Closing the loop in aquaponics: biological and physicochemical recovery of nutrients from aquaculture sludge for circular bioeconomy

Aquaculture sludge from recirculating systems contains a substantial share of feed-derived nutrients, yet it is typically discarded despite nutrient limitations in aquaponics. In this study, tilapia sludge from a commercial RAS was treated by sealed storage, gradual acidification (pH 5), ultrasonication, aerobic mineralization, anaerobic digestion (AD), and two AD-based sequential approaches (AD pH, AD US pH). Total and dissolved fractions were quantified to assess nutrient distribution between solid-bound and dissolved pools, and treatment-dependent mineralization performance was calculated. In untreated sludge, >75–90% of nitrogen, phosphorus, calcium, magnesium, and most trace elements were associated with the solid phase, whereas potassium was largely dissolved. AD variants achieved the highest nitrogen mineralization (up to ~47%) with ammonium-dominated supernatants (≈460–480 mg NH 4 –N L − ¹), while aerobic mineralization promoted nitrification and yielded nitrate-dominant nitrogen (≈257 mg NO 3 –N L − ¹). Phosphorus mobilization was primarily pH-driven, peaking under acidification (≈165 mg PO 4 –P L − ¹) and increasing after post-acidification of AD supernatants. Micronutrient responses were element-specific; manganese increased markedly under acidified and unaerated conditions, whereas dissolved iron remained low across treatments. Compared to ideal hydroponic nutrient media, most supernatants required dilution due to elevated ammonium, and iron supplementation remained necessary. Most potentially toxic elements remained in the low µg L − ¹ range and below irrigation and reported algal toxicity thresholds. Selected sludge-derived nutrient media were suitable for Arthrospira platensis cultivation to support circular nutrient reuse. In growth trials, the non-acidified AD supernatant (AD S) supported the strongest biomass and pigment performance among sludge-derived media, reaching up to 1.32 g DW L − ¹, with a chlorophyll a content of 21.7 mg L − ¹ and a C-phycocyanin content of 24.6 mg g DW − ¹. Overall, sludge mineralization shifted nutrients from bound to dissolved pools and generated distinct nutrient profiles, highlighting the importance of aligning treatment strategy with the intended downstream application in circular RAS–aquaponic systems.

Aquaculture is among the world’s fastest-growing food sectors, averaging around 5% annual growth since 2000, but recently slowing down to about 1.6% per year, with this slowdown linked to environmental and economic constraints (FAO, 2024). Both land-based recirculating aquaculture systems (RAS) and sea-cage production generate waste streams that can affect the receiving water bodies (Wu and Song, 2022; Zhang et al., 2023; Pettersen et al., 2025). From an economic perspective, waste management of nutrient-rich byproducts can account for 30–50% of total production costs in RAS, representing a substantial economic burden (Ende et al., 2024).

This burden is also an opportunity, however, since the same waste stream concentrates much of the nutrient load it is costly to dispose of. The solid fraction of aquaculture effluents, referred to as aquaculture sludge, typically consists of uneaten feed and fish feces. In RAS, this fraction is commonly separated by a drum filter and discarded. Aquaculture sludge contains substantial amounts of macronutrients (N, P, K, Mg) and essential micronutrients (Fe, Mo, Zn, and others) for plant growth (Aboutboul et al., 1995; Chen et al., 1997; Monsees et al., 2017; Goddek et al., 2018; Delaide et al., 2019; Zhang et al., 2021; Sele et al., 2024). Sludge nutrient content is strongly linked to fish feed composition, cultivated species, system design, and management (Pettersen et al., 2025). In general, up to 40% of the nutrients contained in feed can be retained in sludge (Yogev et al., 2016). Regarding phosphorus, more than 50% of phosphorus from fish feed ends up in the sludge (Montanhini Neto and Ostrensky, 2015).

Yet these nutrients remain largely bound in the solid phase. Aquaponic systems – integrating RAS with hydroponic plant production – seek to utilize dissolved nutrients in aquaculture process water to increase resource efficiency (Kloas et al., 2015; Palm et al., 2018; Goddek et al., 2019). However, aquaculture process water is generally deficient in several macro- and micronutrients compared to optimal requirements for hydroponic growth. The extent of required supplementation depends critically on production intensity, crop selection, and system management and leads to a variable nutrient self-sufficiency of aquaponic systems (Goddek et al., 2015, 2019; Bittsanszky et al., 2016; Delaide et al., 2016; Nozzi et al., 2018; Zhang et al., 2021).

These deficiencies are element-specific. Nitrogen in aquaculture process water is often sufficient but imbalanced, with limited ammonium and suboptimal NH 4:NO 3 ratios (Pineda-Pineda et al., 2017; Eck et al., 2019; Aslanidou et al., 2023). Additionally, phosphorus availability is frequently low due to particle binding (Cerozi and Fitzsimmons, 2017; Monsees et al., 2017; Nozzi et al., 2018; Lobanov et al., 2021). Potassium, calcium, magnesium, and several micronutrients are commonly deficient, with iron consistently limiting and zinc, molybdenum, copper, and manganese often below optimal levels, requiring supplementation (Bittsanszky et al., 2016; Yang and Kim, 2020; Stathopoulou et al., 2021; John et al., 2022; Tsoumalakou et al., 2022; Farooq et al., 2023). These limitations highlight the need for alternative nutrient sources, particularly those that enable recovery of nutrients currently retained in solid waste streams such as aquaculture sludge (Zhang et al., 2021; Zhang et al., 2023; Lopchan Lama et al., 2025; Nishanth et al., 2025).

One avenue for using such recovered nutrients is the integration of additional organisms into aquaponic and recirculating aquaculture systems, which receives increased attention within the circular economy framework (Ende et al., 2024; Lopchan Lama et al., 2025). Proposed integrations include duckweed (Paolacci et al., 2021, 2022), microalgae and cyanobacteria (Egloff et al., 2018; Ramli et al., 2020), and zooplankton (Gorzelnik et al., 2023), which can convert residual nutrients into valuable biomass, particularly for feed applications. The cyanobacterium Arthrospira platensis is a promising candidate for integration due to its high protein content, rapid growth, and established use as food and feed biomass (Becker, 2007; Habib et al., 2008). Several studies have shown that Arthrospira spp. can be cultivated on alternative and recycled nutrient sources, including agro-industrial effluents and other nutrient-rich waste streams, reducing production costs while maintaining biomass productivity when nutrient composition and inhibitors are properly managed (Markou and Georgakakis, 2011; Madkour et al., 2012; Markou, 2015; Abdelhay et al., 2019; Papadopoulos et al., 2022; Villaró-Cos et al., 2024). In aquaculture systems, recycled process waters and sludge-derived nutrient fractions may provide suitable media due to their high nitrogen, phosphorus, and trace element content (Addy et al., 2017; Egloff et al., 2018; Arumugam et al., 2020; Ende et al., 2024; Nishanth et al., 2025; Zieliński et al., 2025). However, because many nutrients in aquaculture sludge are not readily bioavailable, efficient processes to liberate and convert them into dissolved forms remain a key research challenge (Eck et al., 2019; Zhang et al., 2021; Nishanth et al., 2025).

Mineralization processes convert organic matter into plant-available inorganic forms (Weil and Brady, 2017). Aerobic mineralization involves the oxygen-dependent breakdown of organic matter by heterotrophic microorganisms, and significant nutrient mobilization from aquaculture sludge has been reported, although comprehensive assessments of micronutrients and potential contaminants remain limited (Monsees et al., 2017; Delaide et al., 2018, 2021; Khiari et al., 2019; Panana et al., 2021; Harabiš et al., 2025).

Anaerobic digestion (AD) is another widely used sludge treatment method in which microbial degradation occurs under oxygen-free conditions (Mirzoyan et al., 2010). While AD has traditionally been applied in aquaculture for waste reduction and biogas production, more recent studies have highlighted its potential for macro- and micronutrient recovery (Monsees et al., 2017; Delaide et al., 2018; Goddek et al., 2018; Zhu et al., 2023; Paes et al., 2025). In addition, passive storage of aquaculture sludge in sealed containers can naturally create anaerobic conditions due to microbial degradation (Monsees et al., 2017), and is therefore investigated here as a simple and cost-effective mineralization approach.

Beyond these biological routes, chemical conditions provide a further lever for nutrient release: nutrient solubility and plant availability are strongly pH-dependent (Resh, 2022; Ahmed et al., 2024; Xia et al., 2024). Acidification can enhance nutrient mobilization through leaching processes driven by organic acids from microbial decomposition or by external acid addition (Delaide et al., 2019). Previous studies reported up to 90% nutrient mobilization at pH 4 (Jung and Lovitt, 2011) and up to 86% phosphorus recovery through acidification (Schleyken et al., 2024). Because microbial mineralization is most effective around neutral pH (6–8), while nutrient leaching increases under acidic conditions, a two-stage approach combining biological mineralization followed by acidification has been proposed (Goddek et al., 2018; Delaide et al., 2019).

Ultrasonication (US) represents another potential sludge treatment method. The formation and collapse of microbubbles during ultrasonic treatment generate strong mechanical forces that disrupt flocs and microbial cells, thereby releasing previously bound compounds (Pilli et al., 2011; Le et al., 2015). Ultrasonication of aquaculture sludge has been shown to improve dewaterability, enhance phosphorus release, and increase biodegradability and biogas production when used as a pretreatment before AD (McDermott et al., 2001; Luo et al., 2015; Martínez et al., 2015; Lin et al., 2020). However, its effects on macro- and micronutrient mobilization, either alone or combined with AD, remain poorly studied.

In this study, a broad set of sludge treatment strategies that have rarely been compared under identical conditions are evaluated side-by-side, including sealed storage, gradual acidification, ultrasound disruption, anaerobic digestion, aerobic mineralization, and two sequential approaches combining anaerobic digestion with post-treatment (AD pH and AD US pH). Their effects on the mobilization of macro- and micronutrients, as well as potential contaminants, are assessed, and the resulting supernatants are compared to conventional hydroponic nutrient solution to evaluate their suitability as alternative nutrient sources. In addition, selected supernatants are adapted and tested as growth media for Arthrospira platensis to assess their capacity to support microbial biomass production. The objective is to identify treatment-dependent differences in nutrient mobilization from aquaculture sludge and to evaluate their potential for supporting plant production and additional biomass modules within circular RAS–aquaponic systems.

Sludge collection

Aquaculture sludge was collected from the same commercial tilapia (Oreochromis niloticus) recirculating aquaculture system, and following the same procedure, as described in Elshobary et al. (2026). In brief, sludge separated by drum filtration (Tiny Drum L4, Spranger) was settled in 50 L drums for 24 h and then in a 100 L conical container for a further 24 h, after which the upper and lower layers were combined, homogenized by manual stirring, and used for the experiments. The RAS was operated with a pH-control base stock solution (3.5 g L − ¹ K 2 CO 3, 2.5 g L − ¹ CaCO 3, 3 g L − ¹ NaHCO 3).

Sludge mineralization treatments

An overview of the different sludge mineralization methods and the design of this study are presented in Figure 1.

Aerated treatment at room temperature (AT RT)

Aerobic mineralization was performed by filling 400 mL of homogenized sludge into 1 L glass bottles placed on a magnetic stirrer (300 rpm) aerated by an air pump. Air was supplied through a porous glass diffuser (35 mm diameter) inserted from the top of the bottles. The airflow was set to 1.5 L min -1 and dissolved oxygen levels were monitored (Hach HQ40d) three times per week to ensure aerobic conditions (DO > 2 mg O 2 L -1). The experiment was performed in triplicate and run for 28 d at 22 ± 1 °C. Water loss due to evaporation was equalized by filling the treatments up to their initial volume with demineralized water prior to sampling and analysis. No seed sludge was used as inoculum to focus on the sludge-native microbiome and avoid external nutrient input.

Unaerated treatment at room temperature (UT RT)

Sealed storage of aquaculture sludge was performed using three identical 1 L Erlenmeyer flasks, filled with 400 mL homogenized sludge and sealed with water-filled fermentation air locks, allowing produced gases to exit, while no air was introduced into the containers. Storage conditions were identical to the AT RT treatment described above under a laboratory hood at 22 ± 1 °C for 28 days.

Ultrasound treatment (US)

Mechanical disruption by ultrasound was performed by filling 400 mL of sludge into a 1 L beaker. The beaker was placed in a styrofoam box filled with ice to avoid heating of the substrate. The sonotrode (Hielscher UP 100H) was equipped with the micro tip MS7 (Hielscher), operated at 30 kHz with 100 W nominal output, set to the maximum intensity (100%) and run for 1 h per replicate. The tip of the sonotrode was placed 1 cm above the bottom of the beaker according to the manufacturer’s manual. The experiment was performed in triplicate and temperature was kept at 24 ± 2 °C.

Treatment by pH-adjustment (pH)

Acidification was performed by filling 400 mL of homogenized sludge into 1 L Erlenmeyer flasks, placed on a magnetic stirrer (300 rpm) under a laboratory hood at 22 ± 1 °C. The pH was lowered to 5 ± 0.02 by adding 1 M HCl in steps of 1 mL. After 6 h the resulting pH was recorded and again lowered to 5 ± 0.02 in the same way as described previously. This was repeated after 24 h for all three replicates. The samples were analyzed after another 24 h.

Anaerobic digestion (AD S, AD pH, AD US pH)

Mineralization by anaerobic digestion was performed as a batch fermentation test based on VDI 4630 in a lab-scale batch fermentation setup (Ritter GmbH & Co. Kg, Germany) consisting of a heating cabinet which holds eight identical fermentation vessels. Each vessel contained a stirring mechanism, set at 15 rpm and was connected to an individual Milli-Gas-Counter (MGC-1 PMMA, Ritter GmbH & Co. KG, Bochum, Germany). The volume of the produced biogas was recorded and converted to NL automatically by the software RIGAMO (Ritter GmbH & Co. Kg, Germany). The temperature was set to 37 °C and the experiment ran for 28 days. Each container was filled with the specific sludge type as stated below and allowed to preheat for 2 h inside the cabinet prior to the start of the experiment. Five vessels were filled with 400 mL homogenized sludge (AD S). After digestion, the digestate of each vessel was divided into two shares, either analyzed directly (AD S) or post-treated by acidification (AD pH) as stated earlier. Additionally, three vessels were filled with sludge pre-treated by ultrasonication treated as stated earlier, in a ratio of 3:1 treated sludge to untreated sludge by volume and post-treated with acidification (AD US pH). This approach was used to ensure sufficient microbial activity, as no additional inoculum was introduced, as described in Section 2.2.1.

Analytical methods

Initial sludge analysis

The sludge investigated here was the same batch characterized in the companion study on the same material (Elshobary et al., 2026); its bulk physicochemical parameters — total solids (TS), volatile solids (VS), pH and electrical conductivity (EC) — were determined once on this batch, as described there (TS by drying at 105 °C to constant weight, VS by ignition at 550 °C, and pH and EC with a portable multiparameter meter), and are common to both studies. Total nitrogen (TN) and the elemental composition were determined independently in the present work: TN according to DIN 51732:2014-07 (adapted for dried aquaculture sludge), and the elemental composition according to DIN EN 16171:2017–01 following microwave-assisted acid digestion and inductively coupled plasma mass spectrometry (ICP-MS; Agilent 7850). Dissolved elements in the untreated sludge were analyzed in the same way as the supernatants of the treated sludge (described below).

Sample preparation and analysis

After undergoing the specific treatment, pH, TS, and VS were measured as described above.

Sample preparation prior to elemental analysis consisted of centrifugation at 43,500 × g for 10 min and filtration of the supernatant through a 0.45 µm filter (ROTILABO, Carl Roth GmbH & Co. KG Karlsruhe, Germany).

Dissolved nutrients were directly analyzed by Hach Lange cuvette kits as follows: NO 3 –N LCK 340/339; NO 2 –N LCK341/342; NH 4 –N LCK304; PO 4 –P LCK049; K LCK228 according to the manufacturer’s standard working procedure via a UV/VIS spectrophotometer (Hach Lange DR5000).

Sample conservation consisted of stabilization using 1% v/v HNO 3 to reach a pH < 2 and storage at -20 °C.

Elemental composition of the supernatants was analyzed by ICP-MS (Agilent MS 7850) at the end of the experiment.

Calculations

The total concentration of an element in the wet sludge was calculated based on total content measured in the dried sludge sample as follows:

Where c t o t, L is the total concentration of an element in a sludge volume (wet) in mg L -1; c t o t, T S is the total concentration of an element in dry matter in mg kgTS -1; ρ s l u d g e is the density of (wet) sludge in kg L -1 and TS is the dry matter content of the sludge in kgTS kg -1.

The fraction of the nutrient content bound in solids was calculated as follows:

Where C b, L is the concentration of an element bound in the solid phase per wet sludge volume in mg L -1; c t o t, L is the total concentration of an element in a sludge volume (wet) in mg L -1; c s, L is the dissolved content of an element measured in the supernatant in mg L -1.

The nutrient mineralization performance was calculated according to (Delaide et al., 2018) as follows:

Where η N is the nutrient mineralization performance at the end of a treatment in %; D N o u t is the mass of dissolved nutrient after treatment, derived from measured concentrations and the reactor liquid volumes; D N i n is the mass of dissolved nutrient before treatment, calculated analogously; and T N i n is the total mass of nutrient contained in the untreated sludge volume used per reactor.

Cultivation of Arthrospira platensis in media based on treated aquaculture sludge

Media preparation and composition

For the first trial, five experimental media (AT RT, UT RT, AD S, AD pH, and AD US pH) were selected according to their nitrogen and phosphorus contents for Arthrospira platensis cultivation. The media were prepared using the obtained supernatants from aquaculture sludge after undergoing the specific treatments described above by the addition of 10 g L -1 sodium bicarbonate (NaHCO 3) as a source of dissolved inorganic carbon (DIC) and buffering agent, consistent with the SAG protocol for A. platensis.

The experimental media that did not show any growth were modified by diluting 1:1 (v/v) with sterile deionized water to reduce the initial ammonium (NH 4 –N) concentration. The diluted media were subsequently supplemented with 10 g NaHCO 3, as well as KNO 3, and K 2 HPO 4 to achieve total nitrogen (TN) and total phosphorus (TP) concentrations equivalent to those of the standard SAG medium (TN = 412 mg N L − ¹; TP = 88.9 mg P L − ¹) as shown in Table 1.

The final concentrations of nitrate (NO 3 –N), phosphate (PO 4 –P), and potassium (K) were adjusted to match the composition of the SAG medium as displayed in the results section. The required amounts of KNO 3 and K 2 HPO 4 were calculated based on the measured concentrations of NH 4 –N, NO 3 –N, and PO 4 –P in the diluted supernatants. The additions ensured that the sum of NH 4 –N (present in the supernatant) and NO 3 –N (supplied as KNO 3) equaled the target TN (412 mg L − ¹), while total phosphate reached 88.9 mg L − ¹ PO 4 –P. The K supplied from both KNO 3 and K 2 HPO 4 was included in the potassium balance.

The final per 200 mL additions of KNO 3 and K 2 HPO 4 were: UT RT, 0.350 and 0.022 g; AD S, 0.257 and 0.076 g; AD pH, 0.264 and 0.030 g; and AD US pH, 0.247 and 0.062 g, respectively. All media were supplemented with 10 g L -1 sodium bicarbonate. Salts were weighed using an analytical balance (± 0.1 mg) and dissolved in the diluted supernatant under continuous stirring. The pH of all media was adjusted to 8.0 ± 0.1 using 0.1 M HCl or 0.1 M NaOH prior to sterilization. All media were sterilized by autoclaving at 121 °C for 20 min, then cooled to room temperature to be suitable for algal cultivation.

Cultivation and growth estimation

A 10% (v/v) inoculum from an exponential-phase Arthrospira platensis stock culture was added to each growth medium in triplicate and cultivated under a 12:12 h light: dark regime at 30 ± 2 °C. Continuous aeration was provided via air pumps at a flow rate of 0.5–1 vvm (volume of air per volume of culture per minute). A photograph of the cultures grown on the different media is provided in Supplementary Figure 6.

Biomass dry weight was determined every week by centrifuging a specific volume of the cultures at 4000 × g for 20 min, washing the pellet twice with distilled water, and collecting the pellet onto pre-weighed Whatman GF/F filters. Filters were dried at 60 °C to constant weight, and biomass concentration was calculated gravimetrically as g DW L − ¹. Biomass productivity was estimated according to (Elshobary et al., 2025):

where DW f and DW i were the dry weight (mg L -1) at time T f and T i, respectively.

Chlorophyll a and carotenoid contents were measured using a DMSO extraction method adapted from (Carreto and Catoggio, 1977). Cell pellets, obtained by centrifugation of the culture at 4000 × g for 15 min, were resuspended in equal volume of DMSO (typically 90% v/v, 1:1 v/v), incubated at 65 °C for 15 min in the dark with occasional vortexing, and recentrifuged at 10,000 × g for 5 min to clarify the supernatant. Absorbance of the supernatant was measured at 480, 645, and 663 nm using a UV-VIS spectrophotometer (Biochrom Libra S70), and concentrations (µg mL −1) were calculated with the following equations:

Phycocyanin measurement

C phycocyanin was quantified spectrophotometrically from Arthrospira platensis cultures using optical density measurements at 615 and 652 nm with a Biochrom Libra S-70 UV-VIS Spectrophotometer. Phycobiliproteins were extracted using a double freeze-thaw method in phosphate buffer (pH 8, 1:10 w/v), followed by mechanical disruption with glass beads and centrifugation at 4000 × g for 20 min at 4 °C. Phycocyanin concentration (mg mL − ¹) was calculated according to (Hsieh-Lo et al., 2019):

Statistical analysis

Prior to statistical analysis, normality of model residuals was assessed using the Shapiro–Wilk test and homogeneity of variances using Levene’s test. While deviations from normality were observed for several variables, homogeneity of variances was generally satisfied. Given the balanced experimental design and equal replication among treatments (n = 3), differences among treatments were evaluated using one-way analysis of variance (ANOVA) followed by Tukey’s HSD post-hoc test. Statistical significance was accepted at p < 0.05.

Statistical differences among treatments were evaluated for dissolved nutrient concentrations, selected mineralization efficiency parameters (ηN, ηP, ηK, ηMg, ηB, ηMn, ηNi, ηZn), and biological response variables (biomass productivity and phycocyanin content). Physicochemical parameters (TS, VS, pH, EC) and selected mineralization efficiencies (ηCa, ηS, ηFe) were not included in statistical testing due to their low variability or limited interpretative relevance for treatment comparison, and are therefore presented descriptively.

Initial sludge characteristics and nutrient distribution

Elemental analysis of the initial untreated sludge and supernatant is shown in Table 2. On a dry matter basis, total nitrogen (TN) amounted to 37.5 ± 1.3 g kgTS − ¹ and total phosphorus (TP) to 18.0 ± 0.6 g kgTS − ¹. Calcium, potassium, magnesium, and sulfur were measured at 123.3 ± 4.7, 10.5 ± 0.25, 3.0 ± 0.14, and 5.3 ± 0.16 g kgTS − ¹, respectively. Among micronutrients, iron (1233 ± 189 mg kgTS − ¹), zinc (531 ± 9 mg kgTS − ¹), manganese (222 ± 1 mg kgTS − ¹), and copper (47.7 ± 1.3 mg kgTS − ¹) were present at substantial concentrations.

In the initial supernatant, nitrogen was predominantly present as NH 4 –N (11.6 ± 0.3 mg L − ¹), while NO 2 –N and NO 3 –N were negligible. Dissolved PO 4 –P concentration was 45.25 ± 0.05 mg L − ¹. Potassium (231 ± 3 mg L − ¹) and calcium (310.6 ± 1.6 mg L − ¹) dominated the dissolved macronutrient fraction. The untreated aquaculture sludge exhibited a total solids (TS) content of 2.80 ± 0.03% and a volatile solids (VS) fraction of 69.41 ± 0.14% TS. The initial pH was 7.10 ± 0.02 and electrical conductivity (EC) was 1.19 ± 0.04 mS cm − ¹ (Table 3).

The distribution of elements between the liquid supernatant and the solid phase (Figure 2) showed that nitrogen (98.8%), phosphorus (90.7%), calcium (90.6%), magnesium (74.7%), and sulfur (85.2%) were predominantly bound to the solid phase, whereas potassium was largely present in dissolved form (81.7%). For micronutrients, more than 90% of copper, iron, manganese, nickel, and zinc were associated with the solid fraction, while boron and chloride exhibited high dissolved shares (Figure 2).

Effects of treatments on physicochemical properties and solids reduction

Total solids (TS), volatile solids (VS), electrical conductivity (EC), and pH before and after treatment are shown in Table 3 and Supplementary Figure 1. Acidification (pH) and ultrasound (US) did not change TS relative to the untreated sludge (2.80%), and VS fractions remained comparable (69.98 ± 0.36% and 69.63 ± 0.62% TS, respectively). Unaerated storage (UT RT) resulted in only minor TS reduction to 2.72%. Aerobic mineralization (AT RT) resulted in the strongest solids reduction, with TS decreasing to 0.77 ± 0.12% and VS to 45.97 ± 1.39% TS (Table 3; Supplementary Figure 1). Anaerobic digestion treatments achieved moderate TS reductions, ranging from 31.8% to 40.7%, with the highest reduction observed in AD US pH (Table 3; Supplementary Figure 1).

The pH differed strongly among treatments (Table 3; Supplementary Figure 1). Acidification lowered pH to 5.12–5.36, while anaerobic digestion without post-treatment increased pH to 8.00 ± 0.02. Electrical conductivity increased across all treatments, with the highest EC recorded in AD pH (7.54 ± 0.12 mS cm − ¹) and AD US pH (7.71 ± 0.21 mS cm − ¹) (Table 3; Supplementary Figure 1).

Nutrient mobilization and mineralization performance

Dissolved concentrations of macro- and micronutrients following treatment are presented in Figures 3 and 4, while complete numerical datasets are provided in Supplementary Table 1, and statistical groupings are presented in Supplementary Table 2. Mineralization performance (ηN) is displayed in Supplementary Figure 2 for macronutrients and Supplementary Figure 3 for micronutrients. Several elements were below the limit of detection, preventing the calculation of ηN. Molybdenum (Mo < 2 mg kgTS − ¹), lead (Pb < 2 mg kgTS − ¹), and arsenic (As < 0.8 mg kgTS − ¹) were below detection limits in the untreated sludge. The dissolved fractions in the supernatants of mercury (Hg < 0.07 µg L − ¹) and aluminum (Al < 0.07 µg L − ¹) were below detection limits across treatments.

Macronutrients

Nitrogen

Nitrogen speciation and concentration varied significantly across treatments (Figure 3). The lowest dissolved NH 4 –N was observed in AT RT (0.08 ± 0.03 mg L − ¹), while NO 3 –N reached 257.07 ± 64.13 mg L − ¹. Among non-biological treatments, NH 4 –N increased to 26.70 ± 1.02 mg L − ¹ under US and to 47.35 ± 8.13 mg L − ¹ under pH adjustment. UT RT resulted in substantially higher NH 4 –N (338.00 ± 7.48 mg L − ¹), while the highest NH 4 –N concentrations occurred in the anaerobic digestion treatments, which did not differ significantly from one another (457.64 ± 17.08–482.32 ± 7.94 mg L − ¹) (Figure 3).

Nitrogen mineralization performance (ηN) ranged from 1.6 ± 0.2% (US) to 47.1 ± 0.9% (AD US pH) (Supplementary Figure 2). Anaerobic digestion treatments formed a statistically homogeneous group with the highest ηN values (44.6–47.1%), significantly exceeding all other treatments. Intermediate ηN values were recorded for UT RT (32.7 ± 0.8%) and AT RT (24.6 ± 6.5%), while significantly lower values were observed for pH adjustment (3.6 ± 0.9%) and US (1.6 ± 0.2%).

Phosphorus

The highest PO 4 –P concentrations were observed after pH adjustment (164.67 ± 7.13 mg L − ¹) and in UT RT (138.73 ± 47.34 mg L − ¹), which did not differ significantly from one another. AD pH (125.02 ± 2.55 mg L − ¹) formed an intermediate group, and US as well as AD US pH showed lower intermediate concentrations; AT RT, AD S, and the initial supernatant did not differ significantly from one another (Figure 3).

The highest ηP was observed for pH adjustment (27.2 ± 1.6%), which did not differ significantly from UT RT (21.3 ± 10.8%), while AD pH (18.2 ± 0.6%) formed an intermediate group. Lower positive ηP values were observed for US (5.1 ± 0.4%) and AD US pH (5.1 ± 0.3%), while slightly negative ηP values occurred for AT RT (−3.6 ± 5%) and AD S (−0.7 ± 0.6%), both representing the lowest group (Supplementary Figure 2).

Potassium, calcium, magnesium, and sulfur

Potassium concentrations varied within a narrower range across treatments and showed only limited significant differences between groups (Figure 3). The highest dissolved K was measured in AD pH (282.17 ± 0.4 mg L − ¹) and UT RT (273.93 ± 3.32 mg L − ¹), while other treatments ranged from 239–271 mg L − ¹ (Figure 3). Potassium mineralization performance (ηK) ranged from 16.3 ± 2% (US) to 99.1 ± 2.4% (AD pH) (Supplementary Figure 2).

Calcium showed the lowest dissolved concentration under AD S (126.35 ± 2 mg L − ¹) and the highest values after post-acidification (AD pH and AD US pH: 345–346 mg L − ¹) (Figure 3). Calcium mineralization performance (ηCa) remained low across treatments, with values below 1% for pH adjustment, UT RT, and US, and slightly negative values for AD S (−6.1 ± 0.1%) and AT RT (−2.6 ± 2.3%). Post-acidification treatments (AD pH, AD US pH) reached approximately 1% (Supplementary Figure 2).

Magnesium concentrations were highest under UT RT (58.27 ± 0.22 mg L − ¹), AD pH (57.59 ± 0.26 mg L − ¹), and AD US pH (55.34 ± 0.29 mg L − ¹), which did not differ significantly from one another, while other treatments showed significantly lower values (Figure 3). Magnesium mineralization performance (ηMg) ranged from 6.2 ± 0.7% (US) to 62.8 ± 0.5% (UT RT), with similarly high values for AD pH and AD US pH. These treatments did not differ significantly from one another and were significantly higher than US and AT RT (Supplementary Figure 2).

Sulfur showed the highest dissolved concentrations under AT RT (46.3 ± 3.78 mg L − ¹), while UT RT and all AD treatments remained at very low levels (≤0.5 mg L − ¹) (Figure 3). Sulfur mineralization performance (ηS) was negative for UT RT and all AD treatments (−17%), while positive values were recorded for AT RT (20.7 ± 3.1%) and US (2.3 ± 0.9%); pH adjustment resulted in −5.1 ± 1.2% (Supplementary Figure 2).

Micronutrients

Dissolved micronutrient concentrations changed strongly across treatments (Figure 4). Iron showed the highest dissolved concentration under UT RT (1282.67 ± 52.18 µg L − ¹) and the lowest under AT RT (9.36 ± 1.14 µg L − ¹). In AD treatments, iron ranged 202–263 µg L − ¹ (Figure 4). Iron mineralization performance (ηFe) remained low across treatments (−0.74% to 3.13%), with the highest ηFe recorded for UT RT (3.13% ± 0.16%) (Supplementary Figure 3).

Manganese showed the highest dissolved concentrations under pH adjustment (1956.33 ± 90.16 µg L − ¹), followed by UT RT and AD US pH, which formed a significantly lower but still elevated group. In contrast, AT RT and AD S yielded significantly lower manganese concentrations, 20.85 ± 24.65 µg L − ¹ and 34.53 ± 1.07 µg L − ¹, respectively (Figure 4). Manganese mineralization performance (ηMn) ranged from −6.33% ± 0.44% (AT RT) and −6.09% ± 0.02% (AD S) to 28.18% ± 1.61% (pH adjustment) (Supplementary Figure 3).

Zinc concentrations increased markedly under acidified and digestion-based treatments, with the highest dissolved concentration observed in AD US pH (379.04 ± 87.03 µg L − ¹). AD pH (284.51 ± 61.15 µg L − ¹) and pH (246.53 ± 50.21 µg L − ¹) formed an intermediate high group, while US (138.83 ± 17.31 µg L − ¹) showed moderate concentrations. In contrast, the initial supernatant, UT RT, AT RT, and AD S remained in the lower group and did not differ significantly from one another (Figure 4). Zinc mineralization performance (ηZn) remained low overall, with the highest ηZn observed in AD US pH (2.50% ± 0.61%), followed by AD pH and pH, while the remaining treatments showed similarly low values (Supplementary Figure 3).

Nickel concentrations increased during anaerobic digestion, reaching the highest values in AD S (45.79 ± 0.81 µg L − ¹). AD pH and AD US pH showed moderately elevated concentrations, while the remaining treatments remained at lower levels (Figure 4). Nickel mineralization performance (ηNi) ranged from −0.78% to 19.86%, with the highest ηNi observed in AD S, which was significantly higher than all other treatments, while the remaining treatments showed similarly low values (Supplementary Figure 3).

Boron concentrations increased most strongly under AT RT (0.30 ± 0.01 mg L − ¹), which was significantly higher than all other treatments, while the remaining treatments showed lower and partially overlapping values (Figure 4). Boron mineralization performance reached 91.9 ± 2.9% under AT RT, which was significantly higher than all other treatments, whereas all other treatments remained substantially lower (Supplementary Figure 3).

Copper concentrations increased particularly in AT RT (41.22 ± 7.48 µg L − ¹), which was significantly higher than UT RT and most other treatments. The remaining treatments showed moderate to low Cu concentrations with partial overlap between groups (Figure 4). Copper mineralization performance followed a similar pattern, with AT RT showing the highest values, while the remaining treatments did not differ strongly from one another (Supplementary Figure 3).

Mobilization of potentially toxic elements

Dissolved concentrations of potentially toxic elements varied among treatments (Table 4). Arsenic ranged from 1.70 ± 0.37 µg L − ¹ in AT RT up to 8.51 ± 0.10 µg L − ¹ in AD pH and 8.49 ± 0.23 µg L − ¹ in AD US pH. Cadmium and lead remained low across treatments (0.02 – 0.40 µg L − ¹) and (0.16–1.61 µg L − ¹) respectively, while Cobalt ranged from 1.18 ± 0.19 µg L − ¹ (UT RT) to 5.41 ± 0.18 µg L − ¹ (pH). Chromium was elevated in pH (7.42 ± 0.29 µg L − ¹), while all other treatments resulted in concentrations lower or close to the initial supernatant (2.34 ± 0.48 µg L − ¹) (Table 4).

Biogas production

Cumulative biogas production followed a sigmoidal pattern in both treatments (Supplementary Figure 4). During the early digestion phase (days 10–15), AD US showed a steeper increase in cumulative biogas production than AD S. Over time, the production curves converged, and both treatments approached similar plateau levels in the later phase of digestion. Final cumulative biogas yields were 2.89 ± 0.09 NL for AD S (n = 5) and 2.91 ± 0.0005 NL for AD US (n = 3). When normalized to the organic dry matter input (7.85 g VS per reactor), the specific biogas yields achieved were 368 ± 12 NL kg − ¹ VS for AD S and 370 ± 0.06 NL kg − ¹ VS for AD US.

Growth trial with Arthrospira platensis

The growth medium significantly affected the growth and biomass productivity of A. platensis (Figure 5). Both were highest in the SAG and AD S media, where biomass productivity reached nearly 0.06 g L − ¹ d − ¹, significantly higher than in the pH-mobilization media and the UT RT medium (Figure 5).

Chlorophyll (Chl) concentrations followed the same pattern as biomass, with the highest increase in A. platensis grown in SAG and AD S media (Figure 6), whereas Chl concentrations plateaued or even declined in the pH-mobilization media by day 14 of cultivation.

Carotenoid concentration was likewise medium-dependent (Figure 6): the largest increase occurred in SAG (15.68 mg L − ¹), with markedly lower increases in the other media and the lowest in the AD pH medium (final concentration 6.21 mg L − ¹).

C-PC was again highest in A. platensis grown in SAG and AD S (24.75–26.04 mg g DW − ¹), whereas all other media remained below 20 mg g DW − ¹ (Supplementary Figure 5).

Sludge characteristics and nutrient distribution

The untreated sludge contained 2.80% total solids (TS) with a volatile fraction of 69.4% TS, indicating that the material consisted predominantly of organic matter derived from feces, uneaten feed, and associated microbial biomass. Aquaculture sludge remains highly dilute, with reported TS concentrations frequently below 1% when originating directly from drum filter backwash streams (Cripps and Bergheim, 2000; Monsees et al., 2017; Strauch et al., 2018; Zhang et al., 2021), while thickened or dewatered substrates exceeding 10–14% are documented, depending on sludge collection and pre-treatment methods (Goddek et al., 2018; Klein et al., 2024). The sludge analyzed in the present study had already undergone an initial sedimentation step prior to sampling and treatment, resulting in a higher solids concentration than raw backwash sludge but still representing a matrix dominated by water. Treating such dilute sludge means that biological or physicochemical processes effectively operate on a matrix largely composed of water, which reduces volumetric treatment efficiency (Cripps and Bergheim, 2000; Zhang et al., 2021; Paes et al., 2025).

To allow physical thickening, the sludge was stored for 48 h prior to experimental treatment. During this period, oxygen depletion within the sedimented sludge likely promoted the onset of anaerobic microbial processes (Monsees et al., 2017). Partial ammonification and organic matter degradation during storage have been reported for aquaculture sludge and may therefore have influenced the initial chemical and biological composition of the supernatant (Mirzoyan et al., 2010; Zhang et al., 2021).

Nitrogen concentrations in the sludge were comparable to those reported for tilapia sludge by Monsees et al. (2017) (38.7–44.3 g kgTS − ¹), while phosphorus concentrations were closer to values reported by Delaide et al. (2018) for tilapia sludge (17.07 ± 3.3 g kgTS − ¹). Such variability is frequently reported for aquaculture residues and reflects differences in feed composition and system management (Zhang et al., 2021; Sele et al., 2024; Pettersen et al., 2025). The comparatively high calcium and potassium concentrations observed in this study are consistent with the alkalinity management strategy applied in the RAS, where carbonate-based buffering salts were used for pH control. Similar system-specific mineral contents have been reported previously and can influence phosphorus speciation, as sludge phosphorus is often associated with calcium-rich mineral phases such as apatite-type compounds (Lin et al., 2020; Yogev et al., 2020).

The partitioning pattern observed here with nitrogen and phosphorus predominantly solid-associated, potassium showing comparatively higher solubility, and most trace elements strongly bound to solids (Figure 2) highlights the nutrient recovery potential. These findings are well supported by previous studies, stating the majority of nutrients in sludge to be solid-associated (Montanhini Neto and Ostrensky, 2015; Delaide et al., 2018; Goddek et al., 2018; Panana et al., 2021). Yogev et al. (2016) estimated that up to 40% of feed input nutrients are retained in the particulate fraction, and are therefore lost if no sludge mineralization with subsequent utilization is applied.

Treatment-dependent mineralization of major nutrients (N and P)

The side-by-side comparison of treatments revealed pronounced differences in nutrient mineralization pathways and efficiencies. Nitrogen mineralization was strongly treatment-dependent and clearly favored anaerobic digestion. All AD variants achieved ηN of 45–47%, with nitrogen almost exclusively present as NH 4 –N (Figure 3; Supplementary Figure 2), indicating efficient ammonification under reducing conditions. Goddek et al. (2018) reported nitrogen mineralization approaching ~53% in pH-neutral UASB reactors, suggesting that reactor configuration and operational stability may further enhance nitrogen transfer efficiency compared to the batch approach used here.

In contrast, the aerated treatment showed a different performance profile. Although aerobic mineralization resulted in the strongest reduction of organic matter (Supplementary Figure 1) and induced clear nitrification, with nitrogen predominantly present as NO 3 –N (Figure 3), the overall nitrogen mineralization efficiency remained lower (ηN ≈25%) (Supplementary Figure 2) than under anaerobic digestion. Delaide et al. (2018) observed nitrogen mineralization up to 58.75% in fed-batch aerobic reactors, indicating that aeration regime, pH control and reactor configuration can substantially influence nitrogen transfer from the solid to the dissolved phase. However, the mineralization performance of Panana et al. (2021) (ηN≈ 5 – 20%) is lower than the values observed in this study, despite their optimization of retention time and artificial post-acidification. By comparison, the purely physicochemical treatments mobilized only small amounts of nitrogen. Both ultrasound and acidification alone resulted in ηN below 4%, suggesting that short-term disruption of the sludge matrix is insufficient to transfer substantial amounts of organically bound nitrogen into the dissolved phase. This observation supports the interpretation that nitrogen mobilization is primarily governed by microbial decomposition processes rather than solubility equilibria, consistent with findings by Panana et al. (2021) that pH reduction enhanced phosphorus mobilization but had limited effects on nitrogen release.

Phosphorus mobilization followed a different pattern and was largely governed by pH-dependent dissolution–precipitation dynamics. Moderate phosphate release occurred during unaerated storage (UT RT), where slightly reduced conditions and gradual acidification (pH 6.34 ± 0.1) likely destabilized particulate phosphate phases. Similar fluctuations between microbial uptake, precipitation and release during sludge treatment have been previously reported (Panana et al., 2021; Harabiš et al., 2025).

In contrast, aerobic mineralization (AT RT) reduced dissolved phosphorus concentrations, indicating net removal from the dissolved pool (Figure 3). Under oxic conditions, phosphate released during organic matter degradation may be rapidly immobilized through microbial assimilation or adsorption onto newly formed iron or managense oxides, which strongly bind phosphate. Such immobilization processes are commonly reported in aerobic sludge systems, leaving room for optimization e.g. via post-acidification (Goddek et al., 2018; Panana et al., 2021; Harabiš et al., 2025).

The highest phosphorus mobilization occurred in the standalone pH treatment (Figure 3), confirming the strong influence of acidification on phosphate solubility. Increased proton availability promotes dissolution of calcium- and magnesium-bound phosphate phases (Cerozi and Fitzsimmons, 2017). Among the digestion-based treatments, AD pH further increased dissolved phosphorus, indicating that post-digestion acidification enhanced dissolution of mineral phosphate phases formed during digestion. Nevertheless, the phosphorus mobilization observed here remained moderate compared to optimized chemical recovery approaches. Schleyken et al. (2024) reported 53–61% increases in dissolved phosphorus through targeted acid re-dissolution and up to 86% overall recovery after precipitation, highlighting the additional recovery potential when chemical mobilization is combined with downstream separation processes.

Despite similar acidification, AD US pH yielded lower phosphorus mobilization than AD pH (Figure 3). Ultrasonic disruption may alter sludge structure and ion distribution prior to digestion, potentially promoting secondary precipitation of phosphate minerals during subsequent treatment (McDermott et al., 2001; Pilli et al., 2011; Panana et al., 2021).

A comparatively higher standard deviation was observed in the aerated treatment at room temperature (AT RT) for TS reduction, pH evolution, and several dissolved nutrient concentrations (Tables 3, 4; Figure 3). Although dissolved oxygen levels were continuously monitored and maintained above the defined threshold, minor differences in air distribution, bubble formation, or mixing intensity between replicates may have led to heterogeneous oxygen transfer. Because aerobic mineralization is highly sensitive to oxygen availability and micro-scale redox gradients, such variations can influence organic matter degradation and subsequent nutrient transformations (Khiari et al., 2019; Zhang et al., 2021; Harabiš et al., 2025). In parallel, the unaerated room temperature treatment (UT RT) likely developed spatially heterogeneous redox zones driven by endogenous microbial activity, diffusion limitations, and localized substrate availability. The transformation of aquaculture sludge is inherently complex, involving simultaneous aerobic, facultative, and anaerobic pathways, as well as tightly coupled carbon, nitrogen, and phosphorus cycling (Mirzoyan et al., 2010; Monsees et al., 2017; Eck et al., 2019; Zhang et al., 2021). This intrinsic complexity likely contributed to the relatively higher standard deviations observed among biological triplicates in these treatments, even under nominally identical conditions.

Taken together, aerobic mineralization was most effective for organic matter degradation and nitrate formation, anaerobic digestion achieved the highest nitrogen mineralization, while acidification steps were most effective for phosphorus mobilization. Interestingly, the low-technology storage treatment (UT RT) mobilized both nitrogen and phosphorus to a degree comparable to more complex treatments despite requiring no aeration, heating or chemical addition. This highlights that treatment selection should consider not only maximum nutrient release but also process intensity in terms of time, infrastructure, energy and chemical inputs when considered for large-scale applications.

While nitrogen and phosphorus largely determined overall treatment performance, the behavior of other macro- and micronutrients followed more element-specific patterns.

Element-specific behavior of secondary macro- and micronutrients

Accordingly, the mobilization of secondary macro- and micronutrients was primarily controlled by element-specific interactions such as mineral association, pH-dependent dissolution, and redox-driven transformations (Eck et al., 2019; Zhang et al., 2023). Among macronutrients, potassium showed little variation between treatments, which is consistent with its largely dissolved initial fraction. Similar behavior has been reported for aquaculture sludge, where potassium is primarily present in the liquid phase and only weakly affected by mineralization processes (Monsees et al., 2017; Delaide et al., 2018).

Calcium mobilization remained low across most treatments and increased slightly after acidification treatments (pH, AD pH, AD US pH), suggesting that a substantial fraction of calcium was present in stable mineral phases. Calcium-rich minerals are known to stabilize particulate phosphate in aquaculture sludge through calcium–phosphate precipitation (Cerozi and Fitzsimmons, 2017; Yogev et al., 2020). As a result, phosphorus may remain largely bound in the solid phase despite sufficient total concentrations, potentially limiting its immediate availability in the liquid fraction. Magnesium showed somewhat higher mobilization under acidified conditions (UT RT, pH, AD pH, AD US pH), which is consistent with the increased solubility of magnesium-containing minerals at lower pH and with observations reported for anaerobic sludge treatment systems (Goddek et al., 2018). However, magnesium can also contribute to secondary mineral formation (e.g. struvite), indicating that increased mobilization does not necessarily translate into sustained dissolved nutrient availability (Le Corre et al., 2009).

The behavior of sulfur differed from that of the other macronutrients. Dissolved sulfur increased under aerobic mineralization, whereas strongly reduced concentrations occurred in UT RT and AD treatments. Such decreases are consistent with reductive sulfur transformations during anaerobic degradation, where sulfate may be converted to sulfide and subsequently precipitate with metal ions such as Fe² + (Eck et al., 2019).

Micronutrient mobilization remained generally low. Low or negative iron mineralization across most treatments is consistent with literature, as iron readily precipitates as hydroxides under oxic conditions or as sulfides under anoxic conditions, limiting sustained solubility (Panana et al., 2021; Harabiš et al., 2025). However, the pronounced iron mobilization observed under UT RT cannot be conclusively explained based on the available data. Reductive dissolution may have contributed, although comparable behavior was not observed in AD-based treatments. Without iron speciation or redox measurements, mechanistic interpretation remains uncertain. Manganese, in contrast, showed somewhat higher mobilization in reduced environments such as UT RT, which reflects its redox-sensitive behavior: manganese occurs as soluble Mn² + under reducing conditions but forms insoluble oxides under oxic environments (Harabiš et al., 2025).

Biogas production

The steeper initial biogas production observed in AD US compared to AD S (Supplementary Figure 4) suggests that ultrasonic pre-treatment enhanced early hydrolysis kinetics, likely by increasing substrate accessibility through floc disruption, as previously described for pre-treated sludge digestion (McDermott et al., 2001; Pilli et al., 2011). However, the convergence of both treatments toward comparable final specific yields indicates that the total biodegradable fraction of the sludge was not substantially altered by ultrasonication. Similar yield ranges have been reported for mesophilic digestion of RAS-derived fish sludge by Klein et al. (2024), who documented 357 NL kg − ¹ VS for batch mono-digestion of aquaculture sludge and between 112–515 NL kg − ¹ VS for a continuously fed UASB reactor. Optimization of gas yields were achieved by co-digestion with food waste (540–740 NL kg − ¹ VS) (Netshivhumbe et al., 2024) or ultrasonication as pretreatment (55% increase in specific biogas yield compared to untreated aquaculture sludge) (McDermott et al., 2001). A recent systematic review confirms that fish waste is a viable biogas substrate whose yields are frequently improved by co-digestion, while also noting considerable variability between studies and the need for standardized reporting (Shandu et al., 2025).

Beyond the yields obtained here, the recovery of biogas from aquaculture sludge carries broader implications for the sustainability of the sector (Syropoulou et al., 2024). Anaerobic digestion simultaneously reduces sludge mass and recovers both energy and a nutrient-rich digestate, converting a costly waste stream into two usable products in a single step (Mirzoyan et al., 2010). The resulting biogas can be used on-site to offset part of the substantial energy demand of RAS, in particular for water heating, thereby lowering both operating costs and the fossil-energy footprint of production; for this reason warm-water RAS in Germany are commonly coupled to biogas plants, where combined heat-and-power generation closes both the energy and material cycles of the farm, although efficient utilization depends on adequate solids separation and thickening to concentrate the substrate (Klein et al., 2024). A conceptual mass-balance model has further suggested that on-site digestion could render an integrated aquaponic system largely energy self-sufficient (Yogev et al., 2016). Taken together, the dual recovery of energy and nutrients positions anaerobic digestion as a central component of a more circular and sustainable aquaculture industry.

Suitability of sludge-derived supernatants for further bioproduction

Comparison with standard hydroponic nutrient solutions

For comparison, the nutrient composition of the supernatants was evaluated against a standard hydroponic “Hoagland” solution as documented by Maleki et al. (2021). The comparison (Figures 3, 5) confirms a pattern widely reported in aquaponic literature: sludge-derived nutrient streams do not replicate balanced hydroponic formulations but exhibit characteristic imbalances in macro- and micronutrients (Delaide et al., 2016; Goddek et al., 2016b; Ru et al., 2017; Yang and Kim, 2020; Aslanidou et al., 2023). In conventional hydroponics, nitrogen is predominantly supplied as nitrate, while ammonium is kept low due to potential toxicity and rhizosphere acidification. Depending on crop and system, recommended NH 4 –N concentrations generally remain below ~10–20 mg L − ¹ and typically represent <10–15% of total nitrogen (Resh, 2022). Several aquaponic studies have identified suboptimal NH 4:NO 3 ratios as a key constraint for plant performance, emphasizing that excessive ammonium can impair root function and nutrient uptake (Eck et al., 2019; Delaide et al., 2021; Tetreault et al., 2021; Zhu et al., 2023).

To enable consistent comparison, a hypothetical ammonium threshold of 15 mg L − ¹ NH 4 –N was applied to all supernatants, allowing a direct and operationally meaningful comparison. This represents a high, still tolerable concentration for fruiting vegetables (Akl et al., 2003; Resh, 2022). Because dilution proportionally reduces all nutrients, the dilution factor required to meet this ammonium ceiling simultaneously defines the effective nutrient supply. AT RT could be applied without dilution and provided nitrate-dominant nitrogen in a hydroponically relevant range, along with moderate concentrations in phosphorus and potassium. Similar nitrate-oriented sludge supernatants have been shown to support hydroponic growth when integrated in decoupled systems (Goddek et al., 2016a, 2016b; Aslanidou et al., 2023). pH and US treatments would require moderate dilution while retaining relevant phosphorus concentrations. Diluted pH would still supply ~52 mg L − ¹ PO 4 –P and US ~38 mg L − ¹, with potassium and calcium remaining within supplementation-relevant ranges. These streams may therefore function as phosphorus-enriched concentrates, directly addressing the widely reported phosphorus limitation in aquaponics (Cerozi and Fitzsimmons, 2017; Lobanov et al., 2021; Aslanidou et al., 2023).

In contrast, UT RT and AD-based treatments would require 22–32× dilution. After such dilution, PO 4 –P would decrease to ~1–6 mg L − ¹ and potassium to ~8–12 mg L − ¹, resulting in comparatively dilute nutrient concentrations. Although highly effective in ammonium and partial phosphorus mobilization, these streams are less suitable for direct application unless modified by nitrification or staged blending, as similarly discussed for AD-derived streams in decoupled aquaponics (Goddek et al., 2016b; Zhang et al., 2021).

The same dilution effect would apply to micronutrients. Treatments requiring little or moderate dilution (pH, US, AT RT) retain meaningful trace-element contributions, whereas AD-based treatments and UT RT lose most micronutrient value. After ~3-fold dilution, the pH treatment still supplies manganese (~0.6 mg L − ¹) and zinc (~0.08 mg L − ¹) in ranges approaching hydroponic targets. US (≈2-fold dilution) contributes moderate levels of zinc and copper, while AT RT can be applied undiluted and provides small but relevant amounts of boron, zinc, copper and molybdenum. In contrast, dissolved iron remains far below typical hydroponic requirements (~2.5 mg L − ¹) in all treatments, necessitating supplementation as repeatedly documented (Kasozi et al., 2019; Stathopoulou et al., 2021; Farooq et al., 2023).

Performance in cultivation of Arthrospira platensis

In the present study, the highest growth in A. platensis in terms of dry weight increase and biomass productivity was obtained with SAG media and the AD S media (anaerobic digestion of the sludge) while growth parameters in A. platensis in media derived from pH mobilization (AD pH and AD US pH) showed a tendency towards stagnation at 14 days of cultivation and reached significantly lower final dry weight at day 21 (Figure 5). This growth pattern is also reflected by an increase in Chl.concentration which increased in SAG and AD S throughout the entire 21 days of cultivation, while plateauing (AD US pH) or even declined in A. platensis grown in AD pH (Figure 6). Final dry weight reached in the present study in A. platensis grown in SAG and AD S media compares favorably to final dry weights previously reported for this species grown in suitable media; final dry weight of 1.1 g L -1 in A. platensis grown in MGH−1 fortified with micronutrients was obtained after 19 days (Ghanbarzadeh et al., 2022), in contrast in the present study a dry weight of up to 1.3 g L -1 was obtained in A. platensis grown in the AD S after 21 days (Figure 5). These differences may be based on different nutrient compositions in the applied media. In comparison, Chl concentrations in the present study were somewhat lower than previously reported in A. platensis reaching approximately 20 mg L -1 in the SAG and AD S media; previous reports showed Chl of 40 mg L -1 after 5 days (Chini Zittelli et al., 2022). Such differences may be related to differences in illumination between studies. Overall, especially AD S appears as a suitable pre-treatment if aquaculture sludge is to be utilized for microalgae cultivation. The fact that a pH of 8 for AD S media was applied, suggests even scope for higher growth and productivity when AD S media will be adjusted to higher pH; since it is commonly known that pH above 9 significantly improves productivity in A. platensis (Ismaiel et al., 2016).

In addition to higher performance, A. platensis grown in SAG and AD S also showed significantly higher C-PC concentrations compared to all C-PC concentrations obtained in the other media (Supplementary Figure 5). The effect of the cultivation media on C-PC content in the biomass is crucial as C-PC is among the most valuable biomolecules in A. platensis as it can be valorized for multiple applications from colorant in the food and feed industry, and antioxidant in nutraceuticals and in the pharma industry. Present C-PC concentrations remain about 2- fold lower compared to previous values obtained with similar media (SAG, 49.01 mg gDW -1) and those reported for media containing NH 4 (53.31 – 58.54 mg gDW -1) (Beyer et al., 2023). Even higher values of C-PC contents of up to 353.0 mg g −1 are reported in A. platensis (Hsieh-Lo et al., 2019). C-PC production is known to be affected by media composition and light conditions. For example, enhanced C-PC content was found in media containing moderate levels of NH 4. One explanation for the lower C-PC in the present study may be related to the lack of traces of NH 3 in the present media. At present pH in all media, NH 4 remains stable and does not induce NH 3 conversion. NH 3 is generally considered to be toxic at higher concentrations but may function as an additional nitrogen source and allow the storage of nitrogen in biliproteins as C-PC (Beyer et al., 2023). Apart from media effects, light is a key factor for C-PC production; C-PC content in a range from 121.0–167.9 mg g −1 are reported under red light (PPFD: 100 μmol photons m −2 s −1) and up to 353.0 mg g −1 under blue light (PPFD: 75 μmol photons m −2 s −1) (Hsieh-Lo et al., 2019). Positive effects of blue light on C-PC content in A. platensis have also been reported elsewhere (Chini Zittelli et al., 2022).

Higher performance in A. platensis in media derived from microbial mobilization compared to performance in media derived from pH mobilization cannot be explained by differences in macronutrient composition, because all media were supplemented with 10 g NaHCO 3, as well as KNO 3, and K 2 HPO 4 to achieve total nitrogen (TN) and total phosphorus (TP) concentrations equivalent to those of the standard SAG medium. In addition, differences in performance were not related to pH, since all media were adjusted to pH 8.0 ± 0.1. Among the analyzed potentially toxic elements (Table 4), zinc appears to be the only element plausibly linked to growth inhibition. Zinc levels exceeding 0.05 mg L − ¹ were documented to cause inhibitory effects in previous studies on A. platensis (Choudhary et al., 2007). The measured concentration in the pH, AD pH and AD US pH media (0.25 – 0.38 mg L − ¹) may have affected the growth of Arthrospira platensis negatively.

Beyond potential trace-element effects, physicochemical medium properties may also have influenced growth. Electrical conductivity (EC) in growth media for freshwater microalgae is a critical parameter that dictates nutrient availability, osmotic stress, and, consequently, growth rates. While EC is commonly used as a proxy for nutrient concentration, it primarily measures the total ion concentration (salinity) of the medium (Zafar et al., 2022). Physicochemical properties of initial sludge and supernatants after specific treatments indicate a considerable difference in EC between AD S, AD pH and AD US pH (Table 3), which also persists in the subsequent adapted media. Present EC values are substantially below EC values commonly present in scientific media such as SAG 17–20 mS cm -1. All media generated from different mobilization techniques showed a dark color. Since no measurements of media color were performed, potential differences in effects of treatment and resulting growth (by reducing light efficiency) cannot be ruled out.

Mobilization of potentially toxic elements and health/safety considerations

EU fertilizing-product contaminant thresholds (e.g., Cd, Pb, Ni, Hg, Cr) under Regulation (EU) 2019/1009 are defined on a dry-matter basis. For the sludge-derived supernatants analyzed here, comparison with dry-matter-based limits is not directly applicable. Therefore, water-quality-based benchmarks provide a more suitable framework for evaluation. When compared to FAO irrigation water guideline values, all measured concentrations were well below recommended maxima (Table 4) (Ayers and Westcot, 1985).

Cultivation-specific toxicity thresholds provide additional context. Most of those elements analyzed in the present study are about 10-fold lower compared to values showing growth inhibitory effects in A. platensis. For example, concentrations of cobalt and lead remain well below the documented threshold of 0.05 mg L -1 (Rangsayatorn et al., 2002; Choudhary et al., 2007). As well as the thresholds for nickel (1.0 mg L − ¹) and cadmium (1.5 mg L − ¹) reported to result in significant growth inhibition for Arthrospira platensis (Diaconu et al., 2023). Zinc concentration represents a clear exception as discussed in the previous section. Duckweed (Lemna spp.) is also known to be sensitive to dissolved heavy metals under standardized growth-inhibition testing (Drost et al., 2007; Khellaf, 2009). Irfan et al. (2024) report EC 50 values for multiple metals in the sub-mg L − ¹ to low-mg L − ¹ range and note that standardized Lemna tests cited in their study show a median EC 50 of 0.683 mg L − ¹. These concentrations remain well above the µg L − ¹ levels measured in the present study.

Nevertheless, absence of acute toxicity does not preclude bioaccumulation. Microalgae and duckweed can accumulate heavy metals in biomass, with reported uptake capacities reaching several mg g − ¹ dry weight depending on species and exposure conditions (Markou et al., 2018). Repeated reuse in circular systems may promote gradual redistribution and accumulation of contaminants (van der Spiegel et al., 2013). Reviews of Focker et al. (2022) and Pettersen et al. (2025) emphasize that circular food systems can promote transfer and accumulation of persistent chemical hazards, including heavy metals, across plant, animal and aquaculture production chains under long-term reuse scenarios.

This consideration also applies to hydroponic and aquaponic vegetable production systems, where nutrients remain in dissolved form and plant uptake can directly reflect water composition due to the absence of soil buffering. Even when dissolved concentrations remain below FAO irrigation guideline values and below reported acute growth-inhibition thresholds for representative duckweed and cyanobacteria species, such comparisons constitute an initial screening only. Given documented metal accumulation and redistribution risks in circular food systems, definitive conclusions regarding suitability for feed, food or recirculating aquaponic applications require verification of heavy-metal concentrations in harvested biomass under realistic cultivation conditions. Continuous monitoring of both nutrient solutions and produced biomass is therefore essential prior to making safety claims or scaling implementation.

Alternative treatments and considerations for scale-up

One class of treatments not evaluated in the present study is thermal or thermochemical processing, encompassing hydrothermal treatment, incineration and pyrolysis (Zhu et al., 2022). Thermal hydrolysis disrupts cell walls and mineral associations and liberates organically bound phosphorus and nitrogen into the soluble phase, with reported increases in soluble phosphorus of around 50% (Wang et al., 2024). In a related study on the same aquaculture sludge, an autoclaved water extract produced a medium that supported A. platensis growth comparable to a synthetic medium, whereas acid- and alkali-assisted thermochemical extracts inhibited growth despite their higher nutrient release (Elshobary et al., 2026). These routes mobilize nutrients effectively but remain energy-intensive; life-cycle assessment nonetheless indicates they can be environmentally favorable, largely by displacing mineral-fertilizer production (Hosseinian et al., 2024).

The scalability of ultrasonication warrants particular consideration. The probe sonotrode used here represents a laboratory configuration, whereas full-scale applications rely on flow-through tube or double-tube reactors installed upstream of the digester (Xie et al., 2007; Bandelin et al., 2020). Reported full-scale benefits are variable, ranging from substantial biogas gains to only minor and often statistically insignificant increases that barely offset the additional energy and equipment costs (Pilli et al., 2011; Lippert et al., 2021), consistent with the comparable final biogas yields of AD US and AD S observed here. Ultrasonication therefore appears most useful as a pre-treatment to accelerate digestion rather than as a stand-alone nutrient-mobilization step at full scale.

From a techno-economic standpoint, the treatments impose distinct types of cost: sealed storage consumes neither energy nor chemicals, but its long retention time requires substantial holding capacity, so its principal burden is the footprint and infrastructure needed — a constraint that grows with the volume of sludge to be processed. Acidification carries a consumable chemical cost, aerobic mineralization and ultrasonication a recurring energy cost, and anaerobic digestion both a capital cost (sealed, heated reactors) and an operating one. The balance is most favorable for anaerobic digestion, whose biogas can partly offset its own energy demand and thereby improve its cost balance relative to the other routes (Section 4.4). For the other routes such offsets are absent, and although comprehensive techno-economic and life-cycle analyses remain scarce for sludge-based nutrient recovery (Goddek et al., 2015; Zhang et al., 2021), a preliminary assessment of thermal treatment of this same sludge found that the energy required could outweigh the market value of the recovered nitrogen and phosphorus (Elshobary et al., 2026). Ultimately, the viability of any of these treatments at scale will hinge on whether the nutrients — and, for digestion, the energy — they recover generate more value than the treatment itself consumes.

This study demonstrated that aquaculture sludge can be converted into dissolved nutrient streams, but treatment strongly determines nutrient form and usability. Anaerobic digestion achieved the highest nitrogen mineralization (≈45–47%), producing ammonium-rich supernatants (≈460–480 mg NH 4 –N L − ¹), whereas aerobic mineralization promoted nitrification and generated nitrate-dominated solutions (≈257 mg NO 3 –N L − ¹). Phosphorus mobilization was primarily pH-driven and reached its maximum under acidification (≈165 mg PO 4 –P L − ¹), with additional release observed after post-acidification of digested sludge.

Despite effective nutrient mobilization, none of the treatments produced a directly usable hydroponic solution. AD- and storage-derived streams required substantial dilution due to high ammonium concentrations, while iron remained consistently below typical plant requirements across all treatments. Acidified treatments provided elevated phosphorus and manganese concentrations.

Among the tested media, the untreated AD supernatant supported the highest Arthrospira platensis growth and pigment production, indicating that biologically mineralized sludge can serve as a suitable nutrient source after adjustment. At the same time, low concentrations of potentially toxic elements suggest no immediate limitation for cultivation, although long-term accumulation in biomass should be verified.

Overall, the results highlight that sludge mineralization can recover substantial nutrient fractions from solid waste streams, but process selection must be aligned with the intended application, balancing nutrient composition, required post-treatment, and operational complexity.