Effect of Pharmaceutical Sludge Pre-Treatment with Fenton/Fenton-like Reagents on Toxicity and Anaerobic Digestion Efficiency

Sewage sludge is successfully used in anaerobic digestion (AD). Although AD is a well-known, universal and widely recognized technology, there are factors that limit its widespread use, such as the presence of substances that are resistant to biodegradation, inhibit the fermentation process or are toxic to anaerobic microorganisms. Sewage sludge generated by the pharmaceutical sector is one such substance. Pharmaceutical sewage sludge (PSS) is characterized by high concentrations of biocides, including antibiotics and other compounds that have a negative effect on the anaerobic environment. The aim of the present research was to determine the feasibility of applying Advanced Oxidation Processes (AOP) harnessing Fenton’s (Fe2+/H2O2) and Fenton-like (Fe3+/H2O2) reaction to PSS pre-treatment prior to AD. The method was analyzed in terms of its impact on limiting PSS toxicity and improving methane fermentation. The use of AOP led to a significant reduction of PSS toxicity from 53.3 ± 5.1% to 35.7 ± 3.2%, which had a direct impact on the taxonomic structure of anaerobic bacteria, and thus influenced biogas production efficiency and methane content. Correlations were found between PSS toxicity and the presence of Archaea and biogas yields in the Fe2+/H2O2 group. CH4 production ranged from 363.2 ± 11.9 cm3 CH4/g VS in the control PSS to approximately 450 cm3/g VS. This was 445.7 ± 21.6 cm3 CH4/g VS (1.5 g Fe2+/dm3 and 6.0 g H2O2/dm3) and 453.6 ± 22.4 cm3 CH4/g VS (2.0 g Fe2+/dm3 and 8.0 g H2O2/dm3). The differences between these variants were not statistically significant. Therefore, due to the economical use of chemical reagents, the optimal tested dose was 1.5 g Fe2+/6.0 g H2O2. The use of a Fenton-like reagent (Fe3+/H2O2) resulted in lower AD efficiency (max. 393.7 ± 12.1 cm3 CH4/g VS), and no strong linear relationships between the analyzed variables were found. It is, therefore, a more difficult method to estimate the final effects. Research has proven that AOP can be used to improve the efficiency of AD of PSS.


Introduction
Sludge is widely considered to be one of the organic feedstocks processable using anaerobic digestion (AD) [1]. Both research works and full-scale installations have shown that well-maintained AD can be applied to produce CH 4 with high efficiency and partially stabilize sludge via removal of putrescible organics, deodorization, improvement of fertilizing properties, and partial sanitization [2,3]. Although AD is well-known, universal and widely recognized as effective, its widespread deployment has been hamstrung by multiple factors [4,5]. One such barrier is the presence of substances that are resistant to biodegradation, inhibit the fermentation process, or are toxic to anaerobic microorganisms [6,7]. Sewage sludge generated by the pharmaceutical sector and other industries is one such The aim of this study was to assess the applicability of AOPs via Fenton/Fentonlike reactions in the pre-treatment of pharmaceutical sewage sludge (PSS) intended for anaerobic digestion. It tested the extent to which the method reduced PSS toxicity and improved AD performance (biogas yields and methane fractions). The results were used to develop optimization methods to estimate AOP methane production as a function of chemical reagent doses and sludge parameters.

Experimental Design
The Surplus sewage sludge from a pharmaceutical sewage treatment plant (PSS) was used for the study. The experiment was divided into two phases. In phase 1 (P1), the PSS was pre-treated through the advanced oxidation process (AOP). During phase 2 (P2), the PSS was subjected to anaerobic digestion (AD). Phase 1 was executed in two series with different levels of AOP chemical reagents. The Fenton reagent (Fe 2+ /H 2 O 2 ) was tested in series 1 (S1), and a Fenton-like regent (Fe 3+ /H 2 O 2 ) in series 2 (S2). A flowchart of the process is given in Figure 1. Each series was divided into six variants (V) with different doses of chemical reagents applied to the PSS. The experimental design is presented in Table 1. Phase 2 (P2)each variant of AOP of PSS was tested with regard to AD performance by means of respirometric measurements in semi-batch reactors. P2 was divided into the same series (S1, S2) and variants (V1-V5) as P1.

Pharmaceutical Wewage Sludge (PSS) and Anaerobic Sludge (AS)
Gravity-thickened PSS was extracted from the secondary clarifier of an aerobic pharmaceutical wastewater treatment plant (P-WWTP) running a conventional activated sludge process without enhanced nutrient removal. Wastewater directed to P-WWTP was from a pharmaceutical company that produces 42 active substances, including baclofen, aripiprazole, tadalafil, vardenafil, alendronate, risedronate, sildenafil, hydrochlorothiazide, xylometazoline, sildenafil, piracetam, pentoxifylline, metronidazole, hydrochlorothiazide, and ethopyrine. The sewage flow rate averaged 2000 m 3 /d. The P-WWTP generates approx. 30 tonnes of PSS/day. The anaerobic sludge (AS) inoculum was sourced from the enclosed digesters of the Municipal Water Treatment Plant in Olsztyn (Poland). The digester operating parameters are organic load rate (OLR) approx. 2.4 kg VS/m 3 ·d, hydraulic retention time (HRT) 20 days, and temperature 35ºC. Prior to the experiment, the AS was conditioned and adapted for PSS anaerobic digestion for 60 days (three times hydraulic residence time in the reactor). The characteristics of the PSS and AS used in the study are presented in Table 2. At the beginning of the P1 experimental cycle, 200 cm 3 PSS was fed into the reactor, after which the chemical reagents were introduced. The PSS was first amended with the target Fe dose, then, after 10 minutes, with the H 2 O 2 at a constant Fe/H 2 O 2 ratio of 1:4 by weight. The sludge was agitated for 20 min at 150 rpm with vertical, 3-blade, mechanical mixers (Nanostar 7.5 Digital, IKA, Poland) to ensure uniform reagent distribution, then at 50 rpm to allow the PSS to react thoroughly with the reagent. The sludge retention time in the reactor was 24 h. The scheme of the experiment organization in this phase is shown in Figure 2. In phase 2, the PSS was subjected to AD.

Phase 2-Anaerobic Reactors
AD performance was tested using the AMPTS II kit (Bioprocess Control, Lund, Sweden). The semi-batch process was run in reactors with a total volume of 2.0 dm 3 fitted with vertical, 3-blade mechanical mixers rotating at 100 rpm (3 min ON/10 min OFF regime). Prior to AD, the reactors were inoculated with 1.0 dm 3 anaerobic sludge. The organic load rate was 2.5 g VS /dm 3 d; the hydraulic retention time was 20 d. Digested sludge was removed once a day, and the reactor was replenished with an equivalent amount of feedstock. In order to ensure anaerobic conditions at the start of the experiment, the inoculum + PSS mixture was purged with pure nitrogen for 3 min. The temperature was kept at a constant 40 • C by placing the reactors in a water bath. Bioreactors were fitted with a nozzle in the carbon dioxide absorption unit. The resultant biogas was fed into a 100 cm 3 tank filled with a 3M solution of NaOH (Pol-Aura Ltd., Olsztyn, Poland).

Analytical Methods
TS, VS and MS were determined gravimetrically. TS levels in the biomass were determined by drying to a constant weight at 105 • C, then burning it at 550 • C (the loss on ignition was the VS, as per PN-EN 15935: 23022-01) [40]. Biomass samples desiccated at 105 • C were assayed for TC, TOC and TN using a Flash 2000 elemental particle analyzer (Thermo Scientific, USA) [41]. The concentrations of dissolved TOC were quantified using a TOC-L (Shimadzu, Kyoto, Japan) [42]. TP was determined colorimetrically in ammonium metavanadate (V) and ammonium molybdate after sample mineralization in a mixture of sulfuric (VI) and chloric (VII) acids at 390 nm, using a DR 2800 spectrophotometer (HACH Lange, Weilheim, Germany) [43]. Total protein was calculated by multiplying the value of TN by the protein conversion factor of 6.25 [44]. Reducing sugars were determined colorimetrically with anthrone reagent at 600 nm using a DR 2800 spectrophotometer (HACH Lange, Weilheim, Germany) [45]. Lipids were extracted using the Soxhlet method with a Buchi extraction apparatus (Flawil, Switzerland) and then determined by weight difference afterward [46]. The pH value of H 2 O was determined potentiometrically with an 867 pH Module (Metrohm, Herisau, Switzerland) [47]. The FOS/TAC (the ratio of the buffer capacity of the sample to the VFA levels in the sample) was determined using a TitraLab AT1000 titrator (HACH Lange, Weilheim, Germany) [48]. Unreacted H 2 O 2 was quantified iodometrically and with the use of Quantofix Peroxide strips (Macherey-Nagel, Düren, Germany) (range: 1-100 and 50-1000 mg/dm 3 ) [49]. Acute toxicity of the sludge (aqueous extracts) was measured using Vibrio fischeri bacteria in an M 500 Analyzer (Azur Environmental, Delaware, USA), acc. to PN-ISO 11348-2:2008 [50]. The aqueous extract was prepared by adding four volumes of distilled water on top of one volume of sludge and agitating mechanically for 24 h [51]. The molecular analysis aimed to determine the percentage of AD bacteria in the biomass using the fluorescent in situ hybridization (FISH) technique [52]. Four molecular probes were used for hybridization: a Bacteria-universal probe EUB338 [53], an Archaea-universal probe ARC915 [54], a Methanosarcinaceae-targeting probe MSMX860, and a Methanosaeta-targeting probe MX825 [55]. The composition of biogas was measured using a gastight syringe (20 mL injection volume) and a gas chromatograph (GC, 7890A Agilent, Santa Clara, CA, USA) equipped with a thermal conductivity detector (TCD) [56]. The GC was fitted with the two Hayesep Q columns (80/100 mesh), two molecular sieve columns (60/80 mesh), and a Porapak Q column (80/100) operating at a temperature of 70 • C. The temperature of the injection and detector ports were 150 • C and 250 • C, respectively. Helium and argon were used as the carrier gases at a flow of 15 mL/min. Additionally, biogas was analyzed by the GMF 430 Gas Data analyzer. The content of methane (CH 4 ) and carbon dioxide (CO 2 ) was measured. The validated analytical procedure was calibrated against a standard curve. Therefore, the dependence of the analytical signal (peak area) as a function of the concentration of the analyzed component was determined for a series of standard gas mixtures.

Calculation Methods
The digestion coefficient (portion digested), i.e., the ratio of the organic VS load removed in the reactor to the VS load fed into the reactor, was determined using the following equation: where η F -digestion coefficient, %; VS in -concentration of organic compounds in the influent, g/dm 3 ; VS out -concentration of organic compounds in the digestate, g/dm 3 ; ρ in -influent density, g/cm 3 ; ρ out -digestate density, g/cm 3 ; Q in -daily volume of feedstock (in), cm 3 /d; Q out -daily volume of digestate (out), cm 3 /d. Biogas/CH 4 production per VS load was calculated as follows: Biogas/CH 4 production per VS load in the influent was calculated using the following equation: -biogas production per VS in the influent, cm 3 /gVS in ; V b/CH 4 -volume of biogas/CH 4 produced per influent load, cm 3 /d; Q in -single load of influent (by volume), cm 3 ; Q out −specific post-AD digestate out (by volume) cm 3 .

Statistical Methods and Optimization
All experimental variants were conducted in triplicate. Statistical analysis of the results was carried out using STATISTICA 13.1 PL package. The Shapiro-Wilk test was used to verify the hypothesis regarding the distribution of every researched variable. ANOVA was performed to establish the significance of differences between variables. The homogeneity of variance was determined using Levene's test. Significant differences between variants were determined via Tukey's honestly significant difference (HSD) test. A significance level of α = 0.05 was adopted for the tests.
Empirical equations were elaborated using stepwise regression with multiple regression. The equations were then used to estimate the correlation between the amount of methane and the post-AOP PSS parameters. Predictors having a significant impact on the changes in estimated parameters were determined in model systems. In addition, the accuracy of the models' fit to empirical data was estimated via a coefficient of determination. The significance of multiple regression models was verified by the F-test. A lack-of-fit test was conducted to evaluate whether the proposed models are sufficiently detailed by comparing the proposed models with full models (which included the remaining explanatory variables omitted in the proposed models). The developed models were subjected to estimation. Next, their fit to obtained results was evaluated by analysis of residuals. The assumption of normality of residual distribution was verified, and the models' accuracy was evaluated by deleting the residual values with respect to predicted values (Statistica 13.1 PL).

Organic Compounds
The adopted method of pre-treatment had no significant effect on volatile solid (VS) levels in the PSS, with VS values being similar across all series and variants (p > 0.05). VS in raw PSS was 78.7 ± 1.9 % TS (Table 3). In S1 (Fe 2+ /H 2 O 2 ), the VS levels ranged from 77.1 ± 2.7% TS (V1) to 74.4 ± 2.0% TS (V4). In S2 (Fe 2+ /H 2 O 2 ), the range was between 78.5 ± 0.4% TS (V1) and 76.3 ± 0.5% TS (V5). Trends in VS in the PSS are presented in Table 3. Atay and Akbal [57] have demonstrated that AOP can be used for efficient stabilization of sludge from municipal wastewater treatment plants, removing 26.8% VS using 0.07/1 Fe 2+ /H 2 O 2 and 60 g/kgTS H 2 O 2 [57]. Other researchers have reported successful experiments on textile sludge [58] and anaerobically digested sludge [59], among others. The failure to remove VS from the PSS most likely stems from insufficient oxidizing power of the reagents at the doses used [60]. Researchers have suggested that, in cases such as this, free hydroxyl radicals initially react with dissolved substances [61]. This is corroborated by Fontmorin and Sillanpää [59]. However, it is important to note that VS removal from sludge through pre-treatment could negatively affect the biomethane yield if the sludge is to be used as feedstock for AD [62]. Table 3. Changes in concentrations of organic compounds (VS) and dissolved TOC in the PSS after pre-treatment.

Series (S)-Variant (V)
Control S1V1 S1V2 S1V3 S1V4 S1V5 Dissolved TOC was 326 ± 10 mg/dm 3 in the control group, 296 ± 5 mg/dm 3 (the lowest) in S1V5, and 303 ± 8 mg/dm 3 in S2V5. The trends in dissolved TOC in the tested PSS are presented in Table 3. The TOC values were significantly different from the raw PSS (p < 0.05). Multiple studies have shown the Fenton and Fenton-like reactions to be highly effective at removing dissolved organics [63,64]. This has been demonstrated in treatments of various sewage and leachates [65][66][67]. There have also been reports of AOP inducing the partial breakdown of complex organics, which did not affect the TOC results but did result in better biodegradation of pollutants and lower toxicity [68]. One example of this phenomenon in practice has been presented by Catalkaya and Kargi [69], who tested the advanced oxidation of diuron in an aqueous solution by Fenton's reagent. The authors found that only 58% of diuron was mineralized after 240 min under optimal operating conditions, indicating the formation of certain intermediate products. No effect of H 2 O 2 and Fe(II) on TOC removal was found [69]. Similar trends have been reported by Pérez et al. [70] in processing waste from paper pulp treatment effluents.

Toxicity
This study shows that Fenton's reagent (Fe 2+ /H 2 O 2 ) significantly reduced the toxicity of the tested PSS. The toxicity of raw PSS was 53.3 ± 5.1%, but significant changes (p < 0.05) were noted very early into the pre-treatment, with toxicity dropping to S1V2 at 49.6 ± 2.5% (Figure 3a). The lowest statistically comparable levels (p > 0.05) were found in S1V4 and S1V5 at 38.3 ± 4.0% and 35.7 ± 3.2%, respectively ( Figure 3a). Accordingly, these were also the series with the highest toxicity removal rates (28.1 ± 2.9% and 33.1 ± 2.8% removal, respectively) ( Figure 3b). The use of Fe 2+ /H 2 O 2 reagents for reducing the toxicity of biodegraded feedstock has also been successfully tested in other studies [71]. Barbusiński and Filipek [72] demonstrated that Fenton's reagent could completely eliminate the high toxicity of industrial wastewater. Similarly, Gerulová et al. [73] found that preliminary treatment with Fe 2+ :H 2 O 2 at a 1:10 molar ratio can reduce the toxicity of metalworking wastewater fluids (MWF), improving their biodegradability [73]. Finally, Lin et al. [74] have obtained reduced toxicity in acrylic fiber manufacturing wastewater. The reduced toxicity brought on by the Fenton treatment is explained by its ability to convert refractory organic matter to smaller organic/inorganic molecules [75]. PSS treatment with Fe 3+ /H 2 O 2 showed highly variable performance in the present study. Significant toxicity reduction (p < 0.05) was achieved only in S2V3 and S2V4 (48.7 ± 1.5% and 46.0 ± 3.0%, respectively) ( Figure 3a). The highest tested dose of the reagents (S2V5) resulted in higher toxicity levels at 52 ± 2% (p > 0.05) (Figure 3a). This may be attributable to the two-step nature of OH • production in the Fenton-like process [76]. The first step generates OH 2 • and reduces Fe 3+ ions to Fe 2+ [77]. It is only in the second stage that the typical Fenton reaction takes over, with a catalytic decomposition of H 2 O 2 to free hydroxyl radicals [78]. Therefore, a Fenton-like reaction may be slower and less effective at removing organics and reducing toxicity [79,80]. Fe 2+ /H 2 O 2 has been shown to perform better than Fe 3+ /H 2 O 2 at treating textile dyeing wastewater [81]. The more complex nature of Fenton-like reactions poses a risk of incomplete degradation of H 2 O 2 to OH • and the presence of residual H 2 O 2 in the medium [82]. H 2 O 2 has a high oxidation potential, making it harmful to microorganisms [83]. Arslan-Alaton and Gurses tested the Fenton-like oxidation of antibiotic formulation effluent and found high levels of residual H 2 O 2 , which they attributed to the relatively slow and poor COD reduction kinetics [84].

Residual (Unreacted) H 2 O 2 and pH Changes in the PSS
No residual H 2 O 2 was detected in the Fe 2+ /H 2 O 2 variants S1V1-S1V3 (Figure 4a). Some amounts were found in S1V4 (126.7 ± 40 mgH 2 O 2 /dm 3 ) and S1V5 (170 ± 40 mgH 2 O 2 /dm 3 ) (Figure 4a). In contrast, the Fenton-like reaction produced higher rates of residual H 2 O 2 . Traces of unreacted oxidant were found in the PSS quite early into the experiment (S2V3) at 46.7 ± 15.3 mgH 2 O 2 /dm 3 , rising concurrently with the reagent doses in the subsequent variants (Figure 4a). Increases to 263.3 ± 47.2 mgH 2 O 2 /dm 3 and 403.3 ± 25.2 mgH 2 O 2 /dm 3 were found in S2V4 and S2V5, respectively (Figure 4a). The presence of residual H 2 O 2 in media processed using AOP (via Fenton or Fenton-like reactions) has been found by other studies as well [85][86][87]. Verma and Haritash [85] used this process to remove amoxicillin (AMX) wastewater and found that treatment with 375 mg/dm 3 [85]. The presence of residual H 2 O 2 may be an indication of reduced oxidation capacity in the Fenton system and a diminished degradation rate, possibly brought on by the removal of free hydroxyl radicals and generation of the HO • 2 radical [86,87]. The use of salt as an Fe ion donor for radical formation can also lower the pH of the treated media. This has been shown to play a particularly large role when using iron-sulfur compounds such as Fe 2 (SO 4 ) 3 , FeClSO 4 , and FeSO 4 ·7H 2 O [88,89]. The Fenton reaction is more efficient at low pH, as demonstrated by Alalm et al. [90] and Verma and Haritash [85]. The authors posited that the best performance was achieved at the range of pH from 2 to 4, with pH 3 being the most optimal choice [85]. However, it is important to note that lowering pH of sewage sludge intended to be anaerobically digested can be detrimental to digestion performance [91]. Methanogenic bacteria are sensitive to changes in the environment and require neutral pH for optimal metabolic activity [92]. With this in mind, we used chlorides as Fe ion donors in our study, as they do not produce significant reductions in pH in pre-treated PSS. In the Fenton reaction group, the pH varied across experimental variantsfrom 7.16 ± 0.15 in S1V1 to 6.90 ± 0.17 in S1V5 (p > 0.05) (Figure 4b). The Fe 3+ /H 2 O 2 series also showed some variance-from 7.10 ± 0.10 (S2V1) to 6.70 ± 0.10 (S2V5) (p > 0.05) (Figure 4b). By comparison, the raw PSS had a pH of 7.23 ± 0.21 (Figure 4b). There have been reports of highly advanced oxidization performance in near-neutral media. This includes a study by Chen et al. [93], where efficient degradation of wastewater tetracycline was achieved via a Fenton-like reaction at pH between 4 and 8 [93]. Another study [94] showed 100% removal of Rhodamine B (RhB) and 90% removal of tetracycline (TC) by a Fenton-like reaction at neutral pH [94].

Series (S)-Variant (V)
Control S1V1 S1V2 S1V3 S1V4 S1V5 Zawieja and Brzeska also managed to improve biogas yields from the AD of surplus sludge by advanced oxidation of sludge with Fenton's reagent [95], finding that the optimal process parameters were: Fe ion dose = 0.08 g Fe 2+ /gTS and Fe 2+ :H 2 O 2 ratio = 1:5. This configuration produced a biogas yield of 0.53 dm 3 /gVS (which represents a 35% increase in biogas production against the unprocessed sludge) and a digestion coefficient of 59%. The methane fraction in the biogas was unaffected, however, remaining at 70% [95]. A massive increase in biogas production-75% against the control-was achieved by Dewil et al. [96] by pre-treating surplus sludge with Fenton's reagent at a dose of 0.07 Fe 2+ /g H 2 O 2 , 50 mg H 2 O 2 /kgTS. Methane content in the biogas ranged between 65 and 70%, with its calorific value remaining unaltered [96].

pH and FOS/TAC
PSS pre-treatment with Fenton's reagent was found to have no significant effect in S1 (p > 0.05), with pH falling within the narrow range of 6.90 ± 0.12 to 7.10 ± 0.02 (Figure 10a). Lower pH values in the model digesters were noted for S2. In S2V5, the Fenton-like reagent triggered a pH reduction to a level of 6.88 ± 0.10 (Figure 10a). On the other hand, lower doses of Fe 3+ /H 2 O 2 did not reduce pH significantly (p > 0.05), with the pH ranging from 7.1 ± 0.10 to 6.98 ± 0.05 (Figure 10a). For comparison, the digester pH was 7.10 ± 0.11 in the control group (Figure 10a). This finding is supported by Zawieja and Brzeska [95], who did not observe any significant reductions in pH during AD of Fenton-oxidized surplus sludge, noting that the sludge pH was 7.74 [95]. However, according to some authors [103], the Fenton reaction converts organic material into organic acids and can lower pH. As such, pH drops during the Fenton reactions should be controlled to optimize treatment efficiency [103]. The FOS/TAC during the digestion of non-pretreated PSS was 0.39 ± 0.03 (Figure 10b). This variable was significantly reduced by AOP, regardless of whether radical formation was mediated by Fe 2+ or Fe 3+ (p < 0.05). The change was more pronounced in S2, with FOS/TAC ranging from 0.33 ± 0.02 (S2V2) to 0.30 ± 0.03 (S2V3) (Figure 10b). Substantial reductions in FOS/TAC were observed for S2V4 and S2V5 (to 0.28 ± 0.01 and 0.25 ± 0.02, respectively) ( Figure 10b). In the conventional Fenton group (Fe 2+ /H 2 O 2 ), the FOS/TAC ratio remained close to the optimal level for AD across the entire range of reagent doses tested. The lowest value was recorded for S1V5 at 0.31 ± 0.01, the highest-for S1V2 at 0.35 ± 0.02 (Figure 10b). The FOS/TAC is the ratio of volatile organic acid to alkaline buffer capacity, often used to assess process stability in anaerobic digesters. Literature reports [104,105] state that FOS/TAC needs to be between 0.2 and 0.6 for stable AD, as was the case in the present study. The FOS/TAC ratio exceeding 0.6 indicates suboptimal running parameters for anaerobic microbes and reduced biogas production [104,105].

Empirical Model and Correlations
Empirical equations were elaborated using multiple regression to estimate the methane yields. Methane production was found to correlate significantly with factors such as the dose of Fenton's reagent, toxicity, and VS after pre-treatment. The methane production model for series 1 (4) had an estimation error of ±0.4933 and accounted for approx. 99.88% of the biogas yield variation (R 2 coefficient of determination = 0.9988). The methane estimation model for series 2 (5) accounted for approx. 99.58% of the methane yield variation (R 2 coefficient of determination = 0.9958) at an estimation error of ±1.3523. The level of mapping the final methane production in the developed models concerning the results obtained in the experimental work was very high, which indicated that the adopted assumptions and the practical value of the optimization procedure used were correct. T-toxicity (%) VS-volatile solids after processing (%TS).
Most of the strong correlations found in series 1 pertained to the conventional Fenton reaction (Fe 2+ /H 2 O 2 ). No strong and significant relationships were found for the Fe 3+ /H 2 O 2 reagent used in series 2 (p > 0.05). A very strong negative association (R 2 = 0.9329) was noted between toxicity and methane levels in series 1 (Figure 11a). In contrast, this correlation was only moderate in series 1 (R 2 = 0.5034) (Figure 11b). Weak negative correlations (R 2 < 0.4) were found between FOS/TAC and methane levels in both experimental series (Figure 11c,d). A very strong negative correlation (R 2 = 0.9503) was noted between toxicity and the share of Archaea in series 1 (Figure 12a), whereas no such correlation was observed in series 2 (R 2 = 0.1345) (Figure 12b). The only strong positive correlation (R 2 = 0.8035) in series 1 was found between the share of Archaea and methane levels (Figure 12c), whereas the corresponding association for series 2 was only moderate (R 2 = 0.5713) (Figure 12d). The predicted correlated effect of Fenton's reagent dosage and toxicity on biogas and methane production is presented, respectively, in Figure 13a,b for series 1 and Figure 14a,b for series 2. In the Fenton reagent (S1) variants, strong linear correlations between the dose of Fe 2+ /H 2 O 2 and toxicity, presence of Archaea and CH 4 production were revealed. In subsequent variants, a tendency to significantly reduce the toxicity of PSS and an increase in the share of Archaea in the anaerobic bacterial community was observed. These phenomena were closely correlated with the increase in the amount of biogas produced and the per-centage of CH 4 content. Unlike S1, when the like-Fenton reagent (Fe 3+ /H 2 O 2 ) was used (S2), the linear relationships between the analyzed variables were disturbed. In S2, the highest chemical doses resulted in an increase in the residual H 2 O 2 concentration in PSS, which probably resulted in an increase in toxicity. This directly affected the decrease in the average number of Archaea in the population of anaerobic bacteria, as well as the decrease in the content of CH 4 in biogas. However, the high efficiency of biogas production was maintained. The increase in biogas synthesis efficiency in S1 (V4-V5) despite the increase in toxicity may be explained by the disintegration and destruction of the organic substrate structure as a result of AOP [106]. This could lead to an increase in the susceptibility of PSS biomass to biodegradation under anaerobic conditions [107,108]. Certainly, the process did not cause complete destruction of cellular structures, as the concentration of TOC in the dissolved phase did not reflect this. Nevertheless, the highest doses of chemical reagents tested were able to disrupt cell walls and accelerate biogas production. The positive effect of AOP on anaerobic digestion and biogas production has been proven in the literature [100,109,110]. In variants V1-V3, the toxicity was limited, which directly affected the increase in the share of Archaea in the bacterial community and the increase in the content of CH 4 in biogas. Apparently, however, the oxidizing (disintegrating) power was too low in these pre-treatment variants to increase biogas production.   These opposite phenomena (increase in biogas production, decrease in CH 4 content) disturbed the linear correlations between the analyzed variables in S2, which were weaker compared to the Fe 2+ /H 2 O 2 system. The lack of strong linear relationships when the like-Fenton reagent in S2 was used makes it a method in which it is difficult to estimate the achievable final effects. This is another, apart from higher technological efficiency, argument indicating the practical advantage of AOP based on the Fe 2+ /H 2 O 2 reagent system.

Conclusions
AOP pre-treatment of PSS using the conventional Fenton's reagent (Fe 2+ /H 2 O 2 ) and a Fenton-like reagent (Fe 3+ /H 2 O 2 ) was successful in reducing dissolved organics by singledigit percentages and, most importantly-in detoxifying this substrate. Considerably better toxicity removal performance was noted for the Fe 2+ /H 2 O 2 system. The poorer performance of the Fe 3+ /H 2 O 2 in this regard probably stemmed from the lower oxidation capacity of the Fenton-like reaction and the relatively high levels of H 2 O 2 residues in the PSS at the highest chemical reagent doses. The tested AOPs did not reduce the VS levels in the PSS biomass to a significant degree-a promising finding in terms of its use as an AD substrate.
Sludge pre-treated with Fe 2+ /H 2 O 2 proved to be a better feedstock for anaerobic digestion, both with regard to biogas/methane yields and the PSS portion digested. Significantly higher values were observed for the two highest reagent doses. This was 445.7 ± 21.6 cm 3 CH 4 /g VS (1.5 g Fe 2+ /dm 3 and 6.0 g H 2 O 2 /dm 3 ) and 453.6 ± 22.4 cm 3 CH 4 /gVS (2.0 g Fe 2+ /dm 3 and 8.0 g H 2 O 2 /dm 3 ). The differences between these variants were not statistically significant. Therefore, due to the economical use of chemical reagents, the optimal tested dose was 1.5 g Fe 2+ /6.0 g H 2 O 2 . The use of a Fenton-like reagent (Fe 3+ /H 2 O 2 ) resulted in lower AD efficiency (max. 393.7 ± 12.1 cm 3 CH 4 /g VS), and no strong linear relationships between the analyzed variables were found. It is, therefore, a more difficult method to estimate the final effects. Research has proven that AOP can be used to improve the efficiency of AD of PSS.
Strong correlations between the tested parameters were found for the conventional Fenton reaction (Fe 2+ /H 2 O 2 ). There was also a very strong negative correlation between toxicity and methane production, as well as between PSS toxicity and the share of Archaea in the microbial structure. No strong and significant correlations pertaining to reagent dosage were found for the Fe 3+ /H 2 O 2 system. Empirical equations were elaborated using multiple regression to estimate biogas and methane yields. Biogas and methane production was found to correlate significantly with such factors as the dose of Fenton's reagent, toxicity, and initial VS after PSS processing.