Influence of persulfate on transformation of phosphorus and heavy metals for improving sewage sludge dewaterability by hydrothermal treatment

Activated persulfate oxidation has been proven to be an efficient advanced sludge treatment technique to improve sludge dewaterability. This study investigates the influence of persulfate on the transformation of phosphorus (P) and heavy metals (HMs) during the hydrothermal treatment of sewage sludge. The hydrothermal temperature, time, and persulfate concentration are optimized by a Box-Behnken design to obtain the best sludge dewaterability, which is expressed by capillary suction time (CST). The highest CST reduction efficiency is 90.5% at the optimal hydrothermal temperature, time, and concentration of persulfate, which are 145 °C, 2 h, and 150 mg/g dry sludge (DS), respectively. The distribution and transformation of P and HMs with different persulfate concentrations (100–200 mg/g DS) during the hydrothermal process are investigated. Results show that more than 90% of the P and HMs in the sludge are retained in sludge cakes after the hydrothermal treatment. The addition of SPS can make the P in the sludge cakes transform into more stable P species according to the extraction capacity of sequential extracts. It can be found from the ecological risk indexes of the HMs that the addition of SPS during the hydrothermal treatment of sludge can reduce the environmental risk of HMs. This study provides insights into the P and HM distribution and transformation during hydrothermal treatment with persulfate, providing a reference for sludge recovery strategies.


Introduction
With the wide application of biological sewage treatment technology, a considerable quantity of sewage sludge is produced as a by-product during wastewater treatment (Neyens et al., 2004). Sewage sludge is a useful waste containing valuable resources, such as organic carbon, phosphorous (P), and nitrogen (N), along with hazardous substances, such as pathogens, heavy metals (HMs), and persistent organic pollutants (Zhang et al., 2009;Świerczek et al., 2018;Mahmoud et al., 2018). Therefore, sludge requires reasonable treatment and disposal before entering the environment. To reduce the costs of sludge transportation and disposal, a reduction of the sludge volume by solid-water separation is an essential step prior to sludge disposal (Li et al., 2018). Many technologies have been explored to improve sludge dewaterability, which is a key point for sludge volume reduction, including electro-dewatering technology (Saveyn et al., 2005), chemical conditioning (Xu et al., 2018a, b), and thermal drying (Olivier et al., 2015). Notably, hydrothermal treatment has proven to be an effective pretreatment method to enhance sludge dewaterability owing to the solubilization of organic matter (Kepp et al., 2000). Additionally, hydrothermal treatment can reduce sludge toxicity, control HMs, and recover resources (e.g., P and N) and energy from biosolids (Dewil et al., 2007;Funke and Ziegler, 2010).

Responsible Editor: Ricardo A. Torres-Palma
On the other hand, the persulfate advanced oxidation process has been widely used to improve sludge dewaterability due to its high solubility, strong oxidation potential, and resistance to pH changes (Zhen et al., 2012a, b). Persulfate can be activated by heat to produce sulfate radicals with strong oxidizability (Kordkandi and Forouzesh, 2014). Therefore, combing a hydrothermal treatment with persulfate can not only intrinsically enhance sludge dewaterability but also activate persulfate to generate sulfate radicals, which can further improve sludge dewatering.
The effects of hydrothermal treatment conditions on nutrient element (i.e., P and N) transformation, solid product properties, and HM risk have been widely investigated. Yu et al. (2019) investigated the effect of metal ions on P speciation during hydrothermal treatment, and their results showed that P immobilization was significantly influenced by the metal contents. Huang and Yuan (2016a, b) studied the effects of temperature, time, and the addition of catalyst on the fate of HMs during the hydrothermal treatment of sewage sludge. It was concluded that the total concentration of HMs in solid products increased with an increasing temperature, and a proper catalyst addition could enhance the immobilization of HMs in the solid products. Zhang et al. (2017) speculated on the pathway of N transformation during the hydrothermal treatment of sewage sludge, and the results indicated that only 20% N remained in the solid product. Even so, little attention has been given to the influence of persulfate on the transformation of P and HMs during the hydrothermal treatment process to enhance sewage sludge dewaterability.
Therefore, different concentrations of persulfate were used to evaluate the bioavailability of P and HMs in sludge cakes produced from hydrothermal treatment. The objectives of this study were (1) to optimize the hydrothermal temperature, time, and concentration of persulfate to achieve the best sludge dewaterability; (2) to evaluate the P transformation and speciation by sequential extracts during the hydrothermal process; and (3) to investigate the fraction distribution and ecological risk of HMs in the sludge cakes after hydrothermal treatment. Finally, this study sheds light on the transformation of P and HMs during persulfate hydrothermal treatment to enhance sludge dewaterability, providing basic knowledge for sludge reuse.

Materials
Raw sewage sludge was collected from Zhuankou Wastewater Treatment (Hanyang District, Wuhan, China), air-dried in an oven at 105 °C until it was completely dehydrated, and then ground to powder and passed through a 0.154 mm sieve. The dried sludge powder (RS) was stored in airtight bags for subsequent experiments. This pretreatment avoided biochemical reactions during sludge storage and precisely controlled the mass ratio of sludge to deionized water in the hydrothermal conditioning procedure. The initial characteristics of the raw sewage sludge are provided in Table 1.

Hydrothermal conditioning procedure
Hydrothermal conditioning experiments were performed in a well-designed 500-ml stainless steel reactor that could operate at high temperatures and pressures. RS and deionized water were mixed at a solid content of 10 wt%, and then SPS was added and the mixture was stirred at 600 rpm for 60 min before heating. The reactor was heated to the desired temperature and maintained at the final temperature for a certain time. The mixed solution was magnetically stirred at a speed of 150 rpm during the whole hydrothermal conditioning procedure. After conditioning, the reactor was naturally cooled to room temperature, and the suspension was separated by centrifugation at 4000 rpm for 20 min. The solid product was dried at 105 °C for 24 h for further analysis. The liquid product called process water was filtered through a 0.45-µm filter and stored at 4 °C for subsequent analysis.
To investigate the influence of persulfate on the transformation of P and HMs during the hydrothermal conditioning of sewage sludge, six experiments with different SPS concentrations were conducted; these experiments are shown in Table 2. The sludge samples conditioned with 100 mg/g DS SPS, 125 mg/g DS SPS, 150 mg/g DS SPS, 175 mg/g DS SPS, and 200 mg/g DS SPS at 145 °C are labeled S1, S2, S3, S4, and S5, respectively; the collected sludge cakes were denoted as CS1, CS2, CS3, CS4, and CS5, respectively. The liquid samples are denoted as LS1, LS2, LS3, LS4, and LS5, respectively. Sludge conditioned without the addition of SPS at 145 °C was used as the control and denoted as S0; its relevant liquid samples and sludge cakes were denoted as LS0 and CS0, respectively. In addition, 2 h was selected as the conditioning time according to response surface methodology (RSM) optimization.

Optimization design
The hydrothermal conditioning time, temperature, and SPS concentration were optimized by using a Box-Behnken design. The ranges and levels of the three constituents shown in Table S1 were determined by a set of preliminary experiments. The CST reduction efficiency was set as the response. Table S2 shows the details of the BBD experimental design and results.

Quantification of the P and heavy metal content
A combustion method was used to measure the total P content in all solid samples. Specifically, the RS and solid products were first combusted at 600 °C for 2 h in a furnace and then extracted by HCl solution (1 M) for 16 h. Then, the total P in the extract was measured as orthophosphate using the ascorbic acid assay on a UV-vis spectrophotometer (UV-1800).
The total HMs content of all samples was measured using microwave-assisted acid digestion and determined by inductively coupled plasma optical emission spectrometry (ICP-OES, Perking, Elmer8300, USA).

Fractionation procedure of the P and HMs
Hedley sequential fractionation was employed to quantify the major speciation of P (Hedley et al., 1982). Briefly, 0.2 g of solid product was placed into a 50-ml centrifuge tube and extracted sequentially with 20 ml of H 2 O, NaHCO 3 (0.5 M), NaOH (0.1 M), and HCl (1 M) at 25 °C for 16 h. The suspension was centrifuged, and the supernatant was filtered through a 0.45-µm filter and stored at 4 °C for further analysis. The solid residue was used for the next extraction step. The H 2 O and NaHCO 3 extractable P species (H 2 O-P, NaHCO 3 -P) were ascribed to the readily soluble phosphates and organophosphates. NaOH extractable P species (NaOH-P) were attributed to P sorbed on the Fe/Al mineral phases. HCl extractable P species (HCl-P) were attributed to insoluble phosphate minerals, such as Ca, Al, and Fe phosphates.
Finally, P retained in the sample is denoted as Residue-P. The P content in all extracts was analyzed by using an ascorbic acid assay with UV-vis spectrophotometry (UV-1800).
All the experiments were conducted in triplicate and the average value and standard deviation were calculated.
The Community Bureau of Reference (BCR) sequential extraction procedure was conducted to investigate the chemical speciation of HMs (Rauret et al., 1999). The four fractions of metal speciation included HMs that were soluble and exchangeable (F1), HMs that bound to iron and manganese oxides (reducible) (F2), HMs that bound to organic matter and sulfide (oxidizable) (F3), and the residual fraction (F4). All the details of the procedure above were published in our previous report (Xiong et al., 2018).

Sludge dewaterability
The sludge dewaterability was evaluated by the capillary suction time (CST) which was measured using a 304 M CST instrument (Triton, UK). The CST reduction efficiency (Y), which was used to assess the conditioning efficiency of hydrothermal treatment, can be obtained as follows: where Y is the CST reduction efficiency (%), CST b is the initial CST of the RS (s), and. CST a is the CST of the sludge after hydrothermal conditioning (s).

Bioavailability of the P and HMs
The existing forms of the P and HMs in the solid products can indicate their bioavailability. The proportions of H 2 O-P and NaHCO 3 -P represent the bioavailable P. The proportion of the F1 concentration indicated the toxicity and bioavailability of HMs.

Heavy metal risk assessment method
The geo-accumulation index (I geo ), monomial potential ecological risk factor (E r ), and potential ecological risk index (RI) were used to assess the potential ecological risks of the HMs in the RS and sludge cakes after the hydrothermal treatment (Yu et al., 2011;Huang et al., 2011), which can be calculated as follows: where C i is the sum of the F1 and F2 contents of the HM (mg/kg) and B i is the HM mean background value in Chinese soil (mg/kg).
where T r is the toxic response factor of the HMs and C r is the background value of soil in Wuhan city (Cheng et al., 2014). The values of T r and C r are shown in Table S3 (Chen et al., 2020). Table S4 shows the assessment of these three indexes (Huang and Yuan, 2016a, b).

Analysis
SPSS 19.0 was used for the statistical analyses, and the influence of different hydrothermal conditions on the P and HMs transformation was analyzed by one-way ANOVA (Tukey's test); p < 0.05 represented a significant difference.

RSM optimization results
Seventeen experiments with different formulations for RSM were conducted to optimize the temperature, time, and SPS concentration for sludge hydrothermal conditioning, with the CST reduction efficiency as the response value. The secondorder polynomial Eq. (5) was obtained after the data fitting and regression analysis as follows: where Y is the CST reduction efficiency (%), A is the hydrothermal conditioning temperature (°C), B is the hydrothermal conditioning time (h), and C is the SPS concentration (mg/g DS). The model for the CST reduction efficiency was significant for an F-value of 58.60 and Prob > F < 0.0001 according to the analysis of variance regression (Table S5).
The "lack of fit F-value" was 2.91, indicating that there was a 16.41% chance that the lack of fit occurred due to noise. The "predicted R 2 " of 0.8498 was reasonably consistent with the "Adj R 2 " of 0.9701, indicating that the model could reliably describe the effects of hydrothermal temperature, time, and concentration of SPS on the sludge dewatering performance. All the above analyses indicated that the model was significant. In this model, A, B, C, AC, BC, A 2 , B 2 , and C 2 were significant because their values of "Prob > F" were lower than 0.05. The predicted versus actual values for the CST reduction efficiency are shown in Fig. S1, which indicated that the model was adequate and reliable, as the majority of the data points were close to the regression line. Figure 1 shows the contour and 3D response surface plots of the proposed model for CST reduction efficiency. The interaction of the other two variables was investigated when the third factor was retained at zero levels. Figure 1a displays the interaction between the hydrothermal temperature and time which indicated that the CST reduction efficiency increased as the hydrothermal time was within the range of 1-2 h. The highest CST reduction efficiency was recorded when the hydrothermal time ranged from 1.5 to 2 h and the temperature was 140-150 °C. Gao et al. (2019) investigated the effect of the temperature and time of hydrothermal carbonization on sludge dewatering by coupling with mechanical compression. They reported that the optimal temperature was 160-240 °C. In addition, an increase in the time had a negligible or slight effect on the CST reduction efficiency. Figure 1b indicates that at a higher temperature (180-200 °C), the CST reduction efficiency could hardly be affected by the increase in SPS concentration. The CST reduction efficiency demonstrated a U-curve with an increasing SPS concentration when the temperature was in the range of 120-160 °C. Figure 1c shows the interactive influence between the hydrothermal time and SPS concentration. At longer durations (2-3 h), the CST reduction efficiency changed slightly with an increase in time. The CST reduction efficiency showed a U-curve with an increasing SPS concentration. Park et al. (2018) found that the properties of the hydrochar and biomass dewaterability could improve with a hydrothermal treatment time of 1 h. We found that the significant influencing factors on the CST reduction efficiency in the experiment were the hydrothermal temperature and SPS concentration. Han et al. (2019) found that the temperature of the thermal hydrolysis pretreatment had a larger effect on the P complexation of sewage sludge than the reaction time. The optimal experimental conditions obtained from the RSM model of the hydrothermal temperature, time, and SPS concentration were found to be 145 °C, 2 h, and 150 mg/g, respectively. The optimal dewatering conditions were confirmed by five comparison experiments.

Effects of the SPS concentration on the distribution of P between the liquid and solid
The distribution of the P between the liquid and sludge cake is shown in Table S6. The results demonstrated that large proportions of the initial P were retained in the sludge cakes, which was consistent with other reports . The results of their study suggested that hydrothermal carbonization could efficiently immobilize P in cow manure. The unit mass of P in sludge cakes clearly increased after the hydrothermal treatment owing to the volatilization of organic matter (Han et al., 2019). In addition, the hydrothermal treatment process may suppress the P release into liquid; therefore, P was retained on the hydrochar of sewage sludge. P gradually increased in the sludge cakes with an increasing SPS concentration due to the degradation of organic sludge flocs being promoted by SO 4 • − radicals (Zhen, et al., 2012a, b). P was released from the sludge cells during the hydrothermal process and combined with Fe, Al, Mg, and Ca (Appels et al., 2008). Thus, the addition of SPS could inhibit the release of P into the liquid and promote P enrichment in the sludge cakes. Figure 2 shows that P in sequential extracts of the RS and sludge cakes after the hydrothermal treatment with different SPS concentrations showed different fractionation behaviors. Generally, the H 2 O-P and NaHCO 3 -P contents in the sludge cakes after the hydrothermal treatment were lower than those in RS (p < 0.05). This result indicated that the hydrothermal treatment decreased the P bioavailability of the sludge cakes, and the same finding was reported by Huang et al. (2017). After the hydrothermal treatment, H 2 O-P and NaHCO 3 -P decreased with increasing HCl-P and Residue-P contents (p < 0.05). The decrease in the P bioavailability may be because orthophosphates degraded from organophosphates formed phosphate precipitates (e.g., calcium phosphate) or adsorbed to minerals in the sludge; therefore, unstable inorganic P may also be transferred to more stable P species due to dissolution or recrystallization (Huang and Tang, 2016). The amounts of H 2 O-P and NaHCO 3 -P decreased to 0.015 mg/g DS and 0.38 mg/g DS, respectively, after the hydrothermal treatment with 200 mg/g DS SPS compared to 0.917 mg/g DS and 2.81 mg/g DS for RS. These results indicated that the addition of SPS promoted P release from the sludge cells through oxidation by the SO 4 • − radicals, and the released P transformed into more stable P species. For instance, amorphous Ca-Mg P could crystallize to form less soluble phosphates, such as apatite (Xu et al., 2018a, b). In addition, the content of H 2 O-P is commonly used to assess the liability of P loss from solid P fertilizer for land utilization because the discharge of P can result in the eutrophication of aquatic ecosystems . Based on the above analysis, it can be concluded that the addition of SPS during the hydrothermal conditioning process could fix P in sludge and inhibit the eutrophication.

Total HMs contents in the liquid and solid
The total contents of Cu, Zn, Ni, Pb, and Cr in the RS, liquid, and sludge cakes after the hydrothermal treatment are shown in Table 3. The above HMs were studied for their relatively high content in the RS. The results suggested that more than 90% of all HMs were retained in the sludge cakes, which was consistent with other reports (Leng et al., 2014;Liu et al., 2018). According to HMs contents, RS contained higher contents of Cu and Zn, less Pb and Cr, and the lowest concentration of Ni, which was in accordance with the results reported by Shi et al. (2013).
After the hydrothermal treatment, the variation in HMs in the sludge cakes was similar to that of the HMs in RS. Compared with HMs content in RS, there was an increase in the sludge cakes, which was mainly attributed to the degradation of organic matter in the sludge during hydrothermal treatment (Chen, et al., 2017).

Fraction distribution of HMs in the sludge cakes
The distribution of four fractions (F1, soluble and exchangeable; F2, reducible; F3, oxidizable; F4, residue) of Cu, Ni, Pb, Zn, and Cr in the RS and sludge cakes are shown in Fig. 3; these results were obtained by the BCR sequential extraction scheme. The mass percentage of the oxidizable and residual fractions of Cu accounted for 69% in the RS. Hu et al. (2018) also reported that Cu always existed as oxidizable and residual fractions that had high bioavailability and mobility. The percentage in the oxidizable fraction clearly increased after the hydrothermal treatment (48.4% for CS0, 75.1% for CS1, 72.2% for CS2, 69.6% for CS3, 48.7% for CS4, 59.4% for CS5), indicating that Cu was mainly bound to the organic matter and sulfides in the RS and sludge cakes after the hydrothermal treatment. The exchangeable and reducible fraction of Cu decreased, suggesting less mobility and potential bioavailability of Cu after the hydrothermal treatment (Kazi et al., 2005).
The variation trend of Ni was similar to that of Cu except for the sludge cake obtained without the addition of SPS. In the RS, the mass percentage of the oxidizable and residual fractions accounted for as much as 73%. Regarding CS0, the majority of Ni was present in the oxidizable fraction (83%), indicating that it was mainly associated with strong organic ligands. With an increase of SPS concentration, the residual fraction of Ni increased from 0.65% (CS0) to 62% (CS5) which showed that the toxicity of Ni decreased significantly after the hydrothermal treatment with SPS. The key reason for the above results could be the strong physic-chemical effects and crystalline compound immobilization during the hydrothermal process (Sun et al., 2018). With respect to Pb in the RS, the highest mass percentage (49.7%) was found in the reducible fraction, suggesting that Pb mainly bound to the Fe and Mn oxides. The exchangeable fraction in the RS was 19.8% and decreased to 1-2% after hydrothermal treatment. Moreover, the residual fraction increased from 22% in the RS to 40-74% in the sludge cakes after the hydrothermal treatment. These results illustrated that the mobility and toxicity of Pb decreased after the hydrothermal treatment, which was in accordance with the results reported by Yuan et al. (2011) and Shao et al. (2015).
Regarding Zn, the mass percentage of the exchangeable and reducible fractions accounted for 65% in the RS, indicating its higher toxicity and availability. He et al. (2010) also found that Zn in sludge mainly presented as a reducible fraction (43.4%), which bound to the Fe and Mn oxides. After the hydrothermal treatment, Zn was mainly present in the oxidizable fraction (40-70%) in all sludge cakes. Chen et al. (2015) also found that Zn easily transferred to the oxidizable fraction during thermal treatment. As the SPS concentration increased, the residual fraction of Zn increased from 15% in CS1 to 32% in CS5. The exchangeable fraction of Zn decreased from 32.8% in RS to 3.7% in CS5, showing the immobilization of Zn after the hydrothermal treatment with SPS.
Unlike Cu, Ni, Pb, and Zn, the highest percentage of Cr was found in the residual fraction (85%) in the RS, decreasing to 5.8% in the CS0 after hydrothermal treatment, and increasing to 80% with the addition of SPS. Shao et al. (2015) also reported that Cr in municipal sewage sludge mainly existed as residual fraction. Cr was prone to transfer to unstable fractions during the hydrothermal treatment. SPS could promote the oxidation of these insoluble Cr fractions to stable Cr complexes. The exchangeable fraction increased from 1% in RS to 2.7% in CS0 and then decreased to 0.5% in CS5. These results showed that the addition of SPS was good for the immobilization of Cr. It was also reported in our previous work that the addition of SPS could reduce the risk of HMs during the conditioning of sewage sludge by Fe 2+ -sodium persulfate oxidation and rice husk (Xiong et al., 2018).
Comparing the residual fractions of Cu, Ni, Pb, Zn, and Cr in the RS and all sludge cakes, it can be found that the addition of SPS during the hydrothermal treatment of sludge could immobilize HMs. Figure 4 shows the values of I geo , E r , and RI. Regarding the RS, the I geo value of Cu and Pb was between 0 and 1, which indicated that Cu and Pb were uncontaminated to moderately contaminated in the environment. The I geo value of Zn was between 1 and 2, which demonstrated that 6.43 ± 2.13 303.8 ± 21.1 12.05 ± 0.07 289 ± 11.3 11.1 ± 0.07 293 ± 9.90 11.1 ± 0.28 300.5 ± 10.6 12.05 ± 0.07 428 ± 12.7 11.7 ± 0.71 315 ± 11.3 Zn was moderately contaminated. The I geo values of Ni and Cr were below 0, which meant that Ni and Cr demonstrated practically uncontaminated. Except for Zn, the I geo values of Cu, Ni, Pb, and Cr of all the conditioned sludge cakes were below 0, indicating that Cu, Ni, Pb, and Cr demonstrated practically uncontaminated in the environment. E r represents the toxicity of the different HMs. It was obvious that the E r value of the conditioned sludge cakes decreased compared with that of the RS. The E r value of all samples was below 40, indicating that the ecological risk of the RS and sludge cakes was low. The potential ecological risk of the HMs was Pb > Cu > Zn > Ni > Cr based on the E r value. According to the value of RI, the ecological risk level of the RS and all conditioned sludge cakes was low in the environment. These results indicated that the ecological risk indexes of the HMs decreased after the hydrothermal treatment, which was in agreement with the analysis of the fraction distribution in Section 3.3.2.

Conclusions
This study investigated P and heavy metals transformation during the hydrothermal treatment of sewage sludge to improve sludge dewaterability. The optimal hydrothermal temperature, time, and concentration of persulfate were 145 °C, 2 h, and 150 mg/g DS, respectively; these parameters were determined by a three-level and variable Box-Behnken design. Different concentrations of persulfate were added to explore its effects on the P and heavy metals distribution and transformation during hydrothermal treatment with sewage sludge. Only a small proportion of P was transferred into the liquid products after the hydrothermal treatment. The extraction capacity and efficiency of P sequential extraction indicated that the hydrothermal treatment decreased the P bioavailability of the sludge cakes. Residue-P in the sludge cakes increased as the persulfate concentration increased. According to the results of the fraction distribution of HMs in the sludge cakes and the environmental risk assessment of HMs, it can be concluded that the addition of persulfate during the hydrothermal treatment of sludge could immobilize the HMs. These results indicated that the combination of persulfate and hydrothermal technology to treat sludge would have a nonnegligible effect on the sludge dewatering performance and P and heavy metals fractions, which could be considered during sludge reclamation. In addition, the dose of persulfate should be optimized to reduce the cost in the large-scale application.

Supplementary Information
The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s11356-022-18624-1. Funding This work was supported by the National Natural Science Foundation of China (51908233). National Natural Science Foundation of China,51908233,Qiao Xiong Data Availability All data generated or analyzed during this study are included in this published article (and its supplementary information files).

Declarations
Ethics approval and consent to participate Not applicable.

Competing interests
The authors declare no competing interests.