Enhanced dewaterability of waste activated sludge by a combined use of permanganate and peroxymonosulfate

Ever-increasing efforts have been made to develop rapid and practical conditioning methods of sludge dewatering. This study demonstrated an innovative combination of potassium permanganate (KMnO4) and peroxymonosulfate (PMS) for sludge dewatering. The combined use of KMnO4 and PMS (KMnO4/PMS) showed its superiority in improving sludge dewaterability over the separate use of KMnO4 or PMS. By dosing 4 mmol g−1 VSS KMnO4 and 3 mmol g−1 VSS PMS, the dewaterability of waste activated sludge (WAS) significantly enhanced as capillary suction time (CST) decreased from 73.65 s to 24.65 s while the water content of dewatered sludge cake (WC) decreased from 78.96% to 70.47%. Apart from CST and WC, the KMnO4/PMS process could also affect negative zeta potential, sludge flocs size and the concentrations of protein and polysaccharide in extracellular polymeric substances (EPS). The enhanced sludge dewaterability and changes of the physicochemical characteristics of the WAS samples during the KMnO4/PMS process were actually ascribed to sulfate radicals (SO4˙−) and hydroxyl radicals (HO˙) in situ generated via PMS activation by manganese oxides (MnOx) in the states of MnO2 and Mn3O4 transferred from KMnO4 oxidation, which was verified by transmission electron microscopy (TEM), energy dispersive X-ray spectroscopy (EDX) techniques and radical scavenging experiments. Moreover, the Fourier transform infrared spectroscopy (FTIR) analysis further confirmed that the in situ generated SO4˙− and HO˙ could improve sludge dewaterability. Thus, the KMnO4/PMS process could be considered as a promising conditioning method of sludge dewatering.


Introduction
The management of large quantities of waste activated sludge (WAS) produced from activated sludge processes has posed great challenges to municipal wastewater treatment plants (WWTPs) due to great economic and environmental burden. [1][2][3][4] As reported, the treatment and disposal of excess sludge usually accounts for up to 60% of the total operating cost in WWTPs. 2 Consequently, how to substantially cut down this kind of cost has become a signicant issue, and many researchers and practitioners have tried their best to tackle this problem. WAS has a high water content in the form of free and bound water (over 90%). 5 And sludge dewatering has been generally considered as a promising strategy in reducing sludge volume and cost of sludge transport and ultimate disposal. 3,4 However, the sludge dewatering process is usually inhibited due to the presence of extracellular polymeric substances (EPS) that can retain a large amount of water, especially bound water. 6,7 Hence, increasing efforts have been focused on development of efficient sludge dewatering techniques, which can promote the release of bound water from EPS. So far, thermal, ultrasonication, freezing and thawing, and chemical oxidation have been investigated to improve sludge dewaterability. [7][8][9][10][11][12] Among all the developed methods, advanced oxidation processes (AOPs) have recently shown their superiority in sludge disintegration and dewatering.
AOPs usually rely on the produced highly reactive radical species, such as hydroxyl radicals (HOc) and sulfate radicals (SO 4 c À ) to disrupt WAS and degrade EPS, thereby enhancing sludge dewaterability. 2,13,14 It has been reported that the Fenton (combination of H 2 O 2 and Fe 2+ ) and Fenton-like (e.g., simultaneous addition of H 2 O 2 and Fe 3+ ) processes based on HOc production could efficiently improve the dewaterability of WAS. 1,7 However, only under strongly acidic conditions (pH ¼ 2-5) can these processes generate sufficient amounts of HOc for better sludge dewaterability, which signicantly hindered the sludge dewatering process and subsequent disposal. [15][16][17] Compared to HOc (E 0 ¼ 1.8-2.7 V), SO 4 c À owns a higher redox potential (E 0 ¼ 2.5-3.1 V), a longer life time (30-40 ms) and a wider working pH range (pH 4-9). 4,18 Therefore, pretty much attention has been recently paid to the SO 4 c À -based AOPs as new alternative methods of sludge conditioning. SO 4 c À is usually generated through the activation of peroxydisulfate (PDS, S 2 O 8 2À ) and peroxymonosulfate (PMS, HSO 5 À ) by ultraviolet (UV), heat, base and (transition) metal catalysts. [19][20][21][22][23][24][25] Our recent study investigated a-MnO 2 with different morphologies (nanoparticles, nanoowers and nanorods) synthesized via a facile hydrothermal method as activators of PMS for ciprooxacin (CIP) degradation, and found that a-MnO 2 nanoowers achieved much higher degradation efficiency of CIP than the other three MnO 2 (a-MnO 2 nanorods, a-MnO 2 nanoparticles and commercial MnO 2 ), which would be attributed to its higher catalytic activity of PMS and larger quantities of SO 4 c À produced from its activation of PMS. 26 Interestingly, Cui et al. recently studied chemical oxidation of benzene and trichloroethylene (TCE) by an innovative combined use of permanganate (KMnO 4 ) and PMS (KMnO 4 / PMS), and also found that it was mainly colloidal and amorphous MnO 2 in situ generated from KMnO 4 oxidation that activated PMS to trigger powerful SO 4 c À -mediated oxidation of both benzene and TCE. 27 As known to us all, both KMnO 4 and PMS have been widely used as powerful oxidants to destruct organic pollutants in water and wastewater. Recently, the two powerful oxidants were also applied to enhance the lterability of waste activated sludge. 28,29 Wu et al. reported that KMnO 4 could efficiently disintegrate the excess sludge with soluble chemical oxygen demand (SCOD) increasing by 3473%, and that the optimal KMnO 4 /sludge solid mass ratio was 0.1 with a stable disintegration degree (DD COD ) of about 34% while suitable reaction time was 30 min. 28 Niu et al. found that PMS oxidation could effectively break sludge particles, and that EPS increased signicantly and transferred to slime layer aer PMS treatment. 29 Yang et al. also reported that PMS could facilitate the disintegration of WAS and the biodegradability of organics could be enhanced aer treatment by PMS. 30 Although it has been proven that KMnO 4 and PMS could disintegrate WAS and enhance sludge dewaterability, the synergistic effect of both oxidants on the dewaterability of WAS still remains unknown.
In this paper, we proposed a combined use of KMnO 4 and PMS (KMnO 4 /PMS) for sludge dewatering. The dewaterability of WAS during the KMnO 4 /PMS process by the measurement of variations in capillary suction time (CST) and water content of dewatered sludge cake (W C ) was initially investigated. Then, the physicochemical characteristics of WAS (CST, W C , zeta potential, particle size and the concentrations of protein and polysaccharide in EPS) at different dosages of KMnO 4 and PMS were further studied. Finally, the mechanism of sludge dewatering by the KMnO 4 /PMS process was deeply explored.

Raw sludge and chemicals
The raw sludge was collected from Qige municipal wastewater treatment plant in Hangzhou, China. A sieve with 20 mesh was then used to remove big gravels and debris from the collected sludge samples. The supernatant was nally decanted to acquire the denser sludge samples used in this study aer plain sedimentation. The acquired denser sludge samples were stored at 4 C in a refrigerator prior to use, and all the experiments were completed within 48 h. The basic characteristics of the sludge samples were measured and shown in Table 1.

Experimental procedures
All the sludge dewatering experiments were conducted in an Erlenmeyer ask with a volume of 500 mL. In each test, 300 mL sludge sample was added and then mixed with a stoichiometric amount of PMS. When PMS was completely dissolved under thorough stirring, H 2 SO 4 or NaOH was used to adjust the solution pH to 7. Finally, the reaction was immediately initiated by dosing a calculated amount of KMnO 4 in a water bath apparatus (300 rpm), and was ceased aer 120 min. All the experiments were duplicated at least, and the average values with standard deviations were presented.

Analytical methods
2.3.1. CST and W C determination. The dewaterability of WAS is commonly evaluated by CST and W C . CST was measured using a CST instrument (Type 304, Triton Ltd., UK) equipped with an 18 mm diameter funnel and a standard Whatman no. 17 chromatography-grade paper. To determine W C , the sludge samples were pretreated by a vacuum ltration method. Briey, 50 mL sludge sample was poured into a standard Buchner funnel and ltered at a constant vacuum pressure for 15 min. The water content of the vacuum-ltered sludge cake trapped by the lter paper was measured according to the standard method, and calculated according to eqn (1): where m 1 is the weight of the lter paper, m 2 is the weight of the vacuum-ltered sludge cake, and m 3 is the weight of the vacuum-ltered sludge cake completely dried at 105 C.

EPS extraction and analysis.
EPS were extracted using a heat extraction method described by Yu et al. 6 Briey, 50 mL sludge sample was centrifuged at 4000g for 5 min at 4 C to acquire the supernatant as soluble EPS (S-EPS). The sludge pellets were then diluted with a 0.05 w/v % NaCl solution that was pre-heated to 70 C. The diluted sludge solution was immediately sheared by a vortex mixer for 1 min and centrifuged at 4000g for 10 min at 4 C to collect the supernatant as the loosely bound EPS (LB-EPS). The residual sludge pellets were resuspended in a 0.05 w/v % NaCl solution and heated to 60 C in a water bath for 30 min. At short notice, the resuspended sludge solution was centrifuged at 4000g for 15 min at 4 C, and the supernatant was collected as the tightly bound EPS (TB-EPS). Before the analyses of protein and polysaccharide (assumed as dominant components of EPS), the EPS samples were pretreated with an optimal amount of Na 2 SO 3 to quench the residual colored KMnO 4 . The pretreated EPS sample was then ltered through a 0.45 mm cellulose acetate membrane. The protein content was quantied by the modied Lowry-Folin method using bovine serum albumin as the standard. 31 The polysaccharide content was determined by the phenol-sulfuric acid method using glucose as the standard. 32 2.3.3. Other analyses. The zeta potential of the WAS samples was analyzed using a Zetasizer Nano (Nano-ZS90, Malvern Instrument Ltd., UK). The particle size distribution was measured by a laser particle size analyzer (LAP-W2000H) using ultrapure water as the dispersant. The Fourier transform infrared spectroscopy (FTIR) analysis of sludge samples was recorded in a wavenumber range of 4000-400 cm À1 by a Nicolet 6700 spectrometer (Thermo, USA).
Given the difficulty in separating the formed solids from the conditioned WAS samples, the extracted EPS solution was used and treated with the KMnO 4 /PMS process to monitor whether manganese oxides (MnO x ) formed aer KMnO 4 oxidation of the organics in EPS. Briey, aer the extracted EPS solution was treated with the KMnO 4 /PMS process, the reaction solution was ltered through a 0.45 mm cellulose acetate membrane to collect the brown-red solids. Then, the collected solids were rinsed alternately several times using ethanol and ultrapure water, and vacuum-dried at 60 C overnight. Finally, the dried solid samples were scanned by transmission electron microscopy (TEM) to capture the microscale crystallization and structure and energy dispersive X-ray spectroscopy (EDX) to analyze the elemental composition.   4 and PMS (KMnO 4 /PMS), respectively. As observed, the CST increased from 73.65 s to 140.60 s and 278.55 s, respectively when the WAS samples were conditioned by 7 mmol g À1 VSS PMS and KMnO 4 alone, respectively, which indicated that neither PMS alone nor KMnO 4 alone could efficiently improve the sludge dewaterability, and that the two individuals even worsened the dewaterability. The deterioration of sludge dewaterability was likely resulted from the progressive cell lysis and further releasing of intracellular biopolymers under the inuence of KMnO 4 or PMS when the intracellular biopolymers could not be degraded effectively in current condition. 33 Lee et al. also found that the PMS conditioning could deteriorate the dewaterability aer the WAS samples were treated with different dosages of PMS at room temperature. 5 Surprisingly, a combined use of 4 mmol g À1 VSS KMnO 4 and 3 mmol g À1 VSS PMS instantly decreased the CST from 73.65 s to 24.65 s, and apparently showed its synergistic effect on signicantly improving the sludge dewaterability. Besides, the decrease of W C (from 78.96% to 70.47%) also indicated the synergistic effect aer the WAS samples were conditioned by the KMnO 4 /PMS process while the W C values almost remained unchanged aer the samples were treated with PMS alone and KMnO 4 alone (79.25% and 77.07%, respectively). The synergistic effect of the KMnO 4 /PMS process on the WAS dewaterability was interpreted with the decrease of both CST and W C , which also implied the generation of SO 4 c À and HOc during the KMnO 4 /PMS process as mentioned in Section 1.  Fig. 2, when PMS dosage was xed at 1 mmol g À1 VSS, the increase of KMnO 4 dosage from 2 mmol g À1 VSS to 6 mmol g À1 VSS did not decrease the CST, and instead jeopardized the dewaterability, which indicated the absence of the synergistic effect of the KMnO 4 /PMS process at a low PMS dosage. This might be because the in situ generated SO 4 c À and HOc during the KMnO 4 /PMS process was too few to improve the dewaterability of the WAS samples when PMS dosage was too low. However, increasing PMS dosage to 2 mmol g À1 VSS efficiently enhanced the dewaterability, and the CST decreased from 73.65 s to 54.30 s, 46.45 s and 63.45 s, respectively with increasing KMnO 4 dosage from 2 mmol g À1 VSS to 6 mmol g À1 VSS. When PMS dosage continuously increased to 5 mmol g À1 VSS, the CST decreased to 22.45 s, 23.85 s and 22.00 s, respectively as KMnO 4 dosage increased from 2 mmol g À1 VSS to 6 mmol g À1 VSS. This Paper is because the higher PMS dosage, the more SO 4 c À and HOc generated to achieve better sludge dewaterability. 34 The similar results were also reported by Liu et al. when the Fe 2+ -activated PMS was used for sludge conditioning. 4 Remarkably, the CST values did not decrease all the way, and instead decreased rstly and then increased with sustainably increasing KMnO 4 dosage, which might be ascribed to the consumption of the generated SO 4 c À and HOc by KMnO 4 with a too high dosage. 35 Liu et al. also observed the similar results when using the magnetic Fe 3 O 4 -MnO 2 core-shell nanocomposites-activated PMS for 4chlorophenol degradation. 36 3.2.2. Water content of dewatered sludge cake (W C ). The proles of W C of the WAS samples aer conditioning of the KMnO 4 /PMS process with different KMnO 4 and PMS dosages were depicted in Fig. 3. As can be seen, the W C changes in the WAS samples were quite similar to the CST changes discussed in Section 3.2.1. By dosing 1 mmol g À1 VSS PMS, the W C aer conditioning with 2 mmol g À1 VSS, 4 mmol g À1 VSS and 6 mmol g À1 VSS KMnO 4 , respectively (82.11%, 76.73% and 77.56%) almost remained unchanged, compared to the W C of the raw sludge (78.96%). This suggested that the synergistic effect of the KMnO 4 /PMS process with 1 mmol g À1 VSS PMS would not work. However, the W C of the WAS samples would continue to decrease with the ever-increasing PMS dosages, which was also attributed to more SO 4 c À and HOc generated from the KMnO 4 / PMS process. The W C aer conditioning could decrease to about 67.5%, and by 11.46% compared to the W C of the raw sludge, which was higher than the variation in the W C of the WAS samples conditioned by the Fenton-like process (8.7%). 1 3.2.3. Zeta potential. To deeply investigate the effect of the KMnO 4 /PMS process on the WAS dewaterability, the zeta potentials of the WAS samples before and aer conditioning were also measured. The unconditioned raw sludge was negatively charged with the zeta potential of À17.7 mV, which was due to the ionization of the carboxyl and amino in the protein of EPS. 37 It can be obviously observed from Fig. 4 that the zeta potentials of the WAS samples were affected by dosing either KMnO 4 or PMS. The zeta potential tended to increase to 0 mV with increasing KMnO 4 or PMS dosage. By simultaneously dosing 4 mmol g À1 VSS KMnO 4 and 5 mmol g À1 VSS PMS, the zeta potential increased to À2.15 mV. The increase in the zeta potential suggested the ever-increasing sludge dewaterability, which quite matched with the changes of the CST and W C . Zhen et al. also reported that the sludge dewaterability improved as the zeta potential increased in the Fe 2+ -activated persulfate process. 38 According to the DLVO theory, the aggregation of sludge ocs is mainly controlled by the surface charge. The decrease in the negative surface charge of sludge ocs would reduce the electrostatic repulsion and bring about the increment of interaction energy, which led to the aggregation of sludge ocs and nally enhanced sludge dewaterability. Therefore, it can be concluded that sludge ocs are easier for agglomeration and sedimentation, and thus dewatering as the zeta potential increases. 39   3.2.4. Particle size distribution. As depicted in Fig. 5, effect of the KMnO 4 /PMS process on particle size distribution of the WAS samples was also investigated. The D 50 value of the unconditioned sludge ocs was determined to be 39.2 mm. When the WAS samples were conditioned with 2 mmol g À1 VSS KMnO 4 and 1 mmol g À1 VSS PMS, the D 50 value decreased to 32.95 mm. With PMS dosage increasing to 5 mmol g À1 VSS, the D 50 value also decreased to 30.08 mm. By dosing 4 mmol g À1 VSS and 6 mmol g À1 VSS KMnO 4 , the D 50 values of the conditioned sludge ocs decreased from 30.42 mm and 28.35 mm to 25.78 mm and 24.44 mm, respectively with increasing PMS dosage. Therefore, increasing both KMnO 4 and PMS dosages could reduce the size of the conditioned sludge ocs. The decrease of the sludge ocs size suggested that the KMnO 4 /PMS process could efficiently break the sludge ocs to enhance the sludge dewaterability as the ocs breakage could provide more passages for the removal of water in the inner of sludge ocs. 40 It was also reported by Higgins and Novak that the sludge dewaterability was adversely impacted by the ocs in the range of 1-100 mm and decreased as the quantities of ocs in this size range increased. 41 3.2.5. Extracellular polymeric substances (EPS). EPS can have signicant effects on the compressibility of sludge and the ability to bind water molecules of sludge ocs, and thus the dewaterability of sludge. 42 Since EPS are mainly composed of protein and polysaccharide, variations in the concentrations of protein and polysaccharide in different EPS fractions (S-EPS, LB-EPS and TB-EPS) were monitored to reveal why the KMnO 4 / PMS process could improve sludge dewaterability. As illustrated in Fig. 6a, with the increase of PMS dosage, the protein content in total EPS tended to decrease. By dosing 2 mmol g À1 VSS KMnO 4 , the protein content in total EPS decreased from 53.46 mg g À1 VSS to 36.99 mg g À1 VSS while the protein content in S-EPS increased from 21.57 mg g À1 VSS to 29.31 mg g À1 VSS before decreasing to 21.59 mg g À1 VSS, as PMS dosage increased. Moreover, the protein content in LB-EPS decreased from 25.21 mg g À1 VSS to 9.01 mg g À1 VSS while the protein content in TB-EPS decreased from 9.68 mg g À1 VSS to 6.39 mg g À1 VSS. Similar changes of the protein content in total EPS, S-EPS, LB-EPS and TB-EPS could also be observed when 4 mmol g À1 VSS and 6 mmol g À1 VSS KMnO 4 were dosed. In contrast to Fig. 6a and b showed that changes of the polysaccharide and protein content in EPS resembled each other.

Effect of the
The dissolved organics (protein and polysaccharide) in S-EPS increased before decreasing because TB-EPS were dissolved and transferred to S-EPS aer the sludge ocs were destructed by SO 4 c À and HOc generated in the KMnO 4 /PMS process. 29 Nevertheless, more SO 4 c À and HOc were generated to degrade the organics in S-EPS when PMS dosage continued to increase, so the protein and polysaccharide content in S-EPS tended to decrease. In addition, the slump of the protein and polysaccharide content in S-EPS, LB-EPS and TB-EPS accounted for SO 4 c À and HOc generation in the KMnO 4 /PMS process and its high reactivity toward S-EPS, LB-EPS, and even TB-EPS. 4 Fig. 5 Effect of the KMnO 4 /PMS process on particle size distribution of the WAS samples: (a) 2 mmol g À1 VSS KMnO 4 ; (b) 4 mmol g À1 VSS KMnO 4 ; (c) 6 mmol g À1 VSS KMnO 4 (experimental conditions:

Mechanism of WAS dewatering by the KMnO 4 /PMS process
Due to the versatility and efficiency, KMnO 4 was deemed as an important precursor for the fabrication of manganese oxide nanomaterials through chemical reduction, including manganese dioxides, tetraoxides and oxyhydroxides. 43  In order to verify the hypothesis, the formed solids were separated from the KMnO 4 /PMS process. And the samples were observed through TEM and EDX. From Fig. 7a-d, it could be clearly seen that some nanoscale solids did form in the KMnO 4 / PMS process. Moreover, the EDX analysis further conrmed that the formed solids were mainly composed of element Mn, O and K (Fig. 7e), and the weight percentages of Mn, O and K were recorded to be 20.9%, 74.26% and 4.82%. It suggested that KMnO 4 was disintegrate into MnO x under the inuence of EPS. According to the previous study, the solids might also be MnO x in the states of MnO 2 , Mn 3 O 4 and K x MnO 4 . 28 Meanwhile, scavenging experiments using ethanol (EtOH) and tert-butyl alcohol (TBA) as radical scavengers were conducted to conrm the formation of SO 4 c À and HOc in the KMnO 4 /PMS process. As can be seen in Fig. 8, aer extra addition of 300 mmol g À1 VSS EtOH and TBA, the CST decreased from 73.65 s to 38.50 s and 30.25 s, respectively while decreasing from 73.65 to 24.65 s without addition of any radical scavenger. Moreover, the reduction value of CST with the addition of EtOH was higher than that with the addition of TBA. Such difference in CST value drop by the two radical scavengers implied that both SO 4 c À and HOc were generated in the KMnO 4 /PMS process because EtOH can effectively quench both SO 4 c À and HOc while TBA reacts much faster with HOc than SO 4 c À . 25 Based on both the results above and current studies, a possible mechanism was proposed. As depicted in Fig. 9, Firstly, KMnO 4 transformed into MnO 2 , Mn 3 O 4 and K x MnO 4 aer oxidizing the organics in EPS. Then, the in situ generated MnO 2 and Mn 3 O 4 efficiently activated PMS to generate SO 4 c À and HOc (eqn (2)-(5)). In detail, HSO 5 À (the species of PMS) initially integrated with^Mn IV and^Mn III located at the surface of MnO 2 and Mn 3 O 4 by surface hydroxyl groups (eqn (6)). Then,^Mn IV and^Mn III would be reduced to^Mn III and Mn II , respectively with the generation of SO 5 c À (eqn (7)). Meanwhile, the formed^Mn III and^Mn II could be oxidized tô Mn IV and^Mn III with the formation of SO 4 c À (eqn (8) and (9)). Furthermore, HOc could also be produced through the reaction between SO 4 c À and H 2 O (eqn (5)). 46 Finally, SO 4 c À and HOc degraded the organics in EPS to break the sludge ocs so that huge quantities of bound water were released and became free water, thus signicantly enhancing sludge dewaterability.  FTIR spectrometer was also used to observed the variations of WAS functional groups aer conditioning in different conditions. As illustrated in Fig. 10, the broad absorption band from 3200 to 3400 cm À1 could be attributed to the stretching vibrations of O-H group of polymeric substances. 47 The peaks at 2925 and 2852 cm À1 were asymmetric and symmetric vibration of CH 2 of aliphatic structures and lipids, respectively. 48 The absorption band between 1640 to 1660 cm À1 was associated with the stretching vibrations of C]O and C-N, and C-N stretching vibration and N-H deformation vibration at 1550-1560 cm À1 . 49 The typical band located at 1040-1070 cm À1 was due to the stretching vibration of OH. 49 Besides, some peaks (<1000 cm À1 ) monitored in the "ngerprint" region could be attributed to the phosphate or sulfur functional groups, which is the functional groups for the production of nucleic acids. 48 Compared with conditioning by KMnO 4 or PMS alone, it can be clearly seen that the intensity of absorption bands of WAS decreased obviously aer treatment by the KMnO 4 /PMS process, implying that SO 4 c À and HOc generated in the KMnO 4 / PMS process could effectively degrade the proteins and polysaccharides in EPS and then resulted in the release of bound water and the improvement of sludge lterability.

Conclusion
This paper investigated the application of a combined use of KMnO 4 and PMS (KMnO 4 /PMS) for sludge dewatering. The major conclusions can be drawn as followed: (1) Compared with the separate use of KMnO 4 or PMS, the combined use of KMnO 4 and PMS (KMnO 4 /PMS) could signicantly improve sludge dewaterability. Aer conditioning with the KMnO 4 /PMS process, CST of the WAS samples decreased from 73.65 s to 24.65 s. However, the dewaterability of WAS would be worsened by conditioning with KMnO 4 alone or PMS alone because the CST value increased signicantly.
(2) The KMnO 4 /PMS process exerted signicant effects on the physicochemical characteristics of the WAS samples. When the WAS samples were treated by the KMnO 4 /PMS process, both CST and W C would decrease, the zeta potential was less negative, and the particle size would also decrease as the sludge ocs were broken into smaller fragments. Besides, the concentrations of protein and polysaccharide in EPS were also reduced due to the synergistic effect of KMnO 4 and PMS.
(3) The combined use of TEM and EDX techniques veried the generation of MnO x in the states of MnO 2 , Mn 3 O 4 and K x MnO 4 aer KMnO 4 oxidation. The radical scavenging experiments conrmed that the in situ generated MnO 2 and Mn 3 O 4 could activate PMS to generate SO 4 c À and HOc, which were responsible for signicantly enhanced sludge dewaterability. Moreover, the FTIR analysis further indicated that SO 4 c À and HOc generated in the KMnO 4 /PMS process could improve sludge dewaterability.

Conflicts of interest
There are no conicts to declare.