SIMULATED DAIRY WASTEWATER TREATMENT IN A PILOT PLANT SCALE MAGNETO-ACTIVE HYBRID ANAEROBIC BIOFILM REACTOR ( MA-HABR )

The aim of this study was to determine the effects of magneto-active microporous packing media manufactured by extrusion technology and modified by the addition of relevant amounts of metal catalysts and magnetic activation on the effectiveness of simulated dairy wastewater (SDW) treatment and biogas productivity in a pilot plant scale hybrid anaerobic biofilm reactor with full mixing. The best performance was found at an organic loading rate (OLR) in the range 6.0 8.0 kg COD/m3·d, where the chemical oxygen demand (COD) removal was about 80%, the biogas yield ranged from 256.7 310.9 L/kg CODremoved, the methane production ranged from 420.6 557.1 L/d and total phosphorus removal ranged from 82.9 90.7%. The study demonstrated that the application of the innovative packing media in the biofilm-bed reactor enhanced the sorption of organic matter, biogas productivity and the binding of biogenic compounds.


INTRODUCTION
The dairy industry is one of the main agricultural productions in the world (Chokshi et al., 2016).In 2015, the production of milk in the European Union was approximately 168.2 million tonnes, of which 155.1 million tonnes were delivered to dairies (Eurostat, Statistics Explained).In Poland, the collection of milk by dairies in 2015 was 10.9 million tones (Eurostat, Statistics Explained).The indicator of waste water volume produced in dairies is about 0.2-10 liters per liter of processed milk with an average generation of about 2.5 liters of wastewater per liter of the milk processed (Sheteand Shinkar, 2013).The daily discharge of dairy wastewater in Poland is approximately 92,000 m3 (Gugała et al., 2015).Dairy processing effluents are generated in an intermittent way with variable flow rates, but generally are characterized by high strength with a COD concentration up to 95 g/L (Chokshi et al., 2016).Although the method of activated sludge wastewater treatment is an efficient solution and usually allows one to achieve the appropriate effluent quality discharged into the receivers surface waters.However, the extensive energy requirements and many technological problems due to the high organic load and variable quality of wastewater indicate the need to use different methods to reduce the treatment costs and intensify the wastewater treatment efficiency.Anaerobic digestion (AD) has been successfully applied in fullscale for the treatment of dairy wastewater (Tiwary et al., 2015).Several bioreactor types have been used; however, granule and biofilm-based reactors are the most common reactor configurations employed (Karadag et al., 2015;Rajagopal et al., 2013).
Anaerobic biofilm reactors are characterized by high loading capacity, resistance to hydraulic and organic overloads, high content of biomass, no requirement of mechanical mixing and short start-up (Ince et al., 2000;Haridas et al., 2005;Karadaget al., 2015).In anaerobic biofilms reactors treating dairy wastewater, microorganisms are supported on filtering materials such as seashells, charcoal, plastic elements, sintered glass, ceramic elements, natural stone and gravel (Ince et al., 2000;Najafpour et al., 2010;Qazi et al., 2011).However, the organic matter removal efficiency is related with the nature and properties of these materials (Karadag et al., 2015).An alternative to the currently used support materials is the use of packing media that enhance the sorption, precipitation and the binding of biogenic compounds.It has been shown that the effective removal of organic matter and phosphorus from dairy wastewater can be achieved by the introduction of metal ions formed during the corrosion of metal elements (Jędrzejewska-Cicińska and Krzemieniewski, 2010).Similarly, a constant magnetic field may significantly improve the efficiency of dairy wastewater treatment via AD (Krzemieniewski et al., 2004a-b).These were the basis for the development of the study on magnetoactive packing media for anaerobic biofilm reactor treated dairy wastewater, which established the value of magnetic induction (Zieliński et al., 2014) and the ratio of introduced metal ions (Zieliński et al.,in press).A new packing material produced by extrusion method was characterized by microspheres, which increase the active surface accessible to anaerobic microorganisms and catalytic properties due to the presence of metal ions (Cu and Fe) and magnetic activators.In laboratory and small scale, the application of the new magneto-active media in AD of dairy wastewater enhanced organic matter removal and phosphorus removal and improved biogas production (Zieliński et al., 2015;Dębowski et al., 2014).However, this innovative solution has not been used before in dairy wastewater treatment on a large scale.
The aim of this study was to determine the effect of the magneto-active microporous packing media manufactured by extrusion and modified by the addition of metal catalysts and magnetic activation on AD of simulated dairy wastewater (SDW) and biogas production in a hybrid anaerobic biofilm reactor with full mixing operating on a pilot plant scale.

Wastewater
In the present study, simulated dairy wastewater (SDW) was prepared by dissolving milk powder manufactured by Dairy Plant in Ostrowia Mazowiecka (Poland) per litre of distilled water to achieve COD concentrations according to the required OLRs: 5.0; 6.0; 7.0; 8.0; 9.0; 10.0 kg COD/m 3 •d.The method used for the OLR calculation was given by Zeeman and Lettinga (1999) and was based on the active volume of the reactor.Several investigators used the same method for making SDW (Leal et al., 2006;Kushwahaet al., 2010).SDW before feeding into the anaerobic reactor was kept at a temperature of 35ºC (±1ºC) in a hydrolysis tank.The wastewater characteristics used at each stage of experiment are presented in Table 1.

Magneto-active packing media preparation
The biofilm support active media used in the experiment was manufactured by microporous extrusion of granulated transparent and plasticised polyvinyl chloride (PVC).PVC, commercially available as Alfavinyl GFM/4-31-TR, was manufactured by Alfa Sp. z o.o.(Poland).Major parameters of the PVC used in the experiment are as follows: density 1230 kg/m 3 , Young's modulus 2600 MPa, tensile strength 21 MPa, elongation at break 300%, shore A-hardness 80 ºSh.When manufacturing the packing media, PVC was modified by the introduction of 0.8% wt granulated blowing agent Hydrocerol 530 (Clariant Masterbatch).Additionally, the PVC was blended with chemically pure copper and iron powder (Cometox).Each time 5.0% wt. of metal additives was introduced to PVC used for the production of packing elements.The copper to iron weight ratio was 1 Cu :9 Fe and was set following previous studies carried out by the authors (Zieliński et al., 2016).Characteristics of the manufactured active media are as follows: density 874±0.8kg/m 3 , shore A-hardness 24±0.4ºSh,tensile strength 138.1±1.8N, porosity 40±0.6%.Simulated dairy wastewater treatment in a pilot plant scale Magneto-Active Hybrid Anaerobic Biofilm Reactor (MA-HABR) Based on previous experiments, it was assumed that suitable magnetic properties can be achieved by the introduction of magnets to 25% of the plastic elements placed inside the reactor (Zieliński et al., 2015).Characteristics of the neodymium magnets used in the experiments are as follow: diameter 10±0.1 mm, height 5±0.1 mm, magnetic flux ~3952•10 -3 mWb, magnetic moment ~450.528•10 - mWb•m, magnetic field in the geometrical center of the surface of the magnetic pole with a distance of 0.7 mm ~0.368 T, magnetic field next to the edge of the surface of the magnetic pole (max.) with a distance of 0.7 mm ~0.384 T, pull force 0.85 kg.

Experimental station setup and operation
A pilot plant scale station for anaerobic treatment of SDW included a hydrolysis tank and a MA-HAB reactor (Fig. 1).The hydrolysis tank was a DPPL type container of 1000 L total volume and 600 L active volume, fitted with a vertical mixer operating at a speed of 30 rpm.Once every 48 h SDW was prepared in the hydrolysis tank.SDW was fed to MA-HABR with the frequency of 24 times per 24 h for 10 min, using a rotary lobe pump of 75 L/h capacity.
The MA-HABR was designed as a DPPL type container of 1000 L total volume and 600 L active volume, fitted with a vertical mixer operating at a speed of 60 rpm.The mixer worked in a 30 min ON and 30 min OFF cycle per hour.The MA-HABR was initially inoculated at a ratio of 40% (by volume).The inoculum originated from the closed fermentation tanks of a local municipal wastewater treatment plant.The concentration of volatile solids (VS) seeded into the reactor was 62.7% total solids (TS).The MA-HABR was filled with 200 L of an active medium.The density of the active medium was lower than that of the reactor contents (a mixture of anaerobic sludge and SDW), so during the mixing the tank was operated as a contact reactor with a fluidised bed (Fig. 1a).One cycle of MA-HABR reactor operating phase (1 h) consisted of 30 min of mixing and 30 min of reaction phase (without mixing).The reaction phase was divided into flotation of the media and the magneto-active filtering layer formation (10 min) (Fig. 1b), effluent discharge through the magneto-active filtering layer (10 min), filling with SDW (10 min).The temperature in the reaction chamber of the MA-HABR was 35ºC (±1ºC) and was maintained automatically.Biogas was simply vented into the atmosphere.
The MA-HABR after the start-up period was operated under an increasing OLR ranging from 5.0 kg COD/m 3 •d to 10.0 kg COD/m 3 •d (six stages of the experiment).This type of technological procedure allowed determination of the acceptable OLR in the MA-HABR.The reactor was operated to reach a steady state at every OLR tested (the steady state was defined as the state where the standard deviations of COD removal efficiencies were within 5%), a total of 229 days.The start-up period and adaptation of MA-HABR at a loading of 5.0 kg COD/m 3 •d lasted for a period of 37 days.Then, the concentration of biofilm adhered to the support material and suspended biomass inside the reaction chamber were calculated.The concentration of adhered biomass was 1.53 g TS/L of the active medium, while the suspended biomass was 3.75 g TS/L.

Analytical methods
The chemical oxygen demand (COD), total nitrogen (TN), ammonia nitrogen (AN), total phosphorus (TP), orthophosphate (P-PO 4 ) in the SDW and effluent were analysed once every 24 h using a DR 5000 spectrophotometer (Hach, Germany).The content of total solids (TS) was determined according to the gravimetric method.Determination of the amount of biomass immobilized on the surface of magneto-active packing media was based on the weight method.Before/after biomass immobilization, the mass of pure/immobilized packing media was determined.For this purpose, 30 weight measurements of the 10 randomly selected packing media were used.In case of immobilized biomass determination, the packing media was dryied at 105 ºC to evaporate water before weight measurement.TS of the immobilized biomass was determined by the difference between the average weight of the immobilized packing media and the average weight of the pure packing media.The tritation method (Tritlab AT 1000, Hach, Germany) was used for the determination of FOS/TAC ratio (volatile organic acid and buffer capacity ratio).
Biogas flow rate was measured continuously using a digital gas flow meter XFM17S (Aalborg Instruments & Controls, Inc., USA).The biogas composition was analysed once every 24 h using a GMF 430 meter (GasData, England) and a gas chromatograph (GC, 7890A Agilent, USA) equipped with a thermal conductivity detector (TCD).The GC was fitted with two Hayesep Q columns (80/100 mesh), two molecular sieve columns (60/80 mesh) and a Porapak Q column (80/100) operating at a temperature of 70ºC.The temperature of the injection and detector ports were 150ºC and 250ºC, respectively.Helium and argon were used as the carrier gases at a flow of 15 mL/min.
Statistical analysis of results was carried out using the STATISTICA 10.0 PL software package.The hypothesis of the distribution of each analysed variable was verified using the W Shapiro-Wilk test.One-way analysis of variance (ANOVA) was used to determine the significance of differences between variables.The homogeneity of variance in the groups was analysed using Levene's test.The RIR Tukey test was used to determine the significance of differences between the analysed variables.The statistical significance was adopted at p = 0.05.Simulated dairy wastewater treatment in a pilot plant scale Magneto-Active Hybrid Anaerobic Biofilm Reactor (MA-HABR)

Organic matter and nutrients removal
Most hybrid reactors treating dairy effluents have been operated in lab-scale systems (Karadag et al., 2015).In our study the HABR was operated on a pilot plant scale and contained an innovative magnetoactive packing media.
The results showed that a pilot plant scale MA-HAB reactor has a high potential in organic matter and phosphorus removal from SDW.The start-up period and adaptation of MA-HABR at a loading of 5.0 kg COD/m 3 •d was 37 days (Table 2).During the MA-HABR operation, the highest COD removal efficiency of 75.8 -77.5% was achieved under the OLR range of 6.0 -7.0 kg COD/m 3 •d (Fig. 2, Table 2).Increase in the OLR gradually deteriorated the COD removal efficiency; however, the COD removal achieved in stage 5 (OLR of 9.0 kg COD/m 3 •d) was comparable to that at the lowest OLR of 5.0 kg COD/ m 3 •d.The removed COD loading reached the maximum level of 3.5 -3.6 g/d under the OLR of 9.0 -10.0 kg COD/ m 3 •d.The lowest removed COD of 2.0 g/d was achieved in the first stage of the experiment.The efficiency of TS removal exceeded 94% in all stages, and was the highest (over 97%) when the OLR ranged from 7.0 -10.0kgCOD/ m 3 •d.Throughout the study, the concentration of TS in the effluent remained at 0.38 -0.47 g/L.
The highest TN removal of 21.5 -26.1% was achieved when the OLR ranged from 6.0 -8.0 kg COD/m 3 •d (Fig. 2, Table 2).Further OLR increase resulted in nitrogen removal deterioration below 15%, which was similar (p>0.01) to the value achieved at the first stage of the experiment.As regards the phosphorus removal, it exceeded 70% throughout the study (Fig. 2, Table 2).At OLR in the range of 5.0 -8.0 kg COD/m 3 •d, the outflow TP concentration was 12-18 mg/L, while phosphorus removal was over 83%.Clogging was not observed during the operation time, and the catalytic properties of the packing media were not reduced.In the presented study, an innovative active medium made of PCV elements modified with metal catalyst containing Cu and Fe was used.The results demonstrated that the addition of catalysts allowed for the effective removal of organic matter (as COD over 58%) and phosphorus (over 70%).Moreover, dairy effluent was discharged from the reactor via a filtering layer formed by a magneto-active packing media that created an additional surface for the growth of microorganisms, and can play a significant role in the removal of solids suspended in wastewater by prolonging the HRT necessary for the biodegradation of complex organic compounds (Haridas et al., 2005).Thus, the effectiveness of TS removal was over 94% during the study.
Similar conclusions were made by Jędrzejewska-Cicińska and Krzemieniewski (2010), who used steel elements to fill an UASB reactor during the treatment of dairy wastewater.Depending on the applied OLR, the efficiencies of phosphorus and COD removal were respectively higher by 16.4 -68.1% and 1.0 -3.1% in the reactor filled with steel elements as compared to performance rates achieved in a conventional UASB system.Zhang et al . (2011) reported a 53% increase in the COD removal in an UASB reactor filled with iron elements as compared to the efficiency noted in a digester without such elements.Shi et al. (2011) observed an 8.7-14.9%increase in COD removal and an over 68.5% increase in phosphorus removal in an UASB reactor supplemented with zero-valent iron (Fe 0 ).Wu et al. (2015) achieved a 70.9% increase in phosphates removal and a 27.6% increase in COD removal from swine production wastewater using a treatment system supplemented with zero-valent iron (ZVI).Jeon et al. (2003) improved the removal of organic matter and phosphorus from municipal and household wastewater by using steel rings to fill an Upflow Anaerobic Fluidised Bed Reactor (UAFBR).
Corrosion of metal elements stimulates the biochemical degradation of organic compounds by modification of the anaerobic environment inside the reactor through a reduction in the oxidationreduction potential (ORP) and an increase in the buffer capacity (Zhang et al., 2011;Wu et al., 2015).It also enhances abiotic processes of contaminant removal, because the formed colloids (i.e., Fe(OH) 2 and Fe(OH) 3 ) help to eliminate organic suspended solids in flocculation, adsorption and precipitation reactions (Jeon et al., 2003;Noubacept, 2008).Moreover, insoluble vivianite compounds are formed in the reaction of metal ions with phosphates, which reduces phosphorus concentration in the treated wastewater (Shi et al., 2011;Wu et al., 2015).Similar processes can be used for the removal of nitrogen compounds.Ammonium ions can form complexes with iron ions, which leads to a reduction in the concentration of total nitrogen in treated wastewater.However during the study, the TN removal was not high (11.1 -26.1%).The major environmental factors influencing methanogenesis are the volatile fatty acids (VFAs), ammonia nitrogen and pH (Fotidis et al., 2013).During the study, AN concentration in the effluent increased along with increasing OLR.Under the OLR of 5.0 -8.0 kg COD/m 3 •d it ranged from 171.1 to 263.1 mg AN/L.AN reached the maximum value of 318.5 mg/L when the OLR was 9.0 kg COD/ m 3 •d.At the last stage of the experiment it was 302.3 mg AN/L.During the 229 days of operation time, the pH was maintained at 6.8 -7.3 when the OLR was 5.0 -7.0 kg COD/m 3 •d, and then it was dropped to pH 6.7 under the OLR of 8.0 -10.0 kg COD/m 3 •d (Table 2).FOS/TAC ratio was determined at each stage of experiment to evaluate the stability of the anaerobic degradation process (Table 2).The FOS/TAC ratio changed between 0.31 and 0.44.During anaerobic digestion of dairy wastewater, the nitrogen in proteins is converted to inorganic forms, including ammonia.Ammonia, especially free ammonia, is the major toxicant in anaerobic digesters and its concentration over 4 g/L may inhibit methanogenesis (Westerholm et al., 2011).According to Chen et al. (2008), an ammonia concentration below 200 mg/L is beneficial for anaerobic microorganisms as a nitrogen nutrient source.An increase in pH would result in increased ammonia toxicity; thus, the most favorable pH for methane production ranged from 6.5 to 7.5 (Lili et al., 2011).In our study, the concentrations of AN were below inhibiting concentrations and the maximum pH value was 7.3.Thus, AN was at a safe level at the all stages of experiment.Moreover, AD instability due to ammonia often results in VFAs accumulation, which leads to a decrease in pH.There was no acidification in our study, and the FOS/TAC ratio showed an optimal value ranging from 0.31 to 0.37 when the OLR ranged from 5.0 to 8.0 kg COD/m 3 •d.When the OLR increased to 9 -10 kg COD/m 3 •d, the FOS/TAC ratio increased to over 0.4 and biogas yield was suddenly reduced, which reflected the overloading of the digester.

Biogas and methane production
At OLR ranging from 5.0-7.0 kg COD/m 3 •d, the biogas yield achieved about 300 L/kg COD removed and was the highest during all the study (Fig. 3, Table 2).The use of OLR of 9.0 kg COD/m 3 •d significantly reduced the Simulated dairy wastewater treatment in a pilot plant scale Magneto-Active Hybrid Anaerobic Biofilm Reactor (MA-HABR) biogas yield below 180 L/kg COD removed.(Fig. 3).High volumetric biogas production rate (VBPR) was noted at OLR from 6.0 -8.0 kg COD/m 3 •d; however, the VBPR reached the maximum value of 966.8 L/d at an OLR of 7.0 kg COD/m 3 •d (Fig. 4, Table 2).Similarly, the highest volumetric methane production rate (VMPR) was noted at OLR from 6.0 -7.0 kg COD/m 3 •d (Fig. 4, Table 2).Methane content in the biogas was on the average 61 -68% at OLR from 5.0 -7.0 kg COD/m 3 •d and decreased to 47% at higher OLRs (Table 2).
COD removed was achieved with the OLR of 9.5 kg COD/ m 3 •d.An anaerobic hybrid reactor with plastic cut rings for biofilm development was effectively used for the treatment of dairy wastewater by Banu et al. (2008).The methane production varied from 280 to 320 L CH 4 /kg COD•day and the methane content in biogas reached 60-65% under OLR from 8.4 kg COD/m 3 •d to 19.2 kg COD/m 3 •d.Strydom et al. (1997) treated effluents from three dairy factories (cheese, fresh milk and milk powder/butter factories) using an anaerobic hybrid reactor.The methane yield varied from 287 to 359 L CH 4 /kg COD removed at OLR between 0.97 and 2.82 kg COD/m 3 •d.In our study, the maximum biogas yield of 310 L/ kg COD removed was achieved at OLR of 6.0 kg COD/m 3 •d.
The corrosion of the magneto-active microporous mediamay affect methanogenesis.According to Karri et al. (2005), during anaerobic digestion, metal catalysts such as zero-valent iron are electrondonors for methanogenic and denitrifying bacteria.Under anaerobic conditions, iron corrosion is triggered by gaseous H 2 , and the formed CO 2 is reduced to CH 4 by hydrogenotrophic methanogens such as Methanococcus thermolithotrophicus, Methanobacterium thermoautotrophicum, or Methanospirillum hungatei (Wu et al., 2015).Shi et al. (2011) used iron nanoparticles to supplement the UASB reactor and obtained an increase in biogas productivity and methane content in biogas (by 12.9-17.9%and 10.7-12.9%,respectively).In a study of an UASB reactor filled with iron elements, Zhang et al. (2011) achieved a 66.8% content of methane in biogas versus 47.9% in biogas generated in an UASB reactor without elements.An increase in the efficiency of methanogenesis in reactors filled with steel elements was also reported by other researchers (Shi et al., 2011;Liu et al., 2015).
Moreover, the magnetic properties of the active packing media used in the present study may enhance the metabolic activity of anaerobic microorganisms in the sludge.The earlier study showed that the value of magnetic induction of 0.38 T positively affected organic matter removal and biogas production (Zieliński et al., 2014).According to Ji et al. (2010), a suitable level of magnetic induction not greater than 17.8 mT stimulates microbial activity, and thus improves the efficiency of wastewater treatment.For example, Krzemieniewski et al. (2004b) applied a stable magnetic field with induction of 0.4-0.6T and achieved increased efficiency of COD, ammonia nitrogen and orthophosphates removal from municipal, household and dairy wastewater (25-55%, 50-66%,  In recent years, anaerobic biofilm reactors have been widely applied for the treatment of dairy wastewater with the OLR ranging from 0.83 kg COD/ m 3 •d to 31.0 kg COD/m 3 •d (Karadag et al. 2015).High-strength milk permeate wastewater from a cheese making process was treated under mesophilic conditions using a bench-scale anaerobic moving bed biofilm reactor (AMBBR) under OLR from 1.5 kg COD/m 3 •d to 11.0 kg COD/m 3 •d (Wang et al. 2009).The maximum methane yield of 313 L CH 4 /kg 87-90%, respectively) when compared to removal rates in the system without magnetic field assistance.The magnetic field also modifies the properties of fluids, such as viscosity, enthalpy or surface tension (Toledo et al., 2008), which may result in more effective migration of gaseous metabolites of anaerobic microorganisms in the sludge from the reactor.

CONCLUSION
Anaerobic biofilm-based reactors are the most common reactor employed for dairy wastewater treatment.However, the main disadvantages of these reactors are the loss of the reactor performance due to the clogging of the filter material and only partial removal of non-organic pollution such as nutrients.There is a need to search for new biofilm support materials which could enhance the sorption, precipitation and the binding of biogenic compounds.This study demonstrated that the use of an innovative magneto-active packing medium ensured a stable organic matter removal (over 70%), phosphorus removal (over 80%) and biogas production (over 800 L/d) at OLR ranging from 6.0 to 8.0 kg COD/m 3 •d.The method of discharge of the effluent through the magneto-active layer enhanced the TS removal and reduced bacterial biomass losses, without clogging of the reactor.A significant decrease in biogas/methane production was found when the MA-HAB reactor was operated at an OLR greater than 8.0 kg COD/m 3 •d; however, organic matter and phosphorus removal were still high.

Table 1 .
Characteristics of SDW used in the study.

Table 2 .
Mean values of the reactor monitoring data reached at each stage of the experiment.