Treatment of Raw Mixed Dairy Wastewater Using an Attached-Growth Biological Filter ^†

Abstract

This study investigates the implementation of an attached-growth pilot-scale biofilter for the biological treatment of mixed dairy wastewater derived from real industrial effluents, consisting of equal proportions of raw second cheese whey (SCW) and pudding dessert wastewater (PDW). The biofilter was inoculated with indigenous microorganisms derived from the mixed wastewater stream with initial dissolved Chemical Oxygen Demand (d-COD) concentrations ranged from 1000 to 12,500 mg/L. The removal performance of organic and inorganic components was evaluated at a recirculation rate of 1.0 L/min, resulting in d-COD reductions of up to 92.3% and removal rates reaching 194.6 mg/(L·h). High removal rates were recorded for ammonium (up to 99.9%) and TKN (92.2–98.7%), while nitrate removal varied (29.4–89.3%) and solids removal exceeded 92%. d-COD concentrations of treated effluent consistently met discharge or municipal disposal legislation values, demonstrating the system’s efficiency and stability and proposing it as an ideal solution for wastewater treatment in dairy facilities.

1. Introduction

The dairy and confectionery production units are important branches of the food industry, using raw materials like milk, eggs, flour and sugar, to produce various products such as bakery goods, puddings, desserts, cheese, and yogurt [1]. Large volumes of wastewater are generated alongside these products, characterized by biodegradable organic matter, nitrogen, and other components. Although, confectionery wastewater is less toxic than other agro-industrial effluents, it contains high concentrations of biodegradable materials and other substances, like organic and inorganic forms of carbon and nitrogen, suspended and dissolved solids, oil and grease, starch, and sugar content [2], appointing it unsuitable for direct disposal without treatment. Similarly, milk-derived wastewater, particularly cheese whey (CW) and secondary cheese whey (SCW), contain high levels of organic matter, solids, and salts, and has an acidic pH [3]. Therefore, companies are forced to apply treatment to their produced wastewater in order to comply with the legislation, while illegal uncontrolled discharge of dairy wastewater into natural water bodies leads to eutrophication and unbalanced ecosystems.

In recent years, there has been extensive research into treatment technologies for confectionery and dairy wastewater, including physicochemical [4,5], electrochemical [6,7], and biological [2,8] methods. Another well known approach is the multi-stage hybrid systems [9,10], which combine the benefits of different methods to optimize results. The application of physicochemical coagulation, flocculation, or adsorption often results in the production of large amounts of sludge and insufficient reduction in the organic load [1]. Electrocoagulation methods and membrane reactors are more efficient at reducing pollutants, but they require high energy and maintenance costs [11,12].

The biological approach to dairy treatment has been studied for a long time [11]. Specifically, anaerobic biological reactors have been studied extensively because they effectively biodegrade organic content and generate valuable biogas [13,14]. As for aerobic biological systems, attached-growth systems demonstrate advantages compared to conventional activated sludge processes, such as stability under variable hydraulic and organic loadings, operational resilience, and reduced sludge production. Specifically, biofilters seem to be a promising approach based on attached-growth microorganisms that colonize a solid support medium [3]. This results in the formation of a dynamic biofilm capable of effective degradation. Biofilters present resistance and tolerance under wastewater composition and environmental conditions which are factors that must be taken into account when designing and implementing biofilters [3,15]. In addition, they are relatively simple to operate and maintain, suggesting a cost-effective and environmentally sustainable solution, especially for small- and medium-sized confectionery and dairy plants.

The main objective of this study was to evaluate the performance of a pilot-scale attached-growth biological system for the treatment of real mixed dairy wastewater consisting of two distinct effluent streams: second cheese whey and pudding dessert wastewater. This biofilter employed in this study has previously demonstrated high treatment efficiency for second cheese whey, even under elevated organic load [3]. Subsequently, the same system was successfully applied to the treatment of wastewater generated from pudding dessert production, achieving high removal efficiencies for organic load and nutrients [15]. However, to date, the application of an attached-growth biological system for the co-treatment of these two dairy wastewater streams in a single integrated process has not been investigated. The present study addresses this research gap by assessing the feasibility, efficiency, and operational stability of an integrated treatment approach. Co-treatment of the combined effluents is expected to provide significant advantages, including dilution of pollutant loads, improved biodegradability, enhanced process stability, and reduced overall treatment costs.

2. Materials and Methods

2.1. Set Up of Biological Pilot-Scale Attached-Growth Reactor

The column pilot-scale bioreactor, constructed from Plexiglas (160 cm height, 9 cm internal diameter, total volume 10.1 L) was positioned vertically on a wooden base and equipped with ten evenly spaced sampling points (Figure 1a). Hydraulic connections, including piping and valves, were located beneath the base to facilitate integration with a recirculation pump, water supply, sewage network, and a mechanical aeration pump. The system was operated in batch mode with recirculation (1.0 L/min) and aerobic conditions maintained via the mechanical air pump installed at the lowest sampling point (Figure 1b). The working volume of the reactor was 7.0 L, with hollow plastic tubes providing a specific surface area of 500 m²/m³ and porosity of 0.8 [3]. The reactor was initially operated in batch mode with mixed dairy wastewater, and following attachment and growth of indigenous microorganisms, experiments were conducted under non-sterile conditions at room temperature without pH control. It should be mentioned that no external inorganic or organic nutrients, such as nitrogen and phosphorus, were added during the experiments.

2.2. Wastewater Characteristics

The wastewater examined in this research work consisted of equal parts by second cheese whey (SCW) and pudding dessert wastewater (PDW) from a local dairy factory and confectionery industry, respectively, both located near the city of Agrinio, Greece. SCW derives from the production of “feta” cheese, while PDW comprises the washing waters from the equipment cleaning (using only water without cleaning chemicals) during the preparation of traditional Greek Pudding Dessert.

As previously described by Patsialou et al. [15,16], both PDW and SCW were characterized by a high organic load, elevated solids concentration, and a neutral to slightly acidic pH. The wastewater was collected immediately after production and stored at 2 °C in opaque plastic bottles for a few days prior to use. Variations in the composition of SCW and PDW were observed each time the wastewater was collected and attributed to different quantities of water used during the washing process of tanks and pipelines after cheese production and Pudding Dessert processing. Consequently, different proportions of SCW, PDW and tap water were mixed to obtain target d-COD concentrations of the mixed dairy wastewater of about 1000, 2500, 5500, 7500, and 12,500 (corresponding to experiments C₁₀₀₀, C₂₅₀₀, C₅₅₀₀, C₇₅₀₀ and C_12,500).

Table 1. Initial concentrations for all parameters investigated for all experiments.

Experiment	Initial Concentration of the Studied Parameters
	d-COD (mg/L)	BOD5 (mg/L)	PO43− (mg/L)	Sugars (g/L)	TSS (mg/L)	TDS (mg/L)	ΤΚΝ (mg/L)	NO3−-N (mg/L)	NH4+-Ν (mg/L)	Salinity (g/L)
C1000	1127 ± 36.5	750 ± 55	3.42 ± 0.07	0.90 ± 0.08	1040 ± 21	1093 ± 155	12.60 ± 1.98	4.08 ± 0.71	1.10 ± 0.11	1
C2500	2888 ± 42.2	2250 ± 212	6.19 ± 0.15	1.60 ± 0.01	1166 ± 69	3693 ± 103	21.88 ± 1.24	6.84 ± 1.4	3.60 ± 0.1	4
C5500	5739 ± 118	4350 ± 353	14.96 ± 0.01	4.66 ± 0.21	2043 ± 187	5113 ± 423	36.41 ± 1.98	14.02 ± 1.4	5.95 ± 0.2	5
C7500	7568 ± 45	5700 ± 707	21.01 ± 1.44	5.84 ± 0.10	3066 ± 292	4916 ± 485	73.54 ± 4.95	34.95 ± 1.44	10.08 ± 0.11	6
C12,500	12,596 ± 344	9600 ± 705	24.44 ± 3.28	11.76 ± 1.23	6843 ± 61	10,205 ± 149	86.14 ± 2.97	22.98 ± 1.90	8.39 ± 0.20	9

presents the main characteristics of the mixed dairy wastewater used (Table 1).

2.3. Analytical Methods and Procedures

Daily samples were collected for analysis during the experiments from the recirculation flow. Physicochemical parameters, including pH, dissolved oxygen (DO), and salinity were measured to ensure the correct operation of the biofilter, using a pH meter (Hanna pH 211, with HI1131 electrode; Hanna Instruments, Woonsocket, RI, USA), a DO analyzer (Hanna HI9146, with HI76407 electrode; Hanna Instruments, Woonsocket, RI, USA), and an analog refractometer (Type ORA 1SA; KERN & SOHN GmbH, Balingen, Germany), respectively. All samples were filtered prior to analysis with 0.45 μm-Millipore filters (GN-6 Metricel Grid 47 mm, Pall Corporation (Port Washington, NY, USA)), except for solids and BOD₅ determination.

Total Suspended Solids (TSS) and Total Dissolved Solids (TDS) were evaluated using a 0.45 μm pre-weighted membrane filter and with a precise sample of well known volume being filtered. The respirometric Method 5210 D (Continuous Oxygen Uptake), following the principles given by the WTW OxiTop system, was used for BOD₅ determination, in the dark, with stable temperature (20 °C) and continuous stirring. Organic components, like dissolved Chemical Oxygen Demand (d-COD) and Total Kjeldahl Nitrogen (TKN), were measured following the Closed Reflux, Colorimetric Method 5220 D and the Persulfate Digestion method SM 4500-N C, respectively. Total sugar content was quantified using the DuBois method [17]. Inorganic forms of nitrogen (NO₃⁻-N, NH₄⁺-N), as well as orthophosphates (PO₄³⁻) were determined using Method 4500-ΝO₃⁻ Β, the Salicylate method [18], and the Ascorbic acid method 4500-P E, respectively. All the aforementioned methods were used as described in ‘Standard Methods for the Examination of Water and Wastewater’ [19].

3. Results and Discussion

3.1. Biotreatment of Mixed Dairy Wastewater Under Different Initial d-COD Concentrations

Five different initial d-COD concentrations were examined, approximately 1000, 2500, 5500, 7500, and 12,500 mg/L. Figure 2a presents the d-COD removal over time for all experimental conditions. The shortest operation cycle (9 h) was achieved in the C₁₀₀₀ experiment, while the required processing time was comparable for C₂₅₀₀ and C₅₅₀₀ experiments (20 and 24 h, respectively). A further increase in initial concentration prolonged the operation cycle, reaching 36 and 60 h for C₇₅₀₀ and C_12,500, respectively.

Despite the variation in treatment duration, d-COD removal remained consistently high in all experimental series, ranging between 87.4% and 92.3%. Notably, in the C₁₀₀₀ experiment the final d-COD concentration was 88 mg d-COD/L, well below the limit for the disposal of treated wastewater to aquatic bodies (120 mg/L COD, [20]). Comparable findings were reported by Carta et al. [21] and Silva et al. [8], although their studies achieved such low effluent COD concentrations at considerably lower initial d-COD levels. Carta et al. [21] employed an attached-growth bioreactor for real dairy wastewater, a system with notable similarities to the present study, while Silva et al. [8] achieved final concentration below 100 mg COD/L using a hybrid biological anaerobic-aerobic system treating artificial dairy wastewater. It is worth mentioning that results meeting the legislation criteria for safe disposal are not usually reported in the literature, making these methods valuable for industrial applications. For higher initial concentrations, C₅₅₀₀, C₇₅₀₀ and C_12,500, the final effluent d-COD values recorded were 724.3, 938.4, and 1021.5 mg d-COD/L, respectively. These concentrations approach the legislative limit for disposal to the sewage system (1000 mg/L COD, [20]).

TKN removal was also high in all cases studied, ranging from 92.2% to 98.7%. Organic forms of nitrogen and ammonia proved to be highly biodegradable by mixed dairy microorganisms, with sharp removal in all experiments within the first hours (Figure 2b). These results exceed the TKN removal that have been recorded by Rezaee et al. [22], despite that the biological configuration presented by them consisted of a combination of aerobic/anaerobic/anoxic steps. Similarly, Hamdani et al. [23], using a vertical-flow fixed bed bioreactor, succeeded the optimum TKN removal of 87.5% after 96 h operating cycle of two anoxic/aerobic phases. This study’s TKN removal rates confirm the system’s high effectiveness, although it is a single-step method.

Regarding the d-COD degradation rate (Figure 2c), the maximum value of 194.6 mg/(L·h) was observed in the C₅₅₀₀ experiment. The degradation rates at C₇₅₀₀ and C_12,500 were very close to this maximum, reaching approximately 184 mg/(L·h) and 193 mg/(L·h), respectively. This indicates that beyond the initial d-COD concentration of 5500 mg/L, no appreciable variations in the degradation rate were observed. A maximum degradation rate at an initial d-COD of 5500 mg/L was also reported by Patsialou et al. [15] using an identical bioreactor configuration for PDW treatment. The degradations rates achieved in the present study also fall within the range reported by Tatoulis et al. [3], (i.e., 110–267 mg/(L·h), where the same type of reactor was applied for the treatment of SCW.

Salinity measurements of filtrated samples were also performed at both the beginning and end of each experiment, using an analog refractometer, as a NaCl range (‰). High salinity, up to 9‰, was recorded in the C_12,500 experiment, whereas initial salinities in the remaining experiments ranged from 1‰ to 6‰ (Table 1), reflecting the naturally high salt content of SCW. At the end of all runs, salinity was completely removed, even in the case of the highest initial value of 9‰. This reduction appears to be correlated with the substantial removal of solids (

Table 2. Final removal efficiencies (%) for all investigated pollutants for all experimental conditions.

Experiment	Removal (%)
	d-COD	BOD5	PO43−	Sugars	TSS	TDS	ΤΚΝ	NO3−-N	NH4+-Ν
C1000	92.2 ± 0.5	76.3 ± 0.7	97.2 ± 0.7	94.5 ± 1.4	88.1 ± 0.9	89.6 ± 2.1	92.2 ± 0.4	89.3 ± 0.9	96.2 ± 0.6
C2500	92.3 ± 7.4	91.4 ± 2.7	87.6 ± 0.1	98.8 ± 0.3	79.4 ± 1.7	83.7 ± 3.1	98.1 ± 0.8	45.5 ± 0.2	99.9 ± 0.1
C5500	87.4 ± 0.2	97.1 ± 1.0	44.5 ± 4.3	98.5 ± 0.8	91.2 ± 0.1	89.6 ± 1.0	94.2 ± 2.4	60.3 ± 5.3	92.9 ± 6.3
C7500	87.6 ± 0.9	96.6 ± 1.1	99.3 ± 1.5	98.3 ± 0.2	66.7 ± 4.5	64.4 ± 3.5	98.7 ± 0.2	86.7 ± 0.8	98.1 ± 0.5
C12,500	91.9 ± 1.0	85.7 ± 0.5	77.1 ± 1.5	96.6 ± 0.4	93.9 ± 0.4	92.8 ± 0.7	96.1 ± 3.0	29.4 ± 2.4	98.4 ± 0.3

), given the well established interrelation between TDS, conductivity, and salinity [24,25]. Similar observation regarding correlation between TDS removal and salinity reduction was recorded by Velmurugan and Pandian [26], using a microbial consortium treatment for dairy effluents. Also, biological processes contribute to salinity reduction by NaCl removal, like assimilation of dissolved minerals incorporated into new biomass or adsorption of ions Na⁺ on EPS. Dairy-associated bacteria can produce Extracellular Polymeric Substances (EPS) containing anionic functional groups with strong tendency to bio-sorb Na⁺. EPS production and composition change under salt stress, supporting bacterial communities to bind dissolved sodium in dairy systems [27,28].

3.2. Removal Efficiencies of Various Pollutants at the Initial d-COD Concentration of 5500 mg/L

As the degradation rates observed at initial d-COD concentrations of 5500, 7500, and 12,500 mg/L were comparable, the C₅₅₀₀ condition was chosen for further study (kinetic experiments), owing to its high degradation performance and shorter operational cycle. On this basis, kinetic studies of various parameters were conducted. Figure 3a,b show the removal profiles of organic carbon forms (d-COD and sugars) and inorganic nitrogen forms (nitrate nitrogen and ammonium nitrogen). Although high removals were achieved for all parameters sugars and ammonium nitrogen were rapidly depleted within the initial hours of the experiment, indicating their higher biodegradability and preferential utilization by the indigenous microbial community. Comparable findings for d-COD and sugar removal were reported by Patsialou et al. [16], in the treatment of raw SCW with marine microalgae and cyanobacteria.

BOD₅ removal was also substantial, reaching 97.1%, confirming the high biodegradability of this mixed wastewater stream. This value exceeds those reported by Patsialou et al. [15], who achieved 86.8% and 78.7% removal during PDW treatment with a biofilter. Organic nitrogen (measured as Total Kjeldahl Nitrogen, TKN) exhibited removal up to 94% (Figure 3c), higher than those recorded by Kushwaha et al. [6], 76.6%, at the same operational time of 24 h using an aerobic sequential batch reactor for artificial dairy wastewater treatment. In contrast, orthophosphate removal was limited to 44.5%, considerably lower than the PO₄³⁻ removals recorded in the other experimental series, which ranged between 77.1% and 99.3% (Table 2).

Finally, solids removal was consistently high, with values of 91.2% and 89.6% for TSS and TDS, respectively (Figure 3d). These values were comparable across all experiments (Table 2), demonstrating the robustness of the system in achieving efficient solids removal irrespectively of the initial organic loading. Patsialou et al. [15] similarly reports high TSS removal (93–96%) for PDW treatment with the same bioreactor, further confirming system’s reliability.

4. Conclusions

This research study evaluated the degradation efficiency of an attached-growth biological filter for the treatment of mixed dairy wastewater, including second cheese whey and pudding dessert wastewater. The filter was inoculated with indigenous microorganisms derived from the same wastewater stream. The aerobic system was operated in batch mode with recirculation under five different initial d-COD concentrations. In the C₅₅₀₀ experiment, the highest d-COD degradation rate of 194.6 mg d-COD/(L·h) was observed, occurring within the shorter operating cycle of 24 h. Under the same conditions, substantial removals of d-COD (87.4%), BOD₅ (97.1%), TSS (91.2%), and TDS (89.6%, respectively, were achieved. A similarly high degradation rate (192.9 mg d-COD/(L·h)) was recorded for the C_12,500 experiment, albeit over a longer operation cycle of 60 h, resulting in a d-COD removal of up to 92%.

Importantly, the final d-COD concentrations in all experiments were below the regulatory limit for discharge to the sewage system (COD < 1000 mg/L), with the C₁₀₀₀ experiment achieving an exceptionally low effluent concentration of 88 mg/L, meeting the stricter standards for direct discharge into natural water bodies (COD < 120 mg/L). Also, the total salinity removal works in addition to safe disposal of treated effluents. Thus, the proposed attached-growth biological system can be considered a sustainable and effective treatment solution for small-scale plants handling mixed dairy wastewater streams. The high removal efficiencies achieved for organic and inorganic compounds, combined with the low operational cost of the system, indicate an economically viable and environmentally friendly treatment option for dairy–confectionery industries.

Future research should focus on the further optimization and integration of the proposed biological system with complementary post-treatment processes in order to develop an efficient hybrid treatment scheme. The combination of the attached-growth biofilter with advanced polishing methods, such as adsorption or constructed wetlands, could enhance effluent quality and broaden the range of potential reuse or discharge options. In addition, molecular characterization of the indigenous microbial communities using metagenomic approaches would provide valuable insight into the microbial structure and functional dynamics governing system performance under varying operating conditions. Finally, a comprehensive assessment of all by-products generated during the biological treatment process is essential for the development of integrated waste management strategies and the sustainable valorization of residual streams.