Evaluation of anaerobic digestion processes for short sludge-age waste activated sludge combined with anammox treatment of digestate liquor

The current focus of wastewater treatment processes is evolving from only achieving organic matter removal towards gaining the additional benefits of energy and nutrient recovery. The main driving force for this change is the increased cost of energy and the fluctuating cost and availability of some nutrients (e.g. phosphorus). Thus there is strong interest in retrofitting existing treatment processes to produce an effluent that complies with water discharge regulations, but with a reduced energy consumption and physical footprint. It is also highly desirable that the new processes can increase opportunities to recover maximal energy and nutrients from wastewater (Batstone & Virdis 2014).

One promising option in this regard is the high-rate activated sludge process (A-stage), which is highly loaded, with short hydraulic and solids retention times (HRT and SRT, respectively) (Chase & Eddy 1944). This process utilizes the maximum biomass growth potential as a means of removing carbon and partial nutrients (i.e. nitrogen, phosphorus) from wastewater, thereby generating a low carbon effluent. More importantly, such a process preserves and captures the removed carbon and nutrients in the biomass (also called activated sludge), rather than being oxidized to carbon dioxide or converted to nitrogen gas and chemical precipitates (e.g. aluminium phosphate) compared with conventional activated sludge processes, which offers a huge potential for energy and nutrient recovery. In particular for phosphorus, high-rate biological phosphorus (Bio-P) removal can allow phosphate to be effectively captured from wastewater (>90%) in the A-stage configuration and be concentrated in a form that is easily amendable for struvite generation, thus maximizing the concurrent carbon and phosphorus recovery (Ge et al. 2015). It has been estimated that approximately 66% of the influent phosphorus can be recovered through downstream anaerobic sludge digestion and struvite crystallization (Ge et al. 2014). Another advantage of the high-rate process is that the short-contact between wastewater and activated sludge reduces the amount of carbon being oxidized aerobically, and thus requires much less aeration (hence less energy consumption) compared with a conventional activated sludge process with a longer SRT (e.g. 10–15 days) (Ge et al. 2013).

Since the activated sludge generated in the A-stage is likely to be relatively degradable, the ideal option for the subsequent sludge stabilization is anaerobic digestion, which can convert the accumulated carbon to methane as a renewable energy source. While the SRT-degradability relationship has been well established some time ago (Gossett & Belser 1982), most previous research focused on the investigation of medium to long SRT (>5 days) activated sludge, or the evaluation of existing full-scale domestic wastewater systems operating under a particular set of conditions, such as the Strass wastewater treatment plant (WWTP) (0.5 day SRT activated sludge) (Wett & Alex 2003). As a result, the expected methane production of an A-stage sludge with different SRTs, as well as the digestibility variations between thermophilic and mesophilic conditions, has not been determined systematically. In response to this, Ge et al. (2013) assessed the degradability of short sludge-age (2–4 days) activated sludge using biochemical methane potential (BMP) tests. However, this study was conducted in batch mode, and therefore could not evaluate the volatile solids (VS) destruction nor digestate nitrogen and phosphorous contents, which can only be evaluated in a continuous feeding system. This meant that energy balance could not be estimated. Anaerobic sludge digestion results in nitrogenous and phosphorus organic compounds contained in the sludge to be released as ammonia and orthophosphate, of which particularly the ammonia will flow into the subsequent sludge dewatering liquor stream. Again, the extent of nitrogen and phosphorus release has not been determined. There is also limited information about whether the sludge dewatering liquor can be treated by anaerobic ammonium oxidation (anammox) which is an emerging low-energy process used for nitrogen removal (Strous et al. 1997).

This paper addresses these limitations on an activated sludge stream with 2 day SRT produced from a laboratory-scale A-stage reactor treating meat processing wastewater. Anaerobic digestion effectiveness of the activated sludge was evaluated under continuous thermophilic and mesophilic conditions. The overall chemical oxygen demand (COD) balance as well as the residual nitrogen removal from the sludge dewatering liquor via the anammox process are also assessed to provide a comprehensive analysis for the short SRT activated sludge treatment.

The substrate used in this study was waste activated sludge generated from a high-rate sequencing batch reactor (SBR) based process with 2 day SRT (organic loading rate of 24 gCOD d⁻¹) treating meat processing wastewater. The details of the SBR process were described in Ge et al. (2013). The activated sludge collected from the SBR was centrifuged to increase the total solids (TS) concentration prior to feeding. The feed sludge was prepared every second day and stored at approximately 4°C to minimize degradation and preserve integrity of the material. Regular analysis was performed to determine the characteristics and consistency of the feed materials, and the results are summarized in Table 1.

At the start-up, the thermophilic and mesophilic digesters were inoculated from a full-scale anaerobic digester (35 ± 1°C, 20-day HRT) in a municipal WWTP in Brisbane, Australia. Both digesters were fed at intervals of 6 hours (four times daily). During feed events, fresh feed sludge was pumped into the digesters and an equal volume of the effluent was withdrawn simultaneously using a multi-head peristaltic pump. The thermophilic digester was operated at three HRTs: (i) 5-day HRT (240 days); (ii) 8-day HRT (50 days); and (iii) 10-day HRT (53 days). The mesophilic digester had been operated for 90 days before the thermophilic digester commenced operation. All comparisons used in this study were calculated from day 102 forward, where both digesters were stable.

BMP tests were conducted to assess the stability of the digester effluents by determining the residual methane potential from digestion residues. The set-up and operation of the tests are the same as in Jensen et al. (2011). Batch tests were performed in 160 mL non-stirred glass serum bottles (100 mL working volume). The inoculum used in the tests was collected from the same full-scale anaerobic digester as stated above. Substrates used in the tests were the digestates collected from the thermophilic digester and the mesophilic digester. The inoculum to substrate ratio used in the tests was 1:1 (VS basis). After loading the substrate and inoculum, the bottles were flushed with high purity nitrogen gas to create anaerobic conditions, sealed and stored in a temperature controlled incubator (35 ± 1°C). Blanks that contained inoculum and MilliQ water without substrate, were conducted to measure the background methane produced from the inoculum. All tests were carried out in triplicate, and all error bars indicate 95% confidence in the average of the triplicates based on two-tailed t-tests. Degradability was estimated by fitting the exponential decay function as described in Jensen et al. (2011).

Specific anammox activity (SAA) tests were conducted to evaluate the effectiveness of anammox treating the ammonium (NH₄⁺) rich sludge dewatering liquor from the thermophilic anaerobic digester. The method is based on the work of Dapena-Mora et al. (2007). The anammox biomass used in the tests was collected from a 50 L anammox biofilm reactor (35°C). This reactor was operated under fed-batch conditions by injecting a concentrated NH₄⁺ solution periodically and a nitrite (NO₂⁻) solution in small pulses to enrich anammox culture on biofilm carriers (BiofilmChip™ M). After sampling a number of carriers from the parent reactor, the biomass was harvested from the carriers by stirring the carriers in nutrient medium at a low speed and then letting the biomass settle. The medium contains CaCl₂ (0.39 g L⁻¹), KH₂PO₄ (0.05 g L⁻¹), MgSO₄ (0.2 g L⁻¹), FeSO₄ (0.00625 g L⁻¹), EDTA (0.00625 g L⁻¹), trace I acidic solution (1 g L⁻¹) and trace II alkaline solution (1 g L⁻¹). The settled biomass was subsequently washed and re-suspended in new nutrient medium. The substrate used in the tests was the dewatering liquor obtained by centrifuging the thermophilic digester effluent.

The SAA tests were performed in 160 mL serum bottles (100 mL working volume). Each bottle contained the pre-determined volumes of the anammox biomass and the sludge dewatering liquor, achieving the initial concentrations of biomass and NH₄⁺ at 1 g VSS L⁻¹ and 60 mg N L⁻¹, respectively. The bottles were then flushed with pure nitrogen gas to obtain anoxic conditions, sealed and placed in a thermostatic shaker, at 150 rpm and 30°C. The initial pH value was fixed at 7.8. NO₂⁻ was added step-wisely to provide sufficient NO₂⁻ for anammox but also to maintain the NO₂⁻ concentration below 30 mg N L⁻¹, which was reported as an inhibition concentration for anammox (Bettazzi et al. 2010).

The control SAA tests were also set-up and operated in the same way as described above, with the exception of adding a prepared NH₄⁺ stock solution as the substrate. All the tests stated above were conducted in triplicate. Error bars in the figures represent 95% confidence in error based on the average of the triplicate (tests two-tailed t-test).

Biogas samples were taken daily from the thermophilic and mesophilic anaerobic digesters and the biogas composition (CH₄, CO₂, H₂) was analyzed by a Perkin Elmer gas chromatography (GC) equipped with a thermal conductivity detector. Slurry samples were collected from both digesters periodically and analyzed for total chemical oxygen demand (TCOD), TS and VS. A part of slurry samples were filtered through Millipore filter units (0.45 μm pore size). The filtered samples were then analyzed for soluble COD (SCOD), phosphate (PO₄³⁻), NH₄⁺and volatile fatty acids (VFAs). TCOD and SCOD were measured by Merck cell tests (Merck KGaA, Germany). NH₄⁺ and PO₄³⁻ were measured using a Lachat Quik-Chem 8000 Flow Injection Analyser (Lachat Instrument, Milwaukee). VFA concentrations were measured by an Agilent Technologies GC with a flame ionization detector.

In the residual BMP tests, biogas production and composition were also monitored throughout each batch test. Accumulated methane production was calculated by the method described in Ge et al. (2011) with blank methane production subtracted. For sludge samples collected at the start and end of each batch test, the substrate, inoculum and combined slurry samples were analyzed for TCOD, SCOD, TS, VS, NH₄⁺, PO₄³⁻ and VFAs.

In the SAA tests, liquid samples were regularly taken from each bottle throughout the experimental period to measure inorganic nitrogen species (NH₄⁺, NO₂⁻ and NO₃⁻). SAA was determined by regression against the linear section of the NH₄⁺ degradation curves.

The control SAA tests were also conducted with externally adding NO₂⁻ and NH₄⁺, and achieved a total nitrogen removal rate of approximately 77 mg N L⁻¹ day⁻¹, which is similar to the removal rate obtained in the SAA tests with the real sludge dewatering liquor. It should also be noted that the actual amount of anammox biomass used in the tests was less than 1 g VSS L⁻¹, as the biomass collected from the parent anammox reactor contained some ammonium oxidizing bacteria (converting a part of NH₄⁺ to NO₂⁻ for anammox). It therefore can be expected that the total nitrogen removal rate can be substantially enhanced when the anammox biomass is increased to the level normally seen in full-scale plants.

Current changes in the wastewater treatment focus is evolving to energy and nutrients recovery and will result in the production of a (very) short SRT activated sludge stream, the anaerobic degradability of which was determined in continuous operation under mesophilic and thermophilic conditions as part of this study. As measured by VS destruction, the digestion temperature had a very clear effect on the sludge degradability with an increase from 50 ± 2% degradability at 35°C (10-day HRT) to 68 ± 4% degradability at 55°C (10-day HRT). This is probably due to the thermophilic temperature enabling increased destruction of specific components or dissolution of complex organics (e.g. waxes and fats), and also improving the lysis and hydrolysis rates of microbial cells in the activated sludge to release more cellular organics into the bulk for subsequent degradation, thereby enhancing the overall degradability. Moreover, the determined sludge degradability is consistent with the reverse correlation between SRT and sludge degradability as reported by Gossett & Belser (1982), and also fills the blank where the degradability of the short SRT (<4 days) sludge has been poorly assessed. The higher degradability of 60–70% measured at 8–10 day HRTs (55°C) is also close to the maximal degradability (nearly 80%) obtained in the BMP tests for the similar sludge-age (2 days) activated sludge (Ge et al. 2013).

Another concern of short SRT sludge stabilization is the degree of residual methane potential from the digestion residues, due to the relatively high degradability of the input material. In this study, except for the 5-day HRT thermophilic digestion process, the residues from all tested digestion processes showed relatively low residual methane potentials, which are all within the commonly applied residue disposal standard of 17% (equivalent to 110 mL methane production per VS_effluent added in this case) (EPA 1994). This indicates the digestion residues are suitable for disposal in terms of nutrients regulations. The thermophilic treatment (55°C), on the other hand, enables the residues to meet the standards for pathogen destruction if semi-batch operation is applied in practise (Paul et al. 2012). It should also be noted that some nutrients (e.g. NH₄⁺, PO₄³⁻) are highly concentrated in the digestion residues. For example, the level of PO₄³⁻ reached 200 mg PO₄³⁻-P L⁻¹ in this case, suggesting a strong potential to achieve efficient recovery in a dedicated struvite crystallizer (Yuan et al. 2012). While struvite precipitation would also remove some limited amount (approximately 30%) of NH₄⁺, the remainder would ideally be eliminated from the sludge dewatering liquor through a dedicated process such as the anammox process tested in this study. The observed total nitrogen removal rate in the SAA tests confirmed that the anammox process operated efficiently with the dewatering liquor, as similar removal rates were measured with real sludge dewatering liquor and pure NH₄⁺ solution. It indicates that the SCOD present in the dewatering liquor was not further consumed for denitrification, and the dewatering liquor also did not have inhibiting/toxic effects on the anammox activity. This is very important for the emerging energy self-sufficient high-rate wastewater treatment processes as the low COD:N ratio effluent streams from the A-stage and the digester could be combined and treated by anammox. This also can optimally utilize the infrastructure investments for anammox reactors and reduce the nitrogen load in the subsequent process following A-stage (by avoiding the nitrogen loading from the dewatering liquor).

The large amount of methane is another valuable product from anaerobic sludge digestion, particularly for the highly degradable short SRT sludge tested in this study. The generated methane can be used in a co-generation process for production of energy (electricity) and heat, which can offset the power and heating demand for thermophilic digestion, as well as other upstream and downstream processes (i.e. high-rate wastewater treatment, secondary thickening, sludge dewatering, anammox). The energy balance of such an integrated high-rate treatment system has been analyzed (details in the Supplementary Material, available with the online version of this paper) based on the method described in Ge et al. (2013). Generally, the energy recovered from the methane production exceeded the energy requirements of the integrated high-rate system, suggesting that a net electric power output can be gained. In particular, approximately 3.5 times more net energy gain can be achieved from the integrated high-rate system with thermophilic digestion compared with mesophilic digestion, resulting in the system energy self-sufficiency reaching to approximately 170%. In addition, the usage of renewable energy (methane) can generate renewable energy credits (e.g. $0.034 per kWh in Australia), which is expected to substantially reduce the operating expenses for such an integrated system.

The thermophilic (55°C) digestion process with 8–10 day HRT achieved 60–68% VS destruction of the 2 day SRT activated sludge, which is significantly higher than a comparable 10 day HRT mesophilic process. Even for this highly degradable material, the results indicated an advantage for thermophilic digestion. The NH₄⁺-rich dewatering liquor from the sludge digestion was treated by an anammox process, with the comparable nitrogen removal rate as the tests using a control synthetic ammonia solution, indicating that the dewatering liquor did not have inhibiting/toxic effects on the anammox activity.

This work was funded by the Australian Meat Processor Corporation (AMPC), with matching funds provided from the Australian Government, via Meat and Livestock Australia (MLA). We thank Barbara Wronski for the assistance in digester operation, and the analytical laboratory in the Advanced Water Management Centre for chemical analysis.