Influence of H₂O₂ Advanced Oxidation Processes on Phosphorus Recovery from Waste Activated Sludge

2.1. Chemicals

Hydrogen peroxide (30% solution, stabilized with phosphate ≤ 5ppm, CAS number: 7722-84-1) was purchased from Carl Roth GmbH + Co. KG, Germany.

2.2. Sample Collection and Preparation

WAS was collected from a wastewater treatment plant in Rostock, Germany. The sample was allowed to settle for 30 min, reducing the volume from 10 L to 4 L by decanting the supernatant. The resulting thickened WWAS was then transported to the laboratory.

Table 1. Characteristics of washed waste activated sludge. Values are given as (mean ± standard deviation) from three replicates.

Dry Residue (DR)	(23.25 ± 0.02) g/kg
Organic Dry Residue (oDR)	(77.74 ± 0.78) % of DR
pH	6.48
Conductivity	90 µS/cm
Total COD	(1230 ± 70) mg/g DR
Total P	(38.88 ± 0.44) mg/g DR
Total N	(75.19 ± 0.13) mg/g DR
Total Fe	(3.53 ± 0.88) mg /g DR

Table 1. Characteristics of washed waste activated sludge. Values are given as (mean ± standard deviation) from three replicates.

Dry Residue (DR)	(23.25 ± 0.02) g/kg
Organic Dry Residue (oDR)	(77.74 ± 0.78) % of DR
pH	6.48
Conductivity	90 µS/cm
Total COD	(1230 ± 70) mg/g DR
Total P	(38.88 ± 0.44) mg/g DR
Total N	(75.19 ± 0.13) mg/g DR
Total Fe	(3.53 ± 0.88) mg /g DR

To remove non-biomass contaminants that could interfere with the H₂O₂ reaction, the thickened WWAS was washed twice via centrifugation at 5,000 RPM for 10 minutes, each time replacing the supernatant with deionized (DI) water. The resulting washed activated sludge (WWAS) was then utilized for subsequent batch tests. Its initial characteristics are summarized in WAS collected from a wastewater treatment plant in Rostock, Germany. The sample was allowed to settle for 30 min, reducing the volume from 10 L to 4 L by decanting the supernatant. The resulting thickened WWAS was then transported to the laboratory.

2.3. Pre-Tests

Pre-tests were conducted to evaluate the effects of various H₂O₂ dosages on sludge disintegration and phosphate release at 20 °C and 60 °C, establishing the optimal dosage for subsequent experiments.

Accordingly, reactors vessels were filled with 400 mL WWAS diluted with DI water to at a concentration of approximately 3.4 g_DR/L. The mixtures were continuously stirred at 300 RPM. For the 60 °C series, the experiments were conducted in a temperature-controlled water bath, injecting H₂O₂ once the sample temperature stabilized. To prevent evaporation while allowing gas escape, reactor openings were partially sealed. After 3 h of reaction, the liquid phase was analyzed for soluble chemical oxygen demand (sCOD), PO₄-P, and pH.

Furthermore, H₂O₂ stability was assessed in DI water (10 g/L) for 3 h; as no significant decomposition was observed at either temperature, thermal degradation was considered negligible.

2.4. Batch Tests

WWAS, equivalent to 0.58 g_DR, was diluted with DI water to a sample volume of 200 ml (2.9 g_DR/L). Four treatment groups were designed to isolate and evaluate the individual and synergistic effects of H₂O₂ at a dosage of 10 g/L (3.45 g_H2O2/g_DR) and thermal stress on sludge disintegration and phosphorus release. Group 1 involved only heating-up the sample from 20 °C to 60 °C within 45 mins without H₂O₂ addition. Groups 2 and 3 replicated the pre-test conditions, with stirred samples maintained at either 20 °C or 60 °C for 3 h, with or without H₂O₂. In Group 4, the thermal pretreated sample (Group 1) was cooling back to 20 °C before the 3-hour reaction period (with or without H₂O₂). The experimental matrix for all treatment groups is summarized in Table 2.

2.5. Analysis

After the batch tests, the dissolved fraction of the samples was analyzed for pH, solubilized COD, PO₄-P, Fe^2+/3+, Ca²⁺, Mg²⁺, NH₄–N, NO₃–N and residual H₂O₂ using HACH test kits (LCK 1414, LCK 350, LCK 320, LCK 327, LCK 327, LCK 305, LCK 339 and HYP-1 (2291700), respectively). The sCOD values were corrected for residual H₂O₂ interference according to (Kang et al. 1999).

For the measurement of total COD, N and Fe with same test kits, the samples were homogenized first using a disperser (T25 basic ULTRA-TURRAX^®, IKA^®-WERKE, Germany) at 11,000 RPM for 5 min. Standard deviation war calculated from triplicate measurements of the WAS.

PP = TP − DP

(1)

DRNP = DP − DRP

(2)

P was fractionated according to procedures illustrated in Figure 1 and subsequently determined using HACH test kits (LCK 350). Total P (TP) concentration corresponds to the value measured after aforementioned sample homogenization and subsequent chemical digestion. Dissolved P is defined as the fraction passing through a 1-2 µm glass fibre filter and is measured following chemical digestion. Dissolved reactive P (DRP, analyzed as PO₄-P) represents a subfraction of DP and is determined directly without prior chemical digestion. The non-directly measured fractions, namely dissolved non-reactive P (DNRP) and particulate P (PP), were calculated according to Equations 1 and 2.

All parameters were measured in triplicates; mean and standard deviation are provided. Availability and mobilization of selected parameters were calculated using equation 3 and 4, respectively, with C_S(end) and C_S(initial) being concentrations in solution after and before treatment and C_total as the total concentration.

$$\mathit{Availability}={\displaystyle \frac{C_{S\left(\mathit{end}\right)}}{C_{\mathit{total}}}}$$

(3)

$$\mathit{Mobilization}={\displaystyle \frac{C_{S\left(\mathit{end}\right)}-C_{S\left(\mathit{initial}\right)}}{C_{\mathit{total}}-C_{S\left(\mathit{initial}\right)}}}$$

(4)

3.1. Pre-Test

The results of the pre-tests are shown in Figure 2. H₂O₂ was applied up to a maximum dosage of 3.0 g, corresponding to a concentration of 7.5 g/L or a specific load of 2.2 g_H2O2/g_DR.

COD solubilization, indicating the extent of sludge disintegration, increased with H₂O₂ dosage at both temperatures, albeit at different orders of magnitude. At 20 °C, the initial sCOD of 14 mg/L rose to 98 mg/L at the highest dosage of 3.0 g H₂O₂. In comparison, thermal treatment alone at 60 °C resulted in an sCOD of 567 mg/L, which further escalated to 2687 mg/L upon the addition of 3.0 g H₂O₂.

At 20 °C, the P release mirrored the trend of sCOD solubilization, increasing from below the detection limit (0.05 mg/L) without H₂O₂ to 32 mg/L at the highest dosage, albeit with a diminishing rate of increase. In contrast, the P release at 60 °C exhibited the opposite trend: thermal treatment without H₂O₂ yielded the highest concentration of PO₄-P. Despite the pronounced effect of H₂O₂ on sCOD solubilization, the PO₄-P concentration stagnated or marginally declined with increasing H₂O₂ dosages.

The H₂O₂ dosage also exerted a distinct effect on the pH value. At 20 °C, the pH slightly increased from 6.2 (without H₂O₂) to 6.6 at the highest dosage. Thermal treatment alone at 60 °C resulted in a marginal pH increase to 6.4; however, the subsequent addition of H₂O₂ led to a continuous decline, reaching a minimum of 4.7 at the maximum dosage.

These results, in particular the declining trend of PO₄-P release with increased H₂O₂ dosage at 60 °C, are well aligned with previous studies, why a more comprehensive analysis could be conducted subsequently.

3.3. Batch Tests

The raw analytical results after all treatments are provided in Table 3. The data of the untreated WWAS before the treatments is shown for comparison (“Heating-up, before treatment”).

Figure 3 summarizes the effect of the four different treatments on the Availability of PO₄-P, DP, sCOD, NH₄-N, NO₃-N, Fe³⁺ and Fe²⁺ with regard to the total mass of P, COD, N and Fe, respectively.

Figure 3. Effect of the different treatments on the availability of various parameters.

Figure 3. Effect of the different treatments on the availability of various parameters.

Figure 4. Injected and residual H₂O₂ concentration after treatments.

Figure 4. Injected and residual H₂O₂ concentration after treatments.

3.3.1. Cod and Residual H₂o₂

The trends observed in the pre-test regarding COD solubilization could be replicated. Thermal treatment alone increased the available sCOD by 17.4% (Figure 3c), whereas H₂O₂ treatment at 20 °C resulted in only 2.9% solubilization (Figure 3b). Combining H₂O₂ addition with an elevated treatment temperature resulted in an elevated 32.8% COD solubilization.

Additional COD solubilization was similarly low, independently if H₂O₂ was added to the raw (Figure 3b) or the thermally pretreated sample at 20 °C (Figure 3d).

Although H₂O₂ did not appear to react effectively at room temperature, no residual H₂O₂ was detected in the sample. In contrast, the treatment that achieved the highest degree of solubilization at 60 °C contained only 1.25 g/L less H₂O₂ than the initial concentration. This difference is attributed to the temperature-dependent activity of catalase, which decomposes H₂O₂ into water and molecular oxygen and loses activity at temperatures above approximately 40 °C (Wang et al. 2009; Gabbita, Kasi V and Hzuang, Jerry YC 1984). In agreement with this explanation, (Wang et al. 2009) observed a rapid decrease in H₂O₂ concentration in WAS at room temperature, whereas H₂O₂ degradation was considerably lower at 60 °C. Thus, the complete depletion of H₂O₂ after 3 h at room temperature observed in the present study appears plausible.

H₂O₂ injected to a thermally pretreated sample at 20 °C was still measured after 3 hours (2.9 g/L). Apparently, the heating-up procedure inactivated parts of the catalase irreversibly, why its degradation activity on H₂O₂ became limited. Nevertheless, an increased solubilization by H₂O₂ was not observed, indicating the missing thermal activation.

3.3.2. Ammonium and Nitrate

The treatments resulted in distinct differences on mobilization and fate of ammonium (Figure 5). The heating-up procedure alone mobilized 5.7% of the total N as ammonium, which probably is directly related to thermally-induced COD solubilization (Figure 3a). When H₂O₂ is added to the thermal treatment, no ammonium but 113 mg/L NO₃-N are detected corresponding to an apparent mobilization of 54% of total N and subsequent oxidation.

Several parallel or consecutive processes may be involved in this case. By disintegrating microbial cells, ammonium and intercellular organic N-compounds become solubilized. H₂O₂ highly likely cleaved amino groups from these amino compounds by oxidative deamination, hydrolysis or broke the amino compound itself into keto acid and free ammonia (Targhan et al. 2021). Also, nitrate was found to be a direct reaction product from AOPs of organic amides (Leitner et al. 2002).

The released NH₄⁺ was seemingly completely oxidized. In clear water, H₂O₂ is to weak for oxidizing ammonia into nitrate. Therefore, Fe ions present in the WAS were highly likely crucial catalysts to initiate Fenton reactions and produce ∙OH. But even then, the reaction kinetics of ∙OH with NH₄⁺ are inherently slow in acidic media (Siddiqui et al. 2013). The observed conversion to nitrate in this study is therefore be due to the specific operational conditions. The combination of an elevated temperature, the extended reaction time of three hours and a large stoichiometric excess of H₂O₂ successfully probably overcame the kinetic limitations, driving the multi-step radical oxidation to its thermodynamic end point (NO₃^-) (Equation 5).

$$\mathrm{N}{\mathrm{H}}_4^{+}+4{\mathrm{H}}_2{\mathrm{O}}_2\to \mathrm{N}{\mathrm{O}}_3^{-}+{5\mathrm{H}}_2\mathrm{O}+2{\mathrm{H}}^{+}$$

(5)

This assumption is also supported by the reducing pH. The oxidation of one mole ammonium with H₂O₂ produces two mole H⁺. Since 113 mg/L NO₃-N correspond to 8 mmol/L, about 16 mmol/L H⁺ must have been produced, which would correspond to an endpoint pH of 2.0 in an unbuffered system. Actually, in this experiment, the pH dropped from 6.5 to 4.4, indicating the presence of buffer systems. This is probably mainly due functional groups of the WAS and its solubilization products, in particular protonation of particular carboxylic groups (COO^- → COOH). The pKa value of many carboxylic acids is in the range between 4 and 5 (Namazian and Halvani 2006) supporting this assumption.

With regard to the intended improvement of anaerobic degradability by oxidative COD solubilisation, the oxidation of ammonia to nitrate is undesirable. The reduction of 1 g nitrate to N₂ requires stoichiometrically 2.86 g COD (Daigger 2014) which is lost for methane production.

No ammonium was detected in the experiment with H₂O₂ dosage at 20 °C after thermal pretreatment either, while nitrate increased to 13 mg/l (6.5% of total N). Considering the remarkably high 53.1% COD solubilization in parallel indicates that most N stayed organically bound N in this case (and could not be oxidized).

3.3.3. Phosphorous

The mobilization of DP and its sub-fraction PO₄-P is compiled in Figure 6. H₂O₂ dosage at room temperature resulted in a slight P-mobilization of 10.5%, of which nearly 90% was dissolved as PO₄-P. This significant increase is in contrast to a very low solubilization of sCOD (2.9%). In this case, it may be due to the specific characteristics of WAS as it comes from a wastewater treatment plant performing enhanced biologic phosphorous removal. Normally, phosphate accumulating organisms (PAO) release PO₄-P under anaerobic conditions and fill their polyphosphate storage under aerobic conditions (Mino et al. 1998). As hypothesis, the extreme oxidative stress and the above discussed probable catalase production/release may have triggered the activation of the phosphate driven energy metabolism with an according P release.

Thermal treatment alone mobilized 26.4% of the total P, of which more than 80 % were released as readily available PO₄-P. This is the highest value for all treatments. Again, the presence of PAOs in the WAS may partly explain this strong release, but here probably initiated by heating stress and oxygen limitation. Additionally, disintegration and improved chemical dissolution of P-salts are probable sources of the increased DRP.

The additional treatment with H₂O₂ at 60 °C resulted in an additional significant increase of DP (40% of P). However, the mobilized PO₄-P is reduced compared to thermal treatment alone (12.2%). That means the strong increase of DP is related to dissolved organic P-fractions. Regarding the high solubilization of COD, the increase of DP is likely caused by cell disintegration. It is probable, that solubilization of organic matter is incomplete and cell internal P remains organically bound. However, this does not explain the reduced availability as PO₄-P compared to thermal treatment alone.

Applying H₂O₂ to the thermally pretreated sample at 20 °C did only insignificantly change DP and PO₄-P compared to thermal treatment alone. That means, in these experiments, no reduction of PO₄-P was observed. This is in line with the COD solubilization. It can be assumed, that the P mobilization was already accomplished by thermal treatment and the low reactivity of H₂O₂ at 20 °C could not access new P sources.

However, the question remains, why originally mobilized PO₄-P at 60 °C is reduced by H₂O₂ dosage. Since biologic uptake can be excluded in these conditions, it suggests that phosphates are somehow chemically bound. Potentially relevant cations for precipitation reactions are Fe³⁺, Ca²⁺ and Mg²⁺.

3.3.4. Ca and Mg and Fe

In the experiment with combined thermal and H₂O₂ treatment at 60 °C, pH dropped to 4.41, mainly due to the oxidation of ammonia to nitrate. This leads also to a nearly complete protonation of orthophosphate (PO₄^3-) to mono-hydrogen phosphate (H₂PO₄⁻). Stoichiometrically, about 99% of the phosphates are available in this form at pH 4.4. This is another relevant buffer system, explaining the dampened pH drop with regard to the stoichiometric H⁺ production by ammonia oxidation. The solubility of Calcium and Magnesium mono-hydrogen Phosphates is in range of several Gramm per Liter. Accordingly, the precipitation with Ca²⁺ and Mg²⁺ ions can be excluded.

The WWAS contained 3.5 mg Fe/g DR even after washing. This is due to iron dosage at the wastewater inflow for sulfide precipitation to mitigate H₂S problems. Accordingly, it can be assumed that a large ratio of iron is bound as FeS. However, since the WAS is aerated at the activated sludge system, some transformation to three valent ion is probable, which itself could already react with phosphate to FePO₄. As shown in Figure 7, all treatments resulted in some mobilization of Fe³⁺.

The increase in case of pure thermal treatment might be due to improved solubility of iron compounds in combination with oxidation of Fe²⁺ with eventually available dissolved oxygen after washing. The strong difference with additional H₂O₂ treatment is probable due to oxidation of Fe²⁺ to Fe³⁺. This effect has been already observed by (Ismail et al. 2022) for WAS containing iron (II) sulfide. They found that H₂O₂ dosage could dissociate FeS, thereby releasing Fe²⁺, which then quickly is further oxidized by H₂O₂ to Fe³⁺ (Fenton reaction).

$${\mathrm{Fe}}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2\to {\mathrm{Fe}}^{3+}+\dot{\mathrm{OH}}+{\mathrm{OH}}^{-}$$

They measured a complete dissociation and oxidization of FeS to Fe³⁺ within 1 min at room temperature, when the molar H₂O₂ dosage was at least two times of FeS content. The ferric iron then undergoes quick hydrolysis in water forming hydrous ferric oxides (HFO), mainly Fe(OH)₃. Available PO₄^3- is then precipitated by the HFO forming ferric hydroxyl phosphates with very low solubility.

$${2\mathrm{Fe}}^{3+}+{2\mathrm{OH}}^{-}+{\mathrm{H}}_2\mathrm{O}\to \mathrm{Fe}(\mathrm{OH}{)}_3\downarrow +{\mathrm{H}}^{+}$$

$$\mathrm{Fe}{\left(\mathrm{OH}\right)}_3+{\mathrm{PO}}_4^{3-}+{3\mathrm{H}}^{+}⟶{\mathrm{FePO}}_4\downarrow +3{\mathrm{H}}_2\mathrm{O}$$

Ismail et al. experimentally verified in their experiments a nearly ideal stoichiometric ratio of one mole of PO₄^3- removed per mole of released Fe³⁺, or a ratio of 0.55 g P/g Fe³⁺ on mass basis. Also other studies show that precipitation of phosphate stops when H₂O₂ dosage exceeds the dosage of Fe²⁺ (Li et al. 2009; Ghodsi et al. 2021).

However, the total iron mass in the sample (2 mg) is clearly a limiting factor and could explain only precipitation of 1.1 mg PO₄-P if a molar ratio of 1:1 is applied.

Still, over proportional binding of phosphorus is possible if iron humic phosphate compounds (Fe-HS-P) are considered. The underlying processes and formed products are well investigated in soil science and plant nutrition context (Gerke 2021) and can be shortly characterized as follows. Fe³⁺ possesses a high affinity to carboxylic acids (-COOH) and phenolic hydroxyls (-OH). These functional groups wrap around the Fe³⁺ iron and forming iron-humic chelates. These provide multiple vacant water ligand spots and can bind more H₂PO^- or HPO₄^2- ions than Fe³⁺ directly. The resulting product is an amorphous, rather colloidal compound. The treatment at 60 °C with H₂O₂ in combination with the assumed protonation of carboxylic groups are exactly fostering conditions for the formation of Fe-HS-P. Also, the strong increase of the organic dissolved P-fraction supports this probable P-fixation pathway.

In the last years Fe-HS-P became also relevant in research on P recovery in sludge treatment, since the ionic and chelatic iron bridges limit the efficiency of standard chemical extraction and subsequent precipitation (Wilfert et al. 2015; Lu et al. 2022). Various approaches have been developed to overcome this challenge, either aiming at destabilizing the Fe-HS-P complex (Wilfert 2018) or direct concentration/drying of the amorphous flocs (Wang et al. 2022).

Summarizing, the apparent reduced PO₄-P mobilization of the thermally-assisted H₂O₂ AOP treatment compared to the sole thermal treatment is probably due to a chain of reactions. COD solubilization produces organic acids. In parallel, NH₄⁺ is oxidized by hydroxylic radicals. The related production of protons yields in a protonation of phosphate to H₂PO₄⁻ and organic functional groups, in particular carboxylic groups of the produced organic acids. The parallel oxidation of Fe²⁺ to Fe³⁺ provides the basis for the formation of ternary iron-humic-phosphate complexes measured here as strong increase of the dissolved organic P-fraction.