SrCo1−xTixO3−δ perovskites as excellent catalysts for fast degradation of water contaminants in neutral and alkaline solutions

Perovskite-like oxides SrCo1−xTixO3−δ (SCTx, x = 0.1, 0.2, 0.4, 0.6) were used as heterogeneous catalysts to activate peroxymonosulfate (PMS) for phenol degradation under a wide pH range, exhibiting more rapid phenol oxidation than Co3O4 and TiO2. The SCT0.4/PMS system produced a high activity at increased initial pH, achieving optimized performance at pH ≥ 7 in terms of total organic carbon removal, the minimum Co leaching and good catalytic stability. Kinetic studies showed that the phenol oxidation kinetics on SCT0.4/PMS system followed the pseudo-zero order kinetics and the rate on SCT0.4/PMS system decreased with increasing initial phenol concentration, decreased PMS amount, catalyst loading and solution temperature. Quenching tests using ethanol and tert-butyl alcohol demonstrated sulfate and hydroxyl radicals for phenol oxidation. This investigation suggested promising heterogeneous catalysts for organic oxidation with PMS, showing a breakthrough in the barriers of metal leaching, acidic pH, and low efficiency of heterogeneous catalysis.

0.4, and 0.6) catalysts calcined at 1000 °C for 6 h was obtained by powder X-ray diffraction (XRD) and their patterns are shown in Fig. 1(a). The characteristic peaks of all samples can be indexed according to the primitive cubic perovskite lattice with a space group Pm _ 3( )m (#221) 30 . In Fig. 1(a), SCT x have stronger intensity of diffraction peaks of perovskites than SCO, which presented higher stability of perovskite-type structure. Compared with the XRD patterns of SCO and TiO 2 (see Fig. 1(a)), SCT x (x = 0.1, 0.2, 0.4, 0.6), which exhibited the single phase, at varying ratio of Ti addition will not change the perovskite-type structure. The characteristic peaks of all perovskites overlap each other, suggesting almost identical lattice parameters for all the perovskites. With increasing Ti content, the most intense characteristic peak at around 2θ of 33° appears to shift slightly toward the lower angles. This indicates an expansion of SCT x lattice, due to the partial substitution of Co by Ti (Note that Co 3+ (VI) HS (high spin) has an ionic radius of 0.61 Å while Ti 4+ (VI) has an ionic radius of 0.75 Å) 37 .
The particulate morphology of SCT x (x = 0.1, 0.2, 0.4, and 0.6) was then carried out by SEM and the typical images are shown in Fig. 1(b). Ti-doped SrCoO 3−δ samples clearly experienced significant agglomeration during calcination at 1000 °C for 6 h, which had a large size of 1-5 μ m. An increase in Ti content (or equivalently, the increase in x) somewhat led to the loosening of grain agglomeration ( Fig. 1(b)). Relatively identical nano-sized grains of 10-20 nm are nonetheless observed for all four samples (See Fig. 1(b)). Nitrogen sorption furthermore reveals an increasing Brunauer-Emmett-Teller specific surface area (S BET ) with increasing Ti content, i.e., the surface area increased from 0.4 m 2 g −1 for x = 0 to 1.2 m 2 g −1 for x = 0.6 (Supplementary Table S1) that supports the SEM observation.
To explore the capability of perovskite oxides for removing phenol, the toxic organic with stable structure, adsorption and/or oxidation of phenol in the presence or absence of different oxide catalysts, with or without PMS were tested, as shown in Fig. 1(c). To provide baselines for the catalytic performance comparison, oxide catalyst self-adsorption and PMS self-oxidation were firstly checked. PMS clearly could not induce any noticeable phenol degradation in the absence of catalysts. In this case, less than 5% of phenol was removed during 90 min. Adsorption of phenol on Co 3 O 4 , TiO 2 , or SCT 0.1 was the only possible mechanism in the absence of PMS. Such adsorption may not translate to effective phenol removal given the low surface areas of these oxides. Accordingly, less than 5% of phenol was removed during the 90 min-period which implies negligible phenol adsorption on these oxides.
Oxidations occurred when both oxide catalysts (i.e., Co 3 O 4 , TiO 2 , or SCT 0.1 ) and PMS were present, in which phenol degradations become more pronounced. These oxidations are related to the activation of PMS to generate sulfate and hydroxyl radicals that can degrade phenol. SCT 0.1 /PMS system showed the largest catalytic activity, degrading the phenol in 15 min. For Co 3 O 4 and TiO 2 , however, only ~60% and ~10% of the phenol were degraded in 90 min, respectively. The catalytic performance thus increased in an order of TiO 2 < Co 3 O 4 < SCT 0.1 (see Fig. 1(c)). Moreover, the total organic carbon (TOC) analysis showed that about 78% of phenol could be Scientific RepoRts | 7:44215 | DOI: 10.1038/srep44215 mineralized to CO 2 after 2 h-oxidation using SCT 0.1 /PMS system. And compared with Co 3 O 4 (14.4 m 2 g −1 ) and TiO 2 (9.5 m 2 g −1 ), SCT 0.1 (0.4 m 2 g −1 ) has the lowest specific surface area (Supplementary Table S1), which all investigated that SCT 0.1 has the excellent catalytic performance for phenol oxidation with PMS activation.
The catalytic performances of SCT x (x = 0.1, 0.2, 0.4, and 0.6) was also investigated in the PMS activation for the oxidative degradation of phenol ( Fig. 1(d)). SCT 0.2 showed the highest catalytic degradation efficiency. Increasing surface area for the perovskite with increasing Ti content did not appear to well correlate with the catalytic improvement for phenol oxidation (Supplementary Table S1). The general trend ( Fig. 1(d)) indicates a decreasing rate of phenol oxidation with increasing titanium (Ti) content on the perovskite although SCT 0.6 can still be considered as a highly active catalyst given its capability to completely oxidize phenol within 45 min in the presence of PMS. The leaching of metal ions from the perovskite matrix to the solution was evaluated using ICP-AES. Co concentration in the filtrate varies from 3.1 to 5.4 mg L −1 (at pH of 3-4) which implies substantial leaching out of Co from SCT x in acidic condition. Co concentration appears to decrease with increasing Ti content in the perovskite. On the other hand, Ti leaching is not as obvious as Co considering its relatively lower concentration, i.e. below 1.0 mg L −1 in all perovskite cases. Therefore, it is a trade-off to obtain both maximized catalytic activity and minimized Co leaching. SCT 0.4 showed a similar Co concentration of 3.1 mg L −1 to SCT 0.6 , indicating either SCT 0.4 or SCT 0.6 could be chosen. The TOC analysis for SCT x (x = 0.1, 0.2, 0.4, and 0.6) showed the maximum TOC reduction on SCT 0.2 (82% TOC reduction) after 5 h phenol oxidation (Inset of Fig. 1(d)). SCT 0.4 could reduce about 77% of TOC over this duration. So given their identical Co leaching level and higher catalytic activity of SCT 0.4 , more tests were then focused on SCT 0.4 .
The effects of operating conditions on SCT 0.4 /PMS system and reusability of SCT 0.4. . As phenol was efficiently degraded by SCT 0.4 -activated PMS, we investigated SCT 0.4 /PMS system for phenol under different operating conditions, such as initial phenol concentration, PMS amount, catalyst loading and reaction temperature. Figure 2(a) displays the kinetic dependence of phenol oxidation on initial phenol concentration (C 0 ). The inset of Fig. 2(a) further shows the dependence of the reaction rate constant (k) for phenol oxidation on the initial phenol concentration (C 0 ). Phenol oxidation at 100% is attained within 30 min when the initial phenol concentration is less than 40 mg L −1 , while the oxidations at only 71% and 64% are attained for the initial phenol concentrations of 60 mg L −1 and 80 mg L −1 , respectively. Considering a fixed amount of catalyst, Fernandez et al. 21 . and Shulkla et al. 38 . provided a pseudo-zero-order kinetics of phenol removal, which is presented by the following equation: where k Co is the apparent zero order constant, C ph is the phenol concentration at any instant time (t), V is the volume of reactor, and W is the mass of catalyst. Therefore, the mathematical model of the degradation profile in Fig. 2(a,b,c,d) is induced by the integration of the above Equation (1). Shukla et al. also studied the kinetics of phenol oxidation by heterogeneous activation of PMS with different Co catalysts and reported that zero-order or first-order kinetics may occur depending on the catalysts used 24,25,38,39 . Table 1 shows the rate constants and the regression coefficients followed the pseudo-zero-order kinetics at four different initial phenol concentrations. The reaction rate decreased with the increasing initial phenol concentration. The reasonably good fittings are indicated by the high values of regression coefficients (R 2 > 0.97). The PMS amount also shows a significantly effect on kinetic phenol oxidation in Fig. 2(b). Higher amount of PMS led to a higher oxidation rate which translates to reduced time for complete oxidation of PMS. For example, when double amount of PMS was added, i.e., 1.5 g of PMS relative to 0.75 g of PMS, phenol oxidation could be completed within 10 min instead of 20 min.
Besides, the phenol oxidation kinetic dependence on the catalyst loading was probed as displayed in Fig. 2(c). Likewise, an increase in catalyst loading also correlates with the increase in the reaction rate. This is consistent with the expected increase of the active sites (Co) for phenol oxidation that enhances the generation of sulfate and hydroxyl radicals 24 .
Phenol degradation by SCT 0.4 -activated PMS was conducted at different reaction temperatures to further study the catalytic activity of SCT 0.4 for the activation of PMS. Figure 2(d) shows phenol oxidations at three  Table 2). The absence of perovskite data for PMS activation means that the activation energies for phenol oxidation on SCT 0.4 /PMS system should be compared to phenol oxidations on the other more widely studied catalyst systems ( Table 2). As we all know, one of critical factors influencing on the reaction activation energy is catalysts. Therefore, Table 2 showed that the phenol oxidation with different catalysts have different activation energies. Co-SiO 2 and Co-ZSM-5, in particular, showed the closest activation energies to SCT 0.4 /PMS system. Co/AC and Co/Fly-ash exhibited the lower activation energy. As a heterogeneous catalyst, the recyclability of SCT 0.4 is critical for practical implication. Figure 2(e) depicts the performance of the recycled SCT 0.4 catalyst for phenol degradation. The catalyst was recovered using filtration followed by mild washing with de-ionized water while the filtrate was stored for further analysis. The reused SCT 0.4 catalyst can still catalyze complete phenol oxidation within 60 min after three subsequent recycles. It showed more pronounced decrease in the activity after the first recycle followed by less decrease in activities in the subsequent recycles. TOC analysis showed that the TOCs were reduced by 76.2% and 75.3% during the first 2 h period of phenol oxidations in the first and second runs, respectively (see Supplementary Table S3). Increasing the reaction time up to 6 h can improve the TOC reduction to above 80% even after the second recycle (see Supplementary Table S3).
The leaching of Co 2+ from perovskite matrix was inevitable given the necessary long term contact and constant mixing between the perovskite catalysts and the aqueous phase. We also tested phenol oxidation using Co 2+ /PMS system in an analogous mole amount to SCT 0.4 as well as phenol oxidation using the filtrate from the previous SCT 0.4 /PMS system of the first run. The phenol oxidation on Co 2+ /PMS system was more rapid than the phenol oxidation on filtrate/PMS. This suggests that cobalt leaching did occur but it was at a much lower rate reflected by its low reaction rate.
We also did some characterizations of the reused SCT 0.4 to study its stability. Figure 2(f) shows the powder XRD patterns of the catalysts before and after the phenol oxidation for 2 h to check the possible changes in the phase compositions. The main perovskite phase of SCT 0.4 was retained although unknown phase(s) appeared afterwards, which are likely due to decomposition of the perovskite phase caused by metal leaching 40 and carbonate adsorption on the solid surface 34 .
The changes in the surface composition of the fresh and used SCT 0.4 catalysts were further evaluated using X-ray photoelectron spectroscopy (XPS). Figure 3(a) displays the survey scan on SCT 0.4 surface, evidencing the presence of Sr, Co, Ti, and O elements. The high resolution Sr 3d XPS spectra for the fresh and used SCT 0.4 catalysts (Fig. 3(b)) further reveal the formation of four peaks at 134.9 eV, 133.8 eV, 133.0 eV, and 131.8 eV. Sr 3d 3/2 components are located at higher binding energy (BE) while 3d 5/2 components are located at lower BE 41 . Accordingly, Sr 3d 5/2 peak at 131.8 eV and Sr 3d 3/2 peak at 133.0 eV correspond to Sr in the perovskite phase 42 . Sr 3d peak at approximately 133.8 eV, on the other hand, is likely related to the strontium carbonate on the surface 34,43,44 . Figure 3(c) further shows the high resolution Co 2p XPS spectra for the fresh and used SCT 0.4 catalysts. The Co 2p spectra contain 2p 1/2 and 2p 3/2 components. The peaks at 786.4 eV and 802.9 eV were the additional satellite peaks for Co 2p 1/2 and Co 2p 3/2 , respectively. The peak located at BE of 781.2 eV (Co 2p 3/2 ) represents Co 3+ while the peak located at BE of 780.1 eV (Co 2p3/2) represents Co 2+ , consistent with the literature 27,43,44 . The peaks corresponding to Co 3+ and Co 2+ can be deconvoluted to determine their peak areas. Accordingly, Co 3+ to Co 2+ ratio increased from 1.93 to 2.93 after the phenol oxidation for 2 h. This may indicate the activity of Co 2+ /Co 3+ redox couple (SCT 0.4 -Co(II)/SCT 0.4 -Co(III)) in catalyzing the phenol oxidation in the presence of PMS. However, the average cobalt oxidation states were calculated approximately to be 2.71 and 2.73 in the fresh and used SCT 0.4 , respectively. The results suggest that the cobalt oxidation states in the phase of perovskite were almost unchanged, which indicates the stability of SCT 0.4 in the catalytic process of phenol oxidation, and the metal leaching is likely to be mainly caused by the perovskite oxides dissolved in the acidic solution.
The high resolution Ti 2p XPS spectra for the fresh and used SCT 0.4 catalysts are depicted in Fig. 3 27,42,44 . Carbonates may be generated during the complex adsorption process of organic compounds or during the oxidation reactions that involves the formation of carbon-based intermediates 34 , likely corresponding to the unknown phase(s) in the XRD of Fig. 2(f). Therefore, the carbonate adsorption on the surface of SCT 0.4 after reaction may cause the deactivation of SCT 0.4 during the subsequent reuses.
The effects of initial pH on SCT 0.4 /PMS system. For the aqueous reaction of PMS activation, initial pH of reaction solution is an important factor to influence the interfacial interactions between PMS, organics and catalysts 45 and was further investigated. Figure 4(a) presents the phenol oxidation on SCT 0.4 /PMS system at different initial solution pH. The pH of the original phenol solution is around 6. We used 0.1 M H 2 SO 4 or 0.1 M KOH to adjust the initial solution pH. Relative to the neutral condition (pH of 7), higher acidity lowers the oxidation rate while higher alkalinity is in the opposite way. The effect of pH on the phenol degradation rate was attributed to the concentration change of the active radicals in the solution. When the pH is below 7, OH − ions react with H + ions in neutralization reaction to generate water. This essentially inhibits the two primary reactions that generate hydroxyl radicals (·OH, See Equations (5) and (6) which will be discussed below). Likewise, increasing the amount of OH − served to shift these reaction equilibria to the products, thus favoring the generation of hydroxyl radicals.
It is interesting that the TOC reduction (after 2 h) trend did not exactly follow the phenol oxidation rate trend. Maximum TOC reduction was observed at the neutral condition (pH of 7) (See inset of Fig. 4(a)). The fact that the TOC reduction for the initial pH = 9 was actually lower than that at the initial pH = 7 can be rationalized in terms of the derived phenoate (that was formed in an alkaline condition) which may slow down the reactivity of phenol and reaction intermediates in oxidation processes 46 .
The initial solution pH also showed a significant effect on metal ion leaching. The concentration of metal ions in the solution sharply decreased by increasing initial solution pH. Taking Co leaching as an example, with respect to the initial pH = 6 where no H 2 SO 4 or KOH was added, the initial pH = 6 resulted in the lower concentration of cobalt ions in the solution (C Co = 3.1 mg L −1 ). Increasing the pH further as represented by the initial pH = 9 led to an even lower concentration of cobalt ions in the solution (C Co = 2.2 mg L −1 ).
Maintaining the solution pH neutral (pH of 7) by constant pH monitoring and periodic addition of KOH) throughout the oxidation process minimized the Co leaching problem. Compared to the initial pH = 6, much lower concentration of cobalt ions in the solution was obtained (C Co = 1.3 mg L −1 ). This additionally resulted in a more rapid phenol oxidation than the initial pH = 9 (Compare Fig. 4(a) with Fig. 4(d) and note the different y-axis representations). By keeping the solution pH at 7 throughout the oxidation process, complete phenol oxidation can be achieved within 10 min as opposed to 20 min required for the initial pH = 9. The preservation of neutral pH during the oxidation course clearly optimized the phenol oxidation performance on SCT 0.4 /PMS system.

Catalytic mechanism.
Several studies have reported that PMS activation using metal-based catalysts involve of two key active radicals, i.e., sulfate radicals (SO 4 · − ) and hydroxyl radicals (·OH) 10,[47][48][49] . To reveal the phenol degradation mechanism in SCT 0.4 /PMS system, different radical inhibitors, ethanol (EtOH) and tert-butyl alcohol (TBA), were employed to evaluate the reaction. It is worth noting that EtOH (with α -H) readily reacts with both SO 4 · − and ·OH whereas TBA (without α -H) reacts mainly with ·OH and is inert to SO 4 40,50,51 . Figure 4(b,c,d) confirms that the phenol degradation rate was reduced with the presence of these radical inhibitors and their increasing concentrations. The presence of 0.2 M EtOH led to k of 0.012 mg L −1 min −1 , lower than k of 0.016 mg L −1 min −1 obtained in the presence of 0.2 M TBA in the system. This suggests that both SO 4 · − and ·OH were generated in the PMS activation process using SCT 0.4 , where normally more SO 4 · − than ·OH formed, as reported in several previous works 27,52,53 . Moreover, Fig. 4(d) reveals that upon maintaining neutral condition (pH = 7) throughout the oxidation course, more rapid phenol oxidation was obtained than that in the original oxidation case where the pH was not controlled. Relative to the original case, the phenol oxidation rate at the neutral condition increased by 1.95 times and 5.84 times for the tests with 0.2 M EtOH and 0.2 M TBA, respectively. This is likely due to the increased active radical species at such neutral condition. In the insert of Fig. 4(c,d), the presence of 0.2 M TBA at the neutral condition k shows more higher value than that without pH control, whereas the presence of 0.2 M EtOH keeps the similar phenol oxidation rate, which could be explained by more SO 4 · − generated than ·OH with pH control at 7, for that TBA reacts mainly with ·OH, but EtOH readily reacts with both them.

Conclusions
We have demonstrated that SrCo 1−x Ti x O 3−δ (x = 0.1, 0.2, 0.4, and 0.6) perovskite oxides exhibit a fast catalytic activity for phenol oxidation in the presence of PMS in a wide pH range. Powder XRD patterns showed cubic perovskite lattice and its stable structure in aqueous reaction. The catalytic performance for phenol oxidation presented an order of SCT 0.6 < SCT 0.4 < SCT 0.1 < SCT 0.2 . The operating conditions can be adjusted to increase the phenol oxidation rate by increasing PMS amount, catalyst loading, and initial pH. Under neutral and alkaline conditions, SrCo 1−x Ti x O 3−δ /PMS presented effective degradation efficiency of aqueous contaminants with less metal leaching and higher TOC reduction rate, which is better than Fenton-like systems and provides promising application for organic oxidation with sulfate radicals.