One of the most reliable methods for dairy effluent purification is biological removal. Such methods can assimilate all dairy wastewater components but they mostly utilise soluble compounds and small colloids. These processes have not been fully studied. Moreover, because of their unlimited adaptation potential, they can be jointly used in various sequences to meet certain component biodegradation requirements (1, 7). Biological treatment has two main branches depending on oxygen requirements: aerobic and anaerobic processes (66).

Aerobic processes. Nowadays, most dairy wastewater treatment plants are aerobic although they have been less efficient, mainly due to filamentous growth and rapid acidification caused by high lactose levels and low water buffer capacity, respectively (4, 12, 67). Problems generally encountered with activated sludge processes are bulking and foaming, which diminish sludge settling, Fe3+ and CO32– precipitation, additional biomass production as well as poor activity at low temperatures. It takes a few months for the sludge adaptation before full operational capacity is reached. Nitrogen from NH3 is easily degraded. Phosphorus removal is less effective and relies on environmental conditions. Aerobic bacteria are less useful in colloid utilisation when compared to anaerobic bacteria. The heightened O2 depletion (>3 kg of O2 per kg of BOD5) requires large energy demands during the aerobic treatment of concentrated dairy wastewater (>2 g of COD per L) (1, 4). Plug flow systems are better than complete-mix processes since they are less sensitive to high organic load problems like bulking sludge, etc. (21). Commonly, dairy effluent OLR, expressed as BOD5, should be less than 0.28–0.30 kg/m3. To enhance biological removal, a proper pretreatment or adequate wastewater dilution should be applied (1, 68).

Aerobic biological systems give a very positive response during synthetic dairy wastewater treatment with 4 g/L of COD and 1 g/L of TKN at pH=11.5, with over 96% of degradation being achieved in a continuous mode (69). An artificial effluent similar to milk powder and butter processing wastewater was treated in an anaerobic- -anoxic-oxic system at HRT of 7 days and a nominal sludge age of 20 days (70). The process was characterised by sludge bulking due to the growth of filamentous bacteria (Sphaerotilus natans, Type 0411 and Haliscomenobacter hydrossis). TN removal remained unchanged at 66% without the improvement in the sludge volume index. TP depended on the anoxic selector relative dimensions (from 49 to 20%) and a respective nitrate rise in the effluent. Nevertheless, more than 90% of COD reduction was achieved.

Aerobic filters are applied to a lesser extent in the treatment of high-strength dairy effluents rich in FOG. High fat and heavy biofilm blockage are possible, which results in biomass loss, filter fouling and corresponding reduction in productivity (1).

The sequencing batch reactor (SBR) is preferred in dairy wastewater treatment because of its various loading capabilities and effluent flexibility. A traditional technology with free sludge flocs is mostly applied. The purification of milk effluents is given by Britz et al. (1). COD was reduced by 91–97, TS by 63, volatile solids (VS) by 66, TKN by 75, and TN by 38%. However, mechanical treatment had to be applied first. Another study shows the aerobic SBR as an excellent example of the combination of activated sludge granulation with dairy effluent treatment (71). Granulation stability is limited by nutrient concentration in the wastewater, while effluent quality depends on the need for preliminary sludge settling, usually 0.25–0.5 HRT. Up to 90% of total COD, 80% of TN and 67% of TP were reached in an 8-hour cycle and 50% volume exchange ratio. The results were obtained after fully activated sludge granulation and consecutive biomass sedimentation. The soluble effluent COD was reduced to 125 mg/L. Industrial effluents are more difficult to treat than synthetic ones. The lower maximum OLRs also reduced the SBR granular sludge efficiency (17). In a bench-scale SBR, raw industrial dairy wastewater was treated with Lactobacillus casei TISTR 1500 (62). Microaerobic conditions maintained in the SBR allow for biomass accumulation in large amounts, leading to 85% lactose reduction via rapid fermentation and subsequent protein coagulation by 90%. As a consequence, 70% of COD degradation can be achieved. Around 2.67 times higher OLR was achieved in two laboratory aerobic SBRs treated with a mixed landfill and dairy effluent than in traditional SBR processes (71). The best BOD5 removal mode was reached at OLR, expressed as BOD5, of 0.8 kg/(m3·day) per a 10- -day HRT. The application of flexible fibre as an activated sludge carrier increases the laboratory SBR reliability and it is possible to treat dairy effluents at very high OLRs. At OLR, expressed as COD, of 0.4 kg/(m3·day), COD was degraded by more than 89% and up to 97% at OLR, expressed as COD, of 2.74 kg/(m3·day) (72). Membrane technologies are successfully applied in the treatment of low-load dairy effluents in an SBR. A high BOD removal (over 97%) and TSS-free wastewater are obtained. Due to low influent loading, TN removal reaches 96% by means of assimilation only. TP elimination reaches only 80% after system optimisation due to the limited excess sludge disposal (73).

Moving bed biofilm reactor (MBBR) shows very high performance when applied to dairy wastewaters: OLR increases dozens of times compared to conventional activated sludge systems. A milk processing effluent was treated in a MBBR with biomass developed on FLOCOR- -RMP® particles (Henderson Plastics Ltd, Norfolk, UK) (74). At OLR, expressed as COD, of 5 kg/(m3·day), more than 80% of total COD degradation was achieved in almost half-order kinetics with partial substrate penetration. TN was decreased by 13.3–96.2%. The small reactor volume and the high OLR encompass process applications including plant renovation and the introduction of new, limited-space treatment facilities (74). A novel MBBR with free-floating plastic elements (with a density slightly less than 1.0 kg/m3) may give 85 and 60% COD reduction at OLRs of 12 and 21.6 kg/(m3·day), respectively. On the basis of test results, we can say that the MBBR should be very suitable for the treatment of dairy industry effluents (75).

Good results can be reached in a membrane bioreactor during the treatment of an ice-cream factory effluent with 13.3 kg/m3 of COD, 6.5 kg/m3 of BOD5 at a temperature of 25 °C. Both indicators are reduced by over 95%, while TKN is decreased by more than 96 and TP by 80%. Under aerobic conditions, the indigenous microflora composed of lactic acid bacteria may reach over 109 CFU/mL, which will downgrade CIP-induced alkaline pH variations (76).

Various alternatives for aerobic treatment of dairy effluents are also used. Pure oxygen is another possibility in the biodegradation of milk wastewater. Oxygen can be applied directly in the homogenisation tank during a traditional physicochemical treatment and stable operation is achieved under a broad initial COD and TSS range. This modification improves effluent quality and reduces process costs. Such oxygen injection systems can replace the expensive anaerobic treatment and are naturally safer (77). Cheese whey can also be successfully utilised as a cheap medium for edible mushroom cultivation. Some authors report the growth of Ganoderma lucidum on protein-free cheese whey. The best soluble COD utilisation was achieved at pH=4.6 and 27.1 °C, while the maximum mycelial yield of 0.35 mg per mg of soluble COD removed was obtained at pH=4.2 and 28.5 °C (78). Although there is information on edible fungal growth, dairy wastewater utilisation has not been studied from a COD point of view (79–82).

Cheese whey effluents can be treated successfully in municipal wastewater treatment plants. Factories with onsite treatment technologies should collect sanitary wastewater independently from processing effluents and discharge them directly into municipal wastewater treatment plants. Nevertheless, such a treatment option can lead to operational problems with secondary treatment units (1, 12). Periodic sludge bulking is possible and is caused by intermittent high soluble COD levels in the receiving sewage plant.

Anaerobic processes. Anaerobic systems are more suitable for the direct utilisation of high-strength dairy wastewater and are more cost-effective than aerobic processes. If properly operated, these systems do not produce unpleasant odours (1, 4). The major problems of anaerobic dairy wastewater treatment include long start-up periods due to complex substrate degradation, preliminary biomass adaptation prior to protein and fat utilisation, fast drop in pH and a resultant inhibition of methane production (as a consequence of the high concentration of easily fermentable lactose and low substrate alkalinity), sludge disintegration by fats in the form of triglyceride emulsions and subsequent biomass flotation, presence of inhibitory compounds (long-chain fatty acids, K+ and Na+ ions), inability of ammonia biodegradation and phosphorus removal, careful management, increased sensitivity to various OLRs and shock loadings, etc. Notwithstanding the little information on industrial-scale anaerobic plants utilising cheese whey, more than 75% COD removal and around 10 kg/(m3·day) of OLRs, expressed as COD, are achieved. The degree of biodegradation depends on the HRT applied (4, 12, 22, 83–85).

Milk processing effluents are predominantly treated in conventional one-phase systems: upflow anaerobic sludge blanket (UASB) reactor and anaerobic filter (AF) are most commonly applied (4). UASB reactors have been used in industrial dairy wastewater treatment for more than 20 years. They are suitable for treatment of overloaded effluents with COD higher than 42 g/L (86). Laboratory scale UASB reactors utilising whey permeates in a continuous regime have been designed (87). Kinetic coefficients using the Monod equation are determined per HRT of 0.4–5 days and an initial wastewater COD of 10.4–0.2 g/L (87). It was shown by a comparative study of the possibility of using flocculent sludge and the effect of different HRTs (6–16 h) on the anaerobic UASB reactor behaviour applied to dairy wastewater treatment that nearly 80% of protein mineralisation, soluble COD and volatile fatty acid degradation as well as over 60% fat removal can be reached at an HRT of at least 12 h and an OLR, expressed as COD, of less than 2.5 g/(L·day) (88). Biomass granulation was also achieved in the UASB reactor within 60–70 days. Of all the elements studied, only Ca2+ ions had any significant effect (89). When treating a synthetic ice-cream effluent in the UASB reactor, TOC was reduced by 86% at an HRT of 18.4 h, with the highest OLR, expressed as TOC, reaching 3.06 kg/(m3·day) (1). High FOG degradation is also possible in an UASB reactor. A couple of bench-scale UASB reactors were successfully employed during the utilisation of a synthetic milk effluent rich in FOG (0.2, 0.6 and 1 g/L) (90). Enzymatic pre-hydrolysis contributed to 8% more COD removal at the highest FOG concentration (90). Cheese effluents are degraded in UASB reactors in laboratory tests and on an industrial scale. A laboratory-scale UASB reactor utilising a cheese factory effluent eliminates around 90% of effluents at an OLR, expressed as COD, of 31 g/(L·day) (91). Organic loads, expressed as COD, over 45 g/(L·day) perform worse (70–80% only). Moreover, chemicals are needed to support a constant pH. Short-shock OLR during operation increases sludge granulation, improving stability in reactor performance. The results of the laboratory tests on an industrial level have been confirmed (1), improving them by 6% per 10% higher load. A full-plant UASB reactor can be applied in cheese factory wastewater treatment. With an initial COD of 33 g/L, HRT of 16 h and OLR, expressed as COD, of 49.5 kg/(m3·day), 86% degradation can be reached. During the utilisation of an industrial effluent from Edam cheese, butter and milk production, a full-scale UASB reactor can be applied, the COD being decreased by 70% (1).

Dairy effluents with a low TSS can be successfully utilised in AFs in an all-scale range. The COD decreased by between 60 and 98% at a HRT of 12–48 h and an OLR, expressed as COD, of 1.7–20 kg/(m3·day) (1). A large specific surface of the filter media creates a precondition for higher biomass accumulation which is less affected by shear stress. A five-time higher load than with the non- -porous filler under the same conditions is achieved. It has been reported that with a couple of mesophilic upflow AFs utilising a milk bottling effluent, the reactor with the porous packing performed better (OLR, expressed as COD, of 21 kg/(m3·day)) than the same reactor with non-porous packing (OLR, expressed as COD, of 4 kg/(m3·day)), which is influenced by shear stress to a greater extent (92). Different temperature regimes can be analysed during the treatment of dairy wastewater in laboratory upflow AFs. At 12.5, 21 and 30 °C and HRT of 4 days on average, the COD removal in each reactor amounted to 92, 85 and 78%, respectively (93). An AF was used to treat ice-cream wastewater in a comparative study with contact process, UASB reactor and fluidised bed bioreactor (FBB) (94). The data showed a COD removal of 67, 80 and 50% at OLR, expressed as COD, of 6, 1 and 2 kg/(m3·day) and 60% of total COD removal, at OLR, expressed as COD, of 2–4 kg/(m3·day). All reactors had a poor biomass retention resulting from FOG loading. An upflow AF performed better, which allowed its full-scale installation in the manufacturing process (94). An upflow AF has been claimed to be unsuitable for the anaerobic digestion of very dilute dairy wastewaters (95). In fact, continuous stirred-tank (CSTR), UASB and baffled reactors also cause problems although experimental data show that the baffled reactor performs better with an OLR, expressed as VS, of 0.117–1.303 g/(L·day) and HRTs between 18.8 and 2 days.

Although a CSTR is a good option for scientific research of complete-mix systems (96), it is difficult to use it on an industrial scale because of HRT restrictions. Such reactors were studied with a cheese effluent consisting of wash water/whey ratio of 4:1 with 17 g/L of COD. However, problems with sludge loss arise if the HRT drops to below 9 days (1).

Milk processing effluents can be treated in hybrid systems too (4). An anaerobic contact digester may reach a COD degradation of over 80–95% under mesophilic conditions. The main disadvantage is the difficult sludge settlement. However, the technology is applied worldwide in dairy plants although it is quite old (1). A laboratory-scale experiment analysed the kinetic performance of anaerobic synthetic ice-cream effluent at 37 °C applying the Monod and Contois equations at an HRT range between 2.99 and 7.45 days. A better explanation of the kinetic coefficients can be achieved in the final pilot-scale plant since it allows variations in the initial substrate concentration (97).

Anaerobic packed-bed bioreactor (PBB) can be successfully applied for dairy wastewater treatment of various organic loads. A downflow PBB was used for treating deproteinised cheese whey with 59 g/L of COD (1). At OLR, expressed as COD, of 12.5 kg/(m3·day), the system decreased the COD to 90–95% at HRT of 2–2.5 days. The influent pH was around 2.9, while the pH in the reactor was almost neutral. Good results were obtained in a pilot--scale plant with an up-flow anaerobic PBB (98). The initial cheese whey COD was 59.4 g/L. A 16-hour HRT was enough to reach 99.4% of lactose conversion. Whey wastewater was degraded to 89% in an anaerobic MBBR at (35±2) °C per 1-day HRT and an OLR, expressed as COD, of 11.6 kg/(m3·day) (99). The cheese whey was decomposed in a laboratory PBB with a polyethylene carrier. The highest COD reduction was achieved at a 3.5-day HRT with OLR, expressed as COD, of 3.8 kg/(m3·day) and biogas production of 0.42 m3 per kg of COD per day (1). The mesophilic anaerobic fluidized-bed bioreactor system degraded 5.2 g/L of COD in the ice-cream wastewater to 94.4% at 35 °C, OLR, expressed as COD, of 15.6 kg/(m3·day) and HRT of 8 h. Under shock loading, the return to steady-state conditions was possible within 6–16 h (100). The fluidized-bed bioreactor was used to treat a low-load milk effluent with 0.2–0.5 g/L of COD. At an 8-hour HRT, 80% of COD was removed (1).

Membrane applications in anaerobic systems are good options for improved effluent filtration combined with a higher concentration and an effective differentiation between HRT and solids retention time. A completely mixed anaerobic microfiltration membrane reactor system was used on cheese whey high in COD (63 g/L) (1). More than 99% of organic matter was utilised when HRT was 7.5 days, which allowed authors to upgrade the studies from the pilot plant to a full-scale demonstration. The application of the ultrafiltration system made it possible to achieve a higher biomass retention for more efficient wastewater treatment.

Different temperature conditions have been tested in order to reach a higher COD anaerobic removal. The psychrophilic anaerobic operation in some laboratory hybrid reactors, utilising whey effluents with low (COD of 1 kg/m3) and high (COD of 10 kg/m3) load, showed a better COD performance when the OLR reached 70–80% in the first reactor (at OLRs, expressed as COD, of 0.5–1.3 kg/(m3·day), in a 20–12 °C range) and more than 90% in the second (at OLRs, expressed as COD, up to 13.3 kg/(m3·day), in a 20–14 °C range) (101). If the high-load reactor was operated at 12 °C, COD removal decreased to 50–60% and biogranule decomposition started. These side effects could be eliminated via an OLR reduction down to 6.6 kg/(m3·day). However, dairy wastewater has higher average temperature, which makes it possible to apply high-load wastewater treatment technologies (6, 18). Another study showed that mesophilic conditions ((36±1) °C) generate more H2 compared to thermophilic ones ((55±1) °C) during the treatment of cheese whey wastewater, with 9.2 and 8.1 mmol of H2 per g of COD, respectively. The specific H2 production was 4.6 times higher at 36 than at 55 °C (102).

Separated-phase systems are preferred from technological point of view. They have the highest organic loading and shortest HRT compared to other anaerobic digesters. The consecutive acidogenic-methanogenic phase division of anaerobic digestion is suitable for the treatment of dairy wastewater with an unbalanced composition (high C:N ratios which acidify very quickly). In such separated-phase systems, the acidogenic reactor has a major role as it supplies short-chain volatile fatty acids which can be easily fermented to CH4 in the methanogenic reactor. The easily utilisable lactose requires a shorter HRT and a smaller volume of the acidogenic reactor than the methanogenic digester (1, 4, 103). Such a system was used to treat a dairy effluent with 50 kg/m3 of COD and pH=4.5. The COD was decreased by 72% at 35 °C and the following operating conditions: OLR, expressed as COD, of 50 and 9 kg/(m3·day), when HRT was only 1 and 3.3 days in the acidogenic and the methanogenic reactors, respectively (1). The CSTR was the preferable model for the acidogenic phase. In a 9-month operation study, a two- -phase anaerobic reactor comprising an acidogenic-phase CSTR and a methanogenic-phase upflow AF was used to treat dairy waste streams (104). The effluent COD was reduced by 90% and the BOD5 by 95%, while an OLR, expressed as COD, of 5 kg/(m3·day) and a 2-day HRT were obtained. The H2 and subsequent CH4 production from fresh cheese whey were achieved in a CSTR, at 35 °C and HRT of 1 day. The mixed liquor was consequently fermented to CH4 in a baffled bioreactor, operated at HRTs of 20, 10 and 4.4 days. At the lowest HRT, the COD reduction reached 94% (105). An acidogenic CSTR and a final methanogenic upflow AF were used to utilise cheese whey. The results showed that a maximum acidogenesis of up to 50%, with the same OLR (expressed as COD) range (0.5–2 g per mixed liquor suspended solids per day) could be achieved at an HRT of 24 h. The effluent was fed subsequently to the upflow AF where the initial soluble COD was decreased by 90% during HRT of 4 days (106). A two-stage hybrid UASB reactor, filled respectively with polyurethane foam and polyvinyl chloride rings in each phase, was supposed to exceed other anaerobic methods in the treatment of dairy effluents. The combined COD removal in the reactor in a stable equilibrium (10.7 to 19.2 kg/(m3·day)) changed from 97 to 99% (39). Anaerobic rotating biological contact reactors are also discussed in the literature for anaerobic separate phase treatment (1). Carrier incorporation into anaerobic reactors for biomass support greatly increases their specific activity. Depending on the operating temperature, dairy wastewater can be treated in a two-phase separation. The basic configuration presupposes that thermophilic acidogenesis is followed by mesophilic methanogenesis. The information on these processes in the literature is scarce (107–109). An experiment compared two couples of anaerobic SBRs working at the following temperatures: the first couple (thermophilic-mesophilic system) at 55–35 °C and the second (mesophilic-mesophilic system) at 35–35 °C. At an OLR, expressed as VS, varying between 2–4 g/(L·day), the thermophilic-mesophilic system performs better (VS removal rate of 43.8–44.1% when HRT is 3 days and 37.1–38.9% when HRT is 6 days) than the mesophilic-mesophilic system (VS removal rate of 29.3– 30.2% when HRT is 3 days and 26.1–29.1% when HRT is 6 days). The overall improved performance showed that the thermophilic-mesophilic system with respect to total coliform reduction, TSS removal and biogas production, is preferable to the mesophilic-mesophilic SBR couple. Despite that, higher energy consumption during the thermophilic phase should be taken into account from an economical point of view (84). During a set of experiments, a high-temperature-based technology including acetic and butyric acid fermentation followed by CH4 production achieved 116% COD reduction and 43% CH4 biosynthesis, thus performing better than single- -phased processes (110).

Combined (anaerobic-aerobic) processes. Since an anaerobic technology reduces mostly C-containing contaminants and has a weaker effect on nutrient removal, it needs to be considered as only a preliminary step which must be polished. This can be achieved by incorporating a local aerobic step or, occasionally, by directly discharging anaerobic effluent into the municipal wastewater treatment plants (4).

A mixed dairy wastewater was purified on a full- -scale level in consecutive UASB reactor and aerobic denitrification steps. When 95% COD removal was achieved, the produced CH4 was sufficient to cover the plant energy requirements (1).

SBR great flexibility makes it an adequate post-aerobic step in combined dairy wastewater treatment. A new downflow-upflow hybrid reactor containing downflow pre-acidification and upflow methanation chambers was designed to treat high-load cheese wastewater at an average OLR, expressed as COD, of 10 g/(L·day). COD (98%) was converted into biogas, while the discharged soluble COD reached 1 g/L. The process was maintained at stable pH values without chemical addition. After treatment in the SBR, more than 90% of COD, nitrogen from NH3 and TP were removed (32). Wastewaters from raw milk quality laboratories, containing milk preservatives (sodium azide or chloramphenicol), were utilised in an industrial-scale plant with an AF and SBR. Influent FOG were completely treated in the anaerobic step without biomass washout for more than 2 years of operation, the COD decrease being more than 90% at an OLR, expressed as COD, of 5–6 kg/(m3·day). However, alkali had to be added to reduce the critical pH drop. The outgoing stream from the anaerobic process was polished in SBR until the final COD dropped to 200 mg/L and the TN to less than 10 mg/L (52).

The consecutive anaerobic-aerobic technology was used to purify reconstituted whey wastewater in a single reactor at low oxygen concentration and 20 °C. Maximum COD removal of (98±2) % was reached at total cycle time of 4 days and OLR, expressed as COD, of 0.78 g/(L·day). In accordance with specific biomass activity, trophic differentiation can be seen in the system: methanogens predominantly live at the bottom of the bulk liquid, while acidogens inhabit suspended flocs. When the soluble O2 rose to 0.5 mg/L during the aerobic phase, the COD was reduced to (88±3) % in a 2-day total cycle time at 1.55 kg/(m3·day) (111).