Source separation and anaerobic co-digestion of blackwater and food waste for biogas production and nutrient recovery

Every year, over one-third of the food produced for human consumption is wasted (Boliko 2019), amounting to approximately 1.3 billion tons globally, with an equivalent CO₂ emission of 30–40 billion tons (Katajajuuri et al. 2014; Counts 2024). Food and kitchen waste (FW) is predominantly disposed of in landfills in many regions of the world, leading to uncontrolled degradation, odor, significant nutrient and energy loss, and substantial greenhouse gas emission (Xue et al. 2024). Besides landfilling, biowaste, including food waste, is often processed through anaerobic digestion (AD) producing biogas and a nutrient-rich digestate that can be used directly or after preprocessing as a fertilizer. Source-separated blackwater (BW), consisting of urine and feces and containing the majority of the nutrients present in municipal wastewater (Wang et al. 2023), is currently diluted with other wastewater streams, and treated at centralized wastewater treatment plants. These plants consume considerable amounts of energy for water pumping and aeration of the activated sludge process (Eshetu Moges et al. 2018).

In the context of a circular economy, one approach involves source separation of BW from other municipal wastewater using minimal transport water followed by anaerobic co-digestion with FW, resulting in a nutrient-rich digestate (Kujawa-Roeleveld & Zeeman 2006). To enable nutrient recovery from municipal wastewater, the collection of BW with water-conserving toilets (e.g., vacuum toilets), followed by local and decentralized treatment, has been proposed and has the potential to facilitate 80–95% nutrient recovery (Gao et al. 2019). In BW, the nutrients are in a concentrated form and are separated from various heavy metals, microplastics, and chemicals present in municipal wastewater collected at centralized wastewater treatment plants (Eshetu Moges et al. 2018). Source-separated BW typically has a high chemical oxygen demand (COD) of >10 g/L and ammonium content of >1 g/L, resulting in a low carbon-to-nitrogen (C:N) ratio, making the AD process challenging due to potential ammonia inhibition (Gao et al. 2019; Zhang et al. 2019). The anaerobic co-digestion of FW and BW improves the buffering capacity and C:N ratio of the process, providing macro and micronutrients for the microbes (Lim et al. 2014; Wang et al. 2020).

Several studies have explored the co-digestion of BW and FW for biogas production (Kjerstadius et al. 2015; Zhang et al. 2019; Wang et al. 2020; Fendel et al. 2022), yet only a few have examined the nutrient fate and recovery (Kjerstadius et al. 2015; Giwa et al. 2022) and mostly in laboratory-scale reactors (Rajagopal et al. 2013; Minale & Worku 2014; Gao et al. 2019; Zhang et al. 2019; Gao et al. 2020; Wang et al. 2020; Zhang et al. 2021; Giwa et al. 2022). To date, the only pilot-scale study reported in continuously stirred tank reactors (CSTRs) has employed a 0.63 m³ reactor for co-digestion of BW and FW, attaining a CH₄ yield of 222 to 332 L per kg-COD removed and a COD removal efficiency of 70–78% (Wasielewski et al. 2016). However, there have been no additional studies on the co-digestion of FW and BW in a pilot-scale in a CSTR system, which is typically used in the industry.

Depending on the feedstock used for AD, digestate can be preprocessed to separate it into liquid and solid fractions using energy-intensive industrial-scale techniques such as screw press, centrifugation, evaporation, or membrane technologies (Fuchs & Drosg 2013). In wastewater treatment, other separation techniques include flocculation followed by sedimentation. Several studies have reported the effectiveness of digestate flocculation (Bauer et al. 2021; Chini et al. 2021; Zuo et al. 2021), with some specifically applying both flocculation and sedimentation (Zhou et al. 2019; Bauer et al. 2021). The resultant solid digestate can be composted, utilized in hydrothermal processes for generating heat and biochar (Fuchs & Drosg 2013; Xia & Murphy 2016), or transformed into fertilizer granules (Czekała et al. 2022). However, there are relatively few studies on the use of liquid digestate. Liquid digestate from various sources has been employed for cultivating microalgal biomass (Erkelens et al. 2014; Parichehreh et al. 2019; Tan et al. 2022), yet no studies have been reported on microalgae cultivation using digestate derived from the co-digestion of BW and FW.

Digestate from AD of FW can be directly applied to the land as fertilizer without preprocessing. However, co-digestion with BW would limit the application of digestate on agricultural land, as in many countries, the use of digestate originating from municipal wastewater treatment and sewage sludge is restricted, for instance, due to its heavy metal content and potential presence of pharmaceuticals, microplastics, etc. The challenges of digestate treatment raise questions about the circularity of the biogas industry when treating municipal wastewater or sewage sludge. Besides direct use as fertilizer, digestate can serve as a nutrient source in other biological processes, such as the biological hydrogen methanation process (Kamravamanesh et al. 2023) or for cultivation of microalgae like Chlorella vulgaris (C. vulgaris) (Parichehreh et al. 2019) and Scenedesmus acuminatus (S. acuminatus) (Alves et al. 2019), for the production of biodiesel and/or bioenergy. Microalgae can utilize the nitrogen, phosphorus, and trace elements from the liquid fraction of digestate as well as CO₂ from the biogas to grow biomass. Utilizing a liquid fraction of the digestate for microalgae cultivation enhances process sustainability by reducing reliance on artificial synthetic nutrients and freshwater (Navarro-López et al. 2020; Ranglová et al. 2021). Nonetheless, the application of the liquid fraction of the digestate is limited by its physicochemical properties, such as opacity, which prevents light penetration, the presence of inhibitors, including bacteria (Bauer et al. 2021) and high ammonium content.

In this study, we demonstrate the pilot-scale co-digestion of source-separated BW and FW for biogas production under mesophilic conditions. The resulting digestate was treated through flocculation and sedimentation to separate the solid and liquid fractions, and the fate of nutrients was analyzed. In addition, we explored the potential of using liquid digestate for the cultivation of two microalgal strains, S. acuminatus and C. vulgaris. The findings presented here may encourage stakeholders to adopt source separation of organic and wastewater streams and may aid in the development of digestate treatment techniques to produce nutrient-rich algal biomass in the future.

BW was obtained from vacuum toilets installed in Hiedanranta, Tampere, Finland, utilizing 1.5 L of water per flush. It was collected into 1,000 L tanks and stored on-site for a minimum of 30 days before usage. The FW was obtained from two local restaurants, from a student restaurant at the university campus (Juvenes Oy, Tampere) and from a lunch restaurant located in Hiedanranta (Zipatta, Tampere) and subsequently minced using a blender (Biltema, Denmark). The preprocessed FW was stored at 2–5 °C and utilized as feedstock. The contents of the two FWs collected varied and the FW from the student restaurant consisted mainly vegetable waste, while the FW from the lunch restaurant also contained food waste from the customers. Digested sewage sludge, serving as the inoculum, was obtained from the Viinikanlahti wastewater treatment facility in Tampere, Finland. Before using it as inoculum in laboratory assays, it was stored at 4 °C, while in the pilot reactor, it was used as such.

The biomethane production potential (BMP) of BW and FW was evaluated separately and in combination using triplicate batch assays. The assays were conducted in 1 L batch reactors with a working volume of 0.7 L operating under mesophilic conditions at 35 °C. Each reactor was equipped with a 5-L Supel™ Inert Foil gas sampling bag (Merck, Germany) to collect the produced biogas. The volatile solid (VS) ratio of BW to FW was 0.41, and the VS ratio of the substrate to inoculum was 1.0 g-VS/g-VS. A total of 80 mL of inoculum as well as NaHCO₃ (with the final concentration of 4 g/L) was added to each batch reactor and the initial pH ranged from 7.2 to 7.8. Before sealing, reactors were purged with nitrogen gas to establish an anaerobic environment. Alongside the primary reactors, the BMP of inoculum was assessed in a separate reactor. The cumulative methane production from the inoculum was subsequently deducted from the cumulative methane production in each reactor. The characteristics of BW, FW, and inoculum used in BMP assays are detailed in Table 1.

A pilot-scale CSTR with a liquid volume of 2 m³ (Metener Ltd, Finland) was operated in a semi-continuous mode for 153 days for co-digestion of BW and FW at a controlled temperature of 35 °C that was established by recirculating the reactor content through a heat exchanger situated outside the reactor. No pH control was used in the reactor. The reactor was fed once a day and each feeding cycle began by removing a specified volume of digestate, after which the substrate was thoroughly mixed in the feed tank using a submersible pump before being transferred to the reactor (despite utilizing mixing, stratification of the feedstock in the feed tank was occasionally observed). This process maintained a hydraulic retention time (HRT) of 25 days and an organic loading rate (OLR) between 0.42 and 0.91 kg-VS/m³/d.

The characteristics of FW and BW used in the reactor studies are presented in Table 2. The feed was prepared in 1 m³ scale weekly and was used over a period of 7 days, and despite the storage temperature of 1–10 °C, the biodegradation of organic matter in the feed was observed from time to time leading to an increase in volatile fatty acid (VFA) concentration in the feed tank and reducing the pH. This, however, had no significant impact on the process performance, but the influent COD and OLR were fluctuating. The co-digestion process commenced with an initial filling of the reactor with 1 m³ of inoculum. Feeding began after a 14-day stabilization period, initially scheduled once every 7 days for the first 3 weeks. Upon achieving stable reactor performance, the feeding frequency was increased to once a day (the start of this feeding frequency, around day 30, was considered as the start of the process).

The volume of the produced biogas was measured using an Itro G4 RF1 (Itron, UK) gas flow meter. The collected gas was stored in a gas bag (IC2 Feeniks Oy, Finland), and its composition was analyzed using a Dräger X-am 8000 (SENSOREX, Finland). Feed and digestate were sampled three times per week, using at least duplicate samples. Before analysis, samples were centrifuged at 4,500 rpm for 30 min (Rotina 420, Hettlich, USA) and filtered through a 0.45 μm pore size filter (Chromafil, Macherey-Nagel, Germany).

The organic matter content in the reactor feed was high, as indicated by the elevated average ratio of VS to total solids (TS) and volatile suspended solids to total suspended solids (VSS/TSS) at 0.92 and 0.94, respectively. The COD of the feed was 21 ± 6 g/L with a soluble COD (CODs) of 14 ± 4 g/L, of which 1.2 ± 0.4 g/L constituted VFAs. The VFAs primarily included acetate, butyrate, propionate, and isobutyrate. The pH of the feed varied significantly, ranging from 4.2 to 7.0. The ammonium nitrogen (NH₄-N) concentration in the feed was high, averaging 1.16 ± 0.26 g/L. The variability in the concentration of organic matter, particularly noted during one feeding cycle, was likely due to hydrolysis and acidification reactions occurring in the feed tank, as evidenced by the decreasing trends in TS and VS values over time.

Flocculation was conducted to remove organic matter and the solid fraction of the digestate using various Flopam cationic polyacrylamide polymers from SNF Finland Oy, including FO-4240 SH, FO-4290 SH, FO-4440 SH, and FO-4350 SH. These polymers, which vary in charge density and molecular weight, were tested at different concentrations with agitation at 300 rpm. The characteristics of the flocculants are detailed in Table 3. The flocculation process was carried out at room temperature (20–24 °C) without pH adjustment. Subsequently, the separation of the solid fraction was achieved through overnight sedimentation at room temperature.

Axenic cultures of C. vulgaris (SAG 211-12) and S. acuminatus (SAG 38.81) were obtained from the George-August Universität culture collection in Göttingen, Germany. Initial cultivation of both S. acuminatus and C. vulgaris was conducted in the N8 mineral media composed of the following: KNO₃ (0.5 g/L), KH₂PO₄ (0.74 g/L), Na₂HPO₄ (0.26 g/L), MgSO₄·7H₂O (50 mg/L), CaCl₂·2H₂O (17 mg/L), FeNaEDTA·3H₂O (11 mg/L), ZnSO₄·7H₂O (3.2 mg/L), MnCl₂·4H₂O (13 mg/L), CuSO₄·5H₂O (18.3 mg/L), and Al₂(SO₄)₃·18H₂O (7 mg/L). This was done in shake flasks at room temperature (25–28 °C) under continuous illumination of 50 ± 5 μmol photon/m²/s in photosynthetically active radiation, with agitation at 150 rpm using a shaking plate (Infors, Switzerland).

For subsequent cultivation in digestate, once S. acuminatus and C. vulgaris reached an optical density (OD₇₅₀) of 2, the cells were centrifuged at 4,500 rpm for 5 min, washed, and re-suspended either in pure digestate or digestate diluted with tap water having the following final digestate concentrations of 10, 13, 17, 20, 25, and 50% (v/v). A 15% (v/v) of cell suspension was used to inoculate 1 L batch reactors with a working volume of 0.7 L. Gas flow was regulated by mass flow meters, maintaining an air flow rate of 10 mL/min. The system operated without pH (at pH 7–9) and temperature (20–35 °C) control. Continuous illumination was done using LEDs at an intensity of 40 μmol/m²/s.

TSS, VSS, TS, and VS were determined gravimetrically following the standard method (APHA 2018a). pH was measured with SenTix 41 electrode and WTW pH 3210 meters. Total COD (CODt) and soluble COD (CODs) were determined using the dichromate method (APHA 2018b). The concentration of VFAs was assessed using Shimadzu Ordior GC-2010 plus gas chromatograph, as described by Kokko et al. (2018), and was calculated in terms of CODs per L. For BMP experiments, CH₄ concentration was measured using a Perkin Elmer 500 GC-FID with a Mol-Sieve 5A PLOT column according to Kokko et al. (2018), while the volume in the gas bags was measured with the water displacement method (Owen et al. 1979).

Cation concentrations were analyzed using a Dionex DX-120 ion chromatograph (ThermoFisher Scientific, USA) and a Dionex Ionpac™ CS12A column (ThermoFisher Scientific, USA). The anions were measured using Dionex AS-DV and an Ionpac™ AS22 column (ThermoFisher Scientific, USA) according to the method described by Jermakka et al. (2021).

Microwave acid digestion of feed and digestate was performed using the MARS6 system (CEM, USA) and concentrated HNO₃ and HCl at 165 °C and 800 psi for 30 min. The elemental concentrations of the digested samples were subsequently measured via inductively coupled plasma mass spectrometry using Thermo Scientific iCAP™ RQ equipment. All measurements were carried out in kinetic energy discrimination mode, employing He as the collision gas in the collision/reaction cell and Ar as the carrier gas.

The CHNS/O analysis of algal biomass was conducted by flash combustion using a FlashSmart organic elemental analyzer (ThermoFisher, USA), with cysteine and methionine serving as calibration standards.

Due to significant variations in the feeding cycles during the reactor experiments, methane yields and OLRs were calculated as averages for each feeding cycle to ensure consistency and reliability in the data interpretation (feeding cycle implies each batch of fresh feed prepared).

To assess the fate of nutrients in the AD process, a detailed analysis of both the feed and digestate was conducted. The analyses focused on macronutrients present in the combined feed of BW and FW, including total phosphorus (P), NH₄-N, sodium (Na), potassium (K), magnesium (Mg), sulfate (SO₄), phosphate (PO₄), and nitrate (NO₃). The reduction in nutrient concentration in the digestate compared with feed can be attributed to the potential stratification of some solids either in the feed tank, in the digester, or the digestate tank, leading to generally lower concentrations in the digestate compared with the initial feed (Table 4). In contrast, the concentration of NH₄-N was higher in the digestate, averaging 1.5 ± 0.15 g/L, compared with 1.1 ± 0.05 g/L in the feed. This could be attributed to ammonification of organic matter during AD. Furthermore, the analysis revealed that the average concentration of several micronutrients, including iron (Fe), nickel (Ni), cobalt (Co), copper (Cu), chromium (Cr), and lead (Pb), exhibited slight increases in the digestate relative to their levels in the feed (Table 4). This could be attributed to the breakdown and transformation of these elements into more soluble forms during the AD process, emphasizing the complex interactions and transformation that organic substrates undergo during AD (Fuchs & Drosg 2013).

The digestate resulting from anaerobic co-digestion of BW and FW exhibited an average TS content of 0.8 ± 0.1% and VS content of 0.4 ± 0.1%. The ammonium concentration in the digestate varied initially, ranging from 1.4 ± 0.1 g/L between days 29–60. This concentration increased slightly to 1.6 ± 0.1 g/L by day 63 and maintained this level until the end of the monitoring period (Figure 3(c)). To effectively separate the liquid and solid fractions of the digestate, a flocculation process was optimized using four different polymers (Table 3). Among these, FO-4240 and FO-4350 achieved complete flocculation of solid matter, forming agglomerates that led to efficient sedimentation of the solid particles and a distinct separation of solid and liquid fractions. In contrast, FO-4440 resulted in only partial agglomeration with more than a 50% volumetric solid fraction and no efficient sedimentation, while FO-4290 showed no agglomeration at all.

Despite the effective flocculation observed with FO-4350, the liquid fraction post-sedimentation appeared turbid, suggesting incomplete separation or the presence of residual colloidal particles. Consequently, polymer FO-4240, at an optimal concentration of 20 mg/L, was selected for the subsequent digestate treatment process due to its superior performance in clarifying the liquid fraction and the low concentration required to achieve high-efficiency flocculation.

After flocculation and sedimentation with FO-4240, the digestate's volumetric composition was 88% liquid and 12% solid. After treatment, the liquid digestate displayed a reduced TS content of 0.3 ± 0.0% and a VS content of 0.09 ± 0.0%. In addition, the CODt of the liquid fraction was measured at 2.5 ± 0.1 g/L, with the CODs of 2 ± 0.1 g/L (Table 4). This indicates a substantial reduction in organic matter, reflecting the effectiveness of the treatment process in managing and reducing the solid content of the digestate.

During the flocculation and sedimentation process, the pH of the liquid fraction of the digestate increased from 7.4 ± 0.2 to 8.4 ± 0.2. The detailed characteristics of both liquid and solid fractions of the digestate are provided in Table 4.

A significant portion of the nutrients was retained in the liquid fraction, with more than 90% of NH₄-N and over 70% of total phosphorus preserved (Table 4). Similarly, high recovery efficiencies were observed for other essential macronutrients, with more than 90% of sodium (Na), potassium (K), magnesium (Mg), calcium (Ca), NO₃, PO₄, and SO₄ remaining in the liquid fraction. This effective nutrient conservation during the flocculation process suggests that these nutrients can potentially be recycled or repurposed from liquid fraction.

A smaller proportion of metals, including iron (Fe), nickel (Ni), zinc (Zn), and cobalt (Co), were found in the solid fraction of the digestate (Table 4). This separation is advantageous as the liquid fraction remains rich in nutrients, making it suitable as feedstock in biological processes where microorganisms require both macronutrients and trace metal salts for growth and production.

In this study, the growth and nutrient utilization of two lipid-producing algal strains, S. acuminatus and C.vulgaris, were investigated using the liquid fraction of digestate as a growth medium. Photoautotrophic growth experiments were performed in 1 L batch reactors with various dilutions of the liquid digestate mixed with tap water to assess the feasibility of using digestate as a nutrient source for algal cultivation.

In contrast, no growth was observed with S. acuminatus cultivated in either the undiluted liquid digestate or its various dilutions (Figure 4(b)), indicating a potential inhibitory effect of the nutrient profile of the liquid digestate for this algal strain.

The CHNS analysis of the dried biomass of C. vulgaris cultivated on the liquid fraction of digestate showed the following composition: 8.6 ± 0.2% nitrogen (N), 47.2 ± 1% carbon (C), 4.7% ± 0.5% hydrogen (H), and 1 ± 0.2% sulfur (S), and the remainder comprising other elements. After the growth of C. vulgaris, the liquid digestate showed an average reduction of 85 ± 3% in NH₄-N and 78 ± 5% in PO₄-P content on day 20 of the cultivation. While no nitrate was detected in the liquid digestate post-cultivation, the composition of Na, K, Mg, and Ca remained unaffected.

The nutrient-rich profile of the liquid digestate underscores its potential as a medium for cultivating nutrient-rich algae, which could be further exploited as biofertilizer or in other biotechnological applications.

The characteristics of the BW and FW analyzed in this study, including TS (2.5 ± 0.5%), and VS (2.2 ± 0.5%) content, pH (7.4 ± 0.2), CODt (22 ± 6 kg/m³), and CODs (13 ± 3 kg/m³), align with values reported previously (Rajagopal et al. 2013; Lavagnolo et al. 2017; Gao et al. 2019). The synergy of BW and FW for co-digestion was evident in the current study. BW, characterized by a very low TS and VS content (approximately 0.5%), contributed minimal organic matter to the digestion process but had a high NH₄-N concentration of 1.4 ± 0.1 kg/m³ and pH of 8.7 ± 0.1. In contrast, FW had a high TS and VS content of 16–18%, serving as the major source of organic matter and thus resulting in C:N ratio ranging from 8.8:1 to 9.5:1 of the combined feedstock.

Despite previous reports of high concentrations of free ammonium (greater than 205 mg/L) inhibiting methanogenesis in AD of BW collected from vacuum toilets (Gao et al. 2019, 2020), no such inhibitions were observed in the current study. This is likely due to the co-digestion with FW, which balanced the concentration of free ammonium in the digestor. The operational OLR of 0.6 ± 0.1 kg-VS/m³/d resulted in an average NH₄-N concentration of 1.1 ± 0.05 kg/m³ in the feed, which was found favorable to biogas production (Figure 3(c)).

In the laboratory trials, the methane production potential, as assessed by BMP tests, revealed a methane productivity of 583.7 ± 70 L/kg-VS for FW, surpassing previously reported yields of 400–496 L/kg-VS (Rajagopal et al. 2013; Ebner et al. 2016). This higher productivity could be attributed to the composition of FW obtained in the current study containing more meat waste, as well as the low OLR used in the batch tests. For BW, the methane yield of 304 ± 44 L/kg-VS (Figure 1) was in the range reported in previous studies (220–400 L/kg-VS) (Lavagnolo et al. 2017; Zuo et al. 2021). When co-digested with FW, a maximum methane yield of 492 ± 21 L/kg-VS, closely aligning with the previously documented range of 520–590 L/kg-VS (Lavagnolo et al. 2017) was attained.

At the pilot-scale, the CH₄ yield of FW co-digested with BW was 0.73 ± 0.1 m³/kg-VS at an OLR of 0.59 ± 0.14 kg-VS/m³/d, which was on average 67% higher than in laboratory-scale BMP studies, which showed 0.49 ± 0.02 m³/kg-VS. The CH₄ production rate in pilot studies ranged from 0.6 to 2.3 m³/m³/d, comparable to the highest CH₄ production rates reported by Gao et al. (2020) of 2.4 ± 0.15 m³/m³/d for co-digestion of BW and FW, despite using a digestor of only 3.5 L size. During the pilot study, the CH₄ content in the biogas was 65 ± 3%, like that observed in the batch assay (64 ± 3%) in this study. The CODt removal efficiency achieved during the pilot study averaged 83 ± 4% (Figure 3(d)), which was at the upper end of the range (52–84%) reported in previous studies (Rajagopal et al. 2013; Gao et al. 2020). In addition, the VS removal efficiency of 82 ± 6% in this study is in a higher range than previously reported similar studies (Gao et al. 2020).

Furthermore, both the average (0.7 ± 0.1 m³/kg-VS) and the maximum (1.1 ± 0.1 m³/kg-VS) methane yields observed in the pilot trials were significantly higher than those previously reported for laboratory-scale digesters (0.21–0.37 m³/kg-VS) (Rajagopal et al. 2013; Gao et al. 2020). In the current study, conducted using a 2 m³ pilot-scale CSTR, the HRT ranged from 20 to 27 days, aligning with the 10–30 days reported for CSTRs with an OLR of 0.59 ± 0.14 kg-VS/m³/d, resulting in an average methane yield of 0.7 ± 0.1 m³/kg-VS. The HRTs for co-digestion of BW and FW vary significantly depending on the reactor system employed. In accumulation systems, HRTs are reported to range from 105 to 280 days, although these studies do not specify the corresponding OLR. Notably, Gao et al. (2019) reported an HRT of 2.6 days in a laboratory-scale 3.5 L UASB, although the scalability of such a system remains unexplored.

The VFAs in the feed ranged from 1.0 to 2.5 g-COD/L and were predominantly consumed during the digestion process. From days 21 to 69, the VFA levels fluctuated but ultimately remained below 0.5 g-COD/L. By day 70, no VFAs were detected (Figure 3(b)), indicating a more efficient VFA degradation compared with previous studies, which reported total VFA concentrations of 0.76–8.89 g/L (Elmitwalli et al. 2006; Kujawa-Roeleveld & Zeeman 2006; Rajagopal et al. 2013; Zhang et al. 2019; Gao et al. 2020; Wang et al. 2020).

Throughout the pilot study, the effluent NH₄-N concentration increased from an initial average of 1.1 ± 0.05 g/L in the feed to 1.6 ± 0.1 g/L in the digestate (Figure 3(c)). This increase is likely due to the ammonification process, which involves the degradation of proteins, nucleic acids, and other nitrogen compounds present in the feed, such as meats and plant material. Solubilization makes the organic nitrogen available for ammonification, resulting in increased NH₄-N concentration in the effluent. Ammonification has been reported in AD of FW where it can inhibit the digestion process due to toxicity to hydrogenotrophic methanogens (Serna-Maza et al. 2014; Chen et al. 2016). Therefore, controlling and understanding ammonification is crucial in managing AD processes using FW as feedstock. Notably, no inhibitory impact for ammonium concentration was detected in this study, aligning with the absence of VFA accumulation toward the end of the pilot trial (Figure 3(b)).

The negligible denitrification observed in the system under anaerobic conditions may indicate insufficient denitrifying bacteria or competing scenarios between denitrifying bacteria and methanogens for carbon sources. Methanogens may have outcompeted denitrifies, leaving the nitrate concentration in the feed unaffected (Deng & Shi 2020). Only very high nitrate concentrations (exceeding 1.5 g/L in AD) can inhibit methanogenesis (Sheng et al. 2013) by creating a competitive environment, reducing the availability of the carbon source for methanogens, and favoring the dominance of bacteria converting nitrate to nitrite, which is reported toxic to methanogens (Klüber & Conrad 1998).

While the 2 m³ CSTR pilot reactor used in this study demonstrated high organics removal efficiency, and CH₄ content a limitation was observed in maintaining a long steady state during the process, a common challenge in large-scale biogas facilities (Leitão 2004; Owusu-Agyeman et al. 2019). In this study, fluctuations in influent CODt and VS content were primarily due to stratification from inefficient mixing and the biodegradation of FW in the feed tank. This issue could be mitigated by real-time monitoring and adaptation of OLR, automated daily feed preparation, or using feed storage tanks and pumps that allow for better mixing of the feedstock, thereby buffering against short-term fluctuations in VS.

Besides biogas, digestate is a valuable byproduct of the AD process (Xia & Murphy 2016). For digestate to be utilized in various applications, some preprocessing is essential. The choice of downstream processing methods depends on the digestate composition, specifically to remove solids and adjust nutrient levels (Bauer et al. 2021). In this study, the aim was to remove suspended organics, including colloids and solid matter, from the digestate while retaining most of the nutrients in the liquid phase to reduce turbidity and facilitate light penetration into the liquid digestate, making it suitable as feedstock for microalgal cultivation. Therefore, various cationic polyacrylamide flocculants differing in charge density and molecular weight were evaluated for the downstream processing of digestate obtained from the co-digestion of BW and FW.

The flocculant FO-4240 SH, characterized by medium charge density and molecular weight, demonstrated the highest efficiency in solid removal from the digestate. Following sedimentation, the resulting liquid fraction contained, on an average, 62 ± 1% less TS and 77.5 ± 2% less VS compared with the whole digestate. In addition, CODt and CODs were reduced by 43 ± 2% and 48 ± 1%, respectively, in the liquid fraction.

Over 90% of macronutrients, including NH₄-N, K, and NO₃, were present in the liquid fraction, whereas only 15–20% of the total phosphorus and some fractions of metals such as Fe, Ni, Zn, and Co remained in the solid fraction. This distribution contrasts with other studies, which reported 20–30% of NH₄-N, 55–65% phosphorus, and 20–30% potassium remaining in the solid fraction post-flocculation of digestate obtained from AD of agricultural and food waste (Bauer et al. 2021).

Chini et al. (2021) demonstrated that chemical flocculants, such as polyphenolic organic polymers and polyacrylamide, exhibit considerably lower nutrient removal efficiencies (20–65%) compared with centrifugation for digestate from the AD of swine manure. Similarly, Chuda & Ziemiński (2021) found that the combination of centrifugation with high-density cationic polyacrylamide flocculants effectively retained only 61% of nitrogen, 75% of potassium, and 33% of phosphorus in the liquid fraction of digestate originating from the AD of sugar beet pulp.

The use of flocculation followed by sedimentation for digestate treatment holds potential for large-scale application due to low investment cost and low energy input in comparison to filtration, centrifugation, and other dewatering techniques commonly done for digestate treatment.

One effective strategy for recovering nutrients from the liquid fraction of digestate is through microalgae cultivation. In this study, to ensure adequate light penetration for algal growth and balance nutrient concentrations, the liquid fraction of digestate was diluted with tap water. Our results demonstrated distinct growth patterns between the two algal strains. No growth was observed for S. acuminatus on the digestate or its dilutions. A possible explanation for this could be the presence of competitive microorganisms in the digestate, which might consume or outcompete the microalgae (Bauer et al. 2021). Previous studies have demonstrated successful growth of S. acuminatus on other types of digestate, where sterilization of liquid digestate through autoclaving was performed before cultivation (Park et al. 2010; Dickinson et al. 2015). Another explanation could be the nutrient profile of the liquid digestate, which may have an inhibitory effect on S. acuminatus. Potential factors contributing to this inhibition could include high concentrations of ammonium, heavy metals, or other compounds present in the digestate that S. acuminatus is less capable of tolerating or metabolizing compared with C. vulgaris.

Similar findings have been reported where specific algal strains are inhibited by ammonia (Cho et al. 2013), organic constituents, COD (Franchino et al. 2016; Tigini et al. 2016), and heavy metals (Wong et al. 1994) in digestates or wastewaters. An ammonium concentration of 0.3 mM at pH 6 has been reported as inhibitory to the growth of Scenedesmus obliquus (Lu et al. 2018).

C. vulgaris grown in 20, 17, and 13% liquid digestate revealed a lag phase of 7–8 days, likely representing the adaptation period required for the cells to acclimate to the new environment. This phenomenon has been reported with other digestates (Cai et al. 2013; Prajapati et al. 2014; Zhu 2015) and wastewaters (Åkerström et al. 2014; Dickinson et al. 2015). The immediate growth and higher final biomass concentration for C. vulgaris at 10% dilution suggest that this concentration of digestate provides an optimal balance of nutrients, minimizing inhibitory effects while supplying adequate resources for rapid growth. This observation aligns with previous studies that have shown C. vulgaris can adapt and thrive in nutrient-rich wastewater, including digestates (Wang et al. 2020).

Notably, the growth of C. vulgaris on 10% liquid digestate was comparable to its growth in control conditions using N8 mineral media, achieving a higher biomass concentration of 1.2 ± 0.1 g/L at the end of the experimental period. The biomass concentrations observed in this study are within the range of 0.1–1.7 g/L reported for various microalgal species grown on different liquid digestates (Xia & Murphy 2016). Some studies have achieved higher biomass concentrations up to 4.0 g/L by employing sterilization methods for digestate, such as autoclaving, or by making dilutions using synthetic mineral media or other wastewater (Åkerström et al. 2014; Cheng et al. 2015). However, sterilization by autoclaving is an energy-intensive and costly process, and not feasible for large-scale applications. In addition, using mineral salts to dilute digestate is neither sustainable nor practical.

Furthermore, growth inhibition at higher concentrations of digestate (greater than 20%), primarily due to photoinhibition and/or nutrient overload, was observed in this study and has been similarly reported in other research (Hollinshead et al. 2014; Fernandes et al. 2020). This highlights the importance of optimizing digestate concentration to balance nutrient availability and prevent inhibitory effects.

In this study, carbon constituted 47.2 ± 1% of the C. vulgaris biomass cultivated on liquid digestate derived from anaerobic co-digestion of BW and FW. Following the removal of C. vulgaris biomass via sedimentation or filtration, the digestate supernatant exhibited an average reduction of 50 ± 10% in both CODs and CODt. Notably, C. vulgaris is known for its mixotrophic growth capabilities (Xia & Murphy 2016), suggesting that the carbon for biomass growth could have originated from both the organic matter in the digestate as well as from CO₂ in the air supplied to the culture.

Biomass analysis indicated that the primary nutrients utilized from the digestate for biomass production were sulfur, nitrogen, phosphorus, and carbon. The liquid fraction of digestate after C. vulgaris growth contained on average 85 ± 3% and 78 ± 5% lower ammonium nitrogen and phosphorus content. This composition highlights the potential of utilizing liquid digestate as a medium for cultivating nutrient-rich algae. The significant nitrogen content is particularly noteworthy, as it is a critical nutrient for algal growth and protein synthesis. The effective uptake of nitrogen by C. vulgaris demonstrates the potential of using digestate as a nutrient source in a circular economy framework. By recycling nutrients from digestate, the need for synthetic fertilizers can be reduced, thereby lowering the environmental footprint of agricultural practices, and promoting sustainable nutrient management.

The immediate and robust growth observed at lower digestate concentrations suggests that optimizing the dilution ratio is crucial for maximizing algal growth and productivity. In contrast, the lack of growth in S. acuminatus highlights the importance of strain selection based on the specific nutrient and inhibitory profiles of the growth medium. Future studies should focus on identifying and mitigating inhibitory components in the digestate, as well as exploring a broader range of algal strains to fully leverage the nutrient potential of the digestate. Overall, our findings demonstrate the potential for integrating algal cultivation with waste management practices, contributing to a circular bioeconomy by converting waste streams into valuable biomass.

In the context of a circular bioeconomy, nutrient-rich digestate derived from anaerobic co-digestion of BW and FW can be efficiently pretreated using a process of flocculation followed by sedimentation. The resulting liquid fraction of the digestate may then serve as a medium for cultivating microalgae biomass. C. vulgaris, in particular, has been widely studied for its lipid production capacity, which can be utilized in biofuel production (Moradi & Saidi 2022). Depending on its composition, the harvested microalgal biomass could be used in biorefinery for bioenergy, biopolymer production, and as a biofertilizer. However, for industrial-scale microalgae cultivation for biorefinery applications to be economically viable, outdoor cultivations in open ponds must be thoroughly investigated.

AD of organic matter typically produces biogas containing 30–45% CO₂, depending on the process's efficiency. Harnessing this biogenic CO₂ is crucial for enhancing the sustainability and efficiency of the process. It can be injected into the microalgae culture as a carbon source (Tripathi et al. 2023) or upgraded to methane via biological hydrogen methanation (Kamravamanesh et al. 2023). In addition, sludge treatment remains a persistent challenge in waste management. AD is commonly employed to treat sewage sludge and food waste, resulting in digestate that can be utilized in agriculture, soil amendments, or composting, provided contaminant levels are within regulatory limits. However, reject water separated from the solid fraction of digestate poses a significant issue due to its high nitrogen content. Directing this water to wastewater treatment plants for nitrogen removal is costly and increases energy and chemical consumption.

Treating reject water through microalgae cultivation offers several advantages: it can reduce nitrogen load and associated treatment costs, and microalgae can efficiently utilize the concentrated nutrients in the reject water. This approach maintains the existing treatments required for the solid digestate. Moreover, using BW instead of sewage sludge or municipal wastewater for microalgae cultivation is advantageous due to the higher concentration of nutrients in forms readily available for microalgal uptake.

This study demonstrated the potential of anaerobic co-digestion of source-separated blackwater and food waste in a 2 m³ pilot-scale CSTR, attaining a CH₄ yield of 0.7 ± 0.1 m³/kg-VS with VS removal efficiency of 82 ± 6%. Beyond biomethane production, the resulting nutrient-rich digestate holds significant value. Flocculation and subsequent sedimentation preserved over 90% of NH₄-N and other macronutrients, along with more than 75% of total phosphorus, in the liquid fraction of digestate. This pretreatment, combined with dilution, facilitated the growth of C. vulgaris on diluted digestate. Biomass analysis indicated that carbon, nitrogen, and sulfur were predominantly utilized by the algae for biomass production and this treatment reduced CODt and CODs of digestate by 50 ± 10%. The liquid fraction of digestate after C. vulgaris growth contained on average 85 ± 3% and 78 ± 5% lower NH₄-N and PO₄-P content. The successful cultivation of C. vulgaris on liquid digestate underscores the feasibility of utilizing waste-derived nutrients for algal biomass production, thereby presenting a sustainable approach to nutrient recycling.

The authors wish to express their gratitude to the Kaute Foundation for funding this research. In addition, we acknowledge the city of Tampere for providing the necessary space, BW from vacuum toilets, and a pilot-scale biogas reactor for the experiments. Special thanks are also extended to Anni Kelola for her diligent work in pretreating the feedstocks and assisting with both laboratory and pilot setups, and to Antti Nuottajärvi for his valuable assistance in operating the pilot-scale biogas reactor.

D.K. and M.K. conceptualized the study and planned the experiments. D.K. performed most of the experiments, the data analysis, as well as the data visualization, and wrote the original draft of the manuscript. M.K. performed the project administration. D.K. and M.K. reviewed and edited the manuscript and read and approved the final manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.