-scale MBBR to treat pulp and paper mill wastewater and the SOL influent of 43.8 gBOD·m À2 ·d À1 was applied to the reactor. This value is much higher than the SOL influent of this study (7-19 gCOD·m À2 ·d À1 ), which can be explained by the filling ratio and wastewater quality. The authors used a filling ratio of 10% to treat a high COD influent of 1,384 mg·L À1 , leading to a reduced total carrier surface available and a high SOL influent. The SOL influent in this study are according to the values applied to the MBBRs (3.2-12.8 gCOD·m À2 ·d À1 ) (Bassin et al. 2016). The SOL influent extremely high can difficult the wastewater treatment by biological processes. Aygun et al. (2008) observed that the SOL increment from 6 to 96 gCOD·m À2 ·d À1 reduced the COD removal from 95.1 to 45.2%. Ødegaard (2006) recommended that SOL values should not exceed 20-25 gCOD·m À2 ·d À1 in MBBR, which corresponds to 65-85 gCOD·m À2 ·d À1 in typical wastewater in high-rate systems. The SOL values applied in this study are very lower than the recommended values, which makes the operation of the MBBRs used in this study a low-middle rate (values higher than 6 gCOD·m À2 ·d À1 ) (Aygun et al. 2008). Carriers with different specific surface and low substrate loading rates also lead to similar performance (Wei et al. (2016)), once a less specific surface is subjected to a higher SOL rate and also to a high COD removal (Ødegaard et al. 2000). These results confirm a high correlation between VOL applied, as well as SOL values, and the COD degradation rates achieved by the reactor.

Respirometric analysis of the biofilm
The respirometry assays aimed to evaluate the DO consumption by the attached biofilm to each type of carrier. The biofilm growth on the carrier surface leads to an increment in the fluid density and consequently, the residence time of the air bubble inside the bioreactor treating wastewater also increases (Collivignarelli et al. 2019). The DO consumptions by biofilm on the carriers are shown in Figure 6.
No significant DO variation was observed for wastewater filtered samples (blank) and the values ranged from 6.5 to 6.7 mg·L À1 . The DO concentration decreased from 6.73 to 5.28 mg·L À1 for K1 carrier, from 6.73 to 6.10 mg·L À1 for corrugated tube carrier, and from 6.82 to 6.5 mg·L À1 for HDPE flakes carrier. Then, the DO Uncorrected Proof consumption observed were 0.022, 0.009, and 0.005 mg of DO for K1, corrugated tube, and HDPE flakes, respectively. These values evidence the activity of the biofilm attached to the carriers, and also that the K1 carriers showed the highest surface oxygen uptake.
The surface DO consumptions of 0.0079 + 0.0013, 0.0033 + 0.0015, and 0.0031 + 0.0026 μgDO·mm À2 were observed for K1, corrugated tube, and HDPE flakes, respectively. HDPE flakes showed the lowest DO consumption mainly due to the lack of a protected surface. The protected surface area present in the corrugated tube and K1 carrier has more ability to be colonized by biofilm (Aygun et al. 2008;Torresi et al. 2017;Tombola et al. 2019).
In this work, the oxygen consumption demonstrated in the respirometry essay is due to the biological organic matter oxidation by heterotrophic microorganisms attached to the carrier surface (Metcalf & Eddy 2016). Then, it was expected a higher COD removal for K1 carriers due to the highest DO consumption. However, as previously shown, no statistical difference was observed for COD removal obtained in three MBBRs. This difference may happen due to irregular and heterogeneous growth of active biofilm on the carrier's surface (Denkhaus et al. 2007). Furthermore, the protected surface of carriers can also show weak correlations with the COD removal by biofilm under some conditions (Dias et al. 2018). Hence, it is not reasonable to assume that all available surface of carriers in each reactor was colonized by active microorganisms.
The results showed that the COD removal was independent of the oxygen uptake obtained in the tests. Moreover, the results found for alternative carriers without protected areas are similar to those reported in the literature for a carrier with a protected area (Azizi et al. 2013;Wei et al. 2016;Santos et al. 2020).

CONCLUSIONS
The present study investigated the organic matter removal from raw domestic wastewater by MBBR using carriers with (K1 and corrugated tube) and without (HDPE flakes) protected surface. The nitrification process was not observed during the operation of the bioreactors. The COD removal of 80 + 5.0, 80 + 3.5, and 78 + 2.4% was achieved by R1, R2, and R3, respectively. The oxygen uptake by biofilm attached on the carriers was 0.0079 + 0.0013, 0.0033 + 0.0015, and 0.0031 + 0.0026 μg DO·mm À2 for the K1, corrugated tube, and HDPE flakes, respectively. No significant differences were observed between the three MBBRs performance in terms of physico-chemical parameters (alkalinity, pH, and dissolved inorganic carbon) and COD removal. Thus, the results showed that the carrier type and its characteristics (total area and with/without protected area) did not affect the organic matter removal.