Introduction

Considering high toxicity of the inorganic arsenic species and serious health risk to human, pollution resulted from arsenic has become a major concern all over the world (Fendorf et al. 2010). Generally, the organoarsenic compounds, such as p-arsanilic acid (p-ASA) and roxarsone (ROX), are thought nontoxic and have been extensively used as feed additives in the poultry and pork industries because of their positive roles in improving feed efficiency, preventing dysentery, and controlling intestinal parasites (Chen and Huang 2012; Garbarino et al. 2003). Nevertheless, when the organoarsenic compounds were discharged into natural environment through the animal’s excrements, it can be decomposed to highly toxic arsenite (As(III)) and arsenate (As(V)) via both biotic and abiotic ways, leading to serious pollution in soil and water (Garbarino et al. 2003; Wang and Cheng 2015; Zhang et al. 2014). In addition, it has been reported that the poultry manure contained ROX concentrations of 14–54 mg kg−1 (Jackson and Bertsch 2001), and the total amounts of arsenic contained in the chicken and swine wastes were roughly 0.8 × 105–5.7 × 106 and 0.9 × 105–2.5 × 107 kg per year, respectively (Xie et al. 2016). At present, organoarsenic compounds have been banned in many countries and deemed as novel organic contaminants (Nachman et al. 2005). Thus, to effectively control their risk to the environment, it is urgent to develop some available treatments to deal with the contaminant.

So far, ample studies have reported water treatment processes for organoarsenic compounds through adsorption by carbon materials (Hu et al. 2012; Mahaninia and Wilson 2017), metal organic frameworks (MOFs) (Lv et al. 2018) and metal oxides (Joshi et al. 2017; Tian et al. 2017), and transformation by chemical and biological treatments (Xu et al. 2007; Xie et al. 2016; Zhang et al. 2014). Although the novel absorbents have exhibited excellent adsorption capacity of organoarsenic compounds, the cost of absorbents and waste management represented significant challenges to widespread implementation (Gupta et al. 2006). In addition, the microorganisms also showed the ability to decompose organoarsenic compounds, but it needed a long period and formed secondary arsenic pollution (Zhang et al. 2014). Besides the biological treatment, the advanced oxidation processes (AOPs) were documented treatment processes for the remediation of organoarsenic compounds, such as photocatalysis (Adak et al. 2015; Czaplicka et al. 2014; Zhu et al. 2014) and Fenton process (Hu et al. 2015; Liu et al. 2016; Xie et al. 2016). The reactive oxygen species (such as HO•, O2−•, and HO2•) generated in the AOPs could rapidly oxidize the organoarsenic compounds into small molecular substances and then mineralize them. Meanwhile, the released inorganic arsenic also could be oxidized to low-toxicity As(V). Among which, the Fenton process based on Fe was regarded as an ideal treatment for the organoarsenic contaminant considering the rapid oxidation efficiency of organoarsenic compounds by hydroxyl radical and the high removal efficiency of inorganic arsenic through the strong adsorption by the formed iron oxides or hydroxides (Xie et al. 2016).

Over the past few decades, the Fenton process was always one of the best options for the remediation of various recalcitrant organic pollutants in water. Due to the simpler operation and smaller amount of sludge than those of traditionally homogeneous Fenton process, the heterogeneous Fenton systems mediated by iron catalysts, such as zero valent iron (ZVI, Fe0) (Keenan and Sedlak 2008; Lv et al. 2016a, b), goethite (α-FeOOH) (Plata et al. 2010), Fe3O4 (Xue et al. 2009), and Fe0/Fe3O4 (Costa et al. 2008), have recently attracted a growing number of concerns. Recently, it has been noticed that the heterogeneous Fenton process mediated by nanoparticles is potentially useful for the remediation of persistent organic pollutants (POPs) due to their large specific surface area and high surface reactivity (Keenan and Sedlak 2008; Lv et al. 2016a, b; Zhang et al. 2009). In our previous studies, the heterogeneous Fenton system mediated by nZVI/Pd particles could rapidly degrade diphenyl ether without extra addition of hydrogen peroxide under room temperature and pressure conditions (Lv et al. 2016a, b). Thus, in the heterogeneous Fenton systems mediated by nZVI, the activities of nZVI particles played an extreme role. According to the reported studies (Xia et al. 2014; Zhuang et al. 2011), due to the strong magnetic effects between the nanoparticles, the ZVI nanoparticles tended to agglomerate, which significantly deadened the activities of nZVI. In addition, our previous work also found that the SiO2-coated nZVI/Pd exhibited lower aggregation between nanoparticles, lower bacterial toxicity, larger special surface area, and higher reactivity than bare nZVI/Pd particles, resulting in better performance of PBDE degradation (Lv et al. 2016a, b). Based on these findings, it can be firmly believed that the SiO2-coated nZVI also exhibits excellent performance in the heterogeneous Fenton systems.

In this work, a heterogeneous Fenton process mediated by the SiO2-coated nZVI (SiO2-nZVI) was employed to simultaneously remove the p-arsanilic acid (p-ASA, a typical organoarsenic compound) and the released arsenic. The removal performance at different pH values was firstly evaluated. In addition, the intermediates during p-ASA degradation were monitored by GC-MS and the degradation pathways of p-ASA by this treatment were explored. To better understand the removal mechanism of arsenic, the particles were characterized with XRD and XPS. Moreover, the effects of dissolved organic material (DOM) on p-ASA and arsenic removal were discussed.

Material and methods

Materials

Analytical grade para arsanilic acid (p-ASA), tetraethyl orthosilicate (TEOS), ferrous sulfate (FeSO4·7H2O, > 99%), and sodium borohydride (NaBH4, > 98%) were purchased from Aladdin Bio-Chem Technology Co., Ltd. (Shanghai, China). The humic acid (CAS no. 1415-93-6; molecular formula, C9H9NO6; molecular weight, 227.17) was purchased from Sigma-Aldrich (Germany). HPLC grade methanol and dichloromethane were separately supplied by Merck (Germany) and CNW (Germany). Water was double-distilled and then deionized with a Milli-Q water purification system (Millipore). The deoxygenated water was prepared by blowing high-purity nitrogen into the double-distilled water in a big bottle (2 L) for 2 h and then the bottle was kept in a glove box (LM1000S, Dellix Industry Co. Ltd., China).

Synthesis of SiO2-coated zero-valent iron nanoparticles

Synthesis of SiO2-coated zero-valent iron nanoparticles (SiO2-nZVI) was carried by reducing FeSO4·7H2O (4.000 g) with NaBH4 (0.4 g) in 200-mL degassed ultrapure water. Details were described in our previous reports (Lv et al. 2016a, b). The washed SiO2-nZVI particles were finally suspended in the deoxygenated water to form SiO2-nZVI slurry for use. Before use, the Fe0 in the slurry (9.201 g L−1) was determined by quantifying hydrogen gas evolved, using gas chromatography (Agilent 7890A GC, USA) with FID detector after digesting an aliquot of the SiO2-nZVI particles with 10 mL of H2SO4 (65 wt%) solution in a serum bottle overnight. The accurate concentration of Fe0 was calculated according to the amount of the generated H2 (55.845 g of Fe0 produced 1 mol of H2).

Degradation of p-ASA

The degradation of p-ASA was conducted under air atmosphere in a balloon flask (250 mL) with 300 mg L−1 of Fe0 and 10 mg L−1 of p-ASA. The detailed operations were as follows: 3.26 mL of SiO2-nZVI slurry and 145.24 mL of deionized water were separately added into the flask. After that, 1.5 mL of p-ASA stock solution (1000 mg L−1 in deionized water) was added into the system. The flask was conducted with mechanical stirrer at 200 rpm and 30 °C. At selected time intervals, 1 mL of liquid sample was taken out and then 10 μL of methanol was immediately added into the sample to stop the Fenton process. After filtering through 0.22-μm glass fiber filters (Tianjin Branch billion Lung Experimental Equipment Co., Ltd., China), the residual p-ASA concentrations were monitored. The system without SiO2-nZVI was set the control. All experiments were conducted in triplicate. To investigate the effects of initial pH on p-ASA degradation, 1 mol L−1 NaOH and 5 mol L−1 H2SO4 were used to adjust the pH values (2.0–7.0) in the system. The influence of dissolved organic matters (DOM) on p-ASA degradation was conducted in the presence of humic acid at different concentrations (0.5–5 mg C L−1).

Toxicity measurements

To evaluate the acute toxicity of the effluent, the Vibrio qinghaiensis sp.-Q67 (Beijing Hamamatsu Photon Technology Co. Ltd., China) was chosen as the test bacteria. Prior to toxicity assessment, the bacteria were reactivated in 1 mL 2.5% NaCl solution and stored at 4 °C in a refrigerator. Subsequently, V1 μL (0, 20, 50, and 100) of the effluent and p-ASA solution (10 mg L−1) (100-V1) μL of ultrapure water and 100-μL reactivated bacteria were separately added to the 96 microporous plate. After exposure to sample for 15 min at 25 ± 1 °C, the bioluminescence was measured by the ELIASA (TECANinfinite200, Switzerland). Toxicity was expressed as the bioluminescence inhibition ratio and it could be described as follows (I = the intensity of bioluminescence):20.

$$ \mathrm{Inhibition}\ \mathrm{ratio}=\left({I}_{\mathrm{control}}-{I}_{\mathrm{sample}}\right)/{I}_{\mathrm{control}}\times 100\%. $$

Analytical methods

The p-ASA concentrations were monitored by ultra-performance liquid chromatography (UPLC, Waters ACQUITY, USA). The organic intermediates during p-ASA degradation were determined by Thermo Trace GC Ultra instrument coupled to a Thermo DSQ II mass spectrometer (GC-MS, Thermo Electron Corporation, Waltham, USA). The concentrations of ammonium ion (NH4+) were monitored by an ion-chromatography system (ICS-90, Dionex). The arsenic species during the Fenton process were determined by 7700x inductively coupled plasma mass spectrometer (ICP-MS) hyphenated with a 1200 high-performance liquid chromatograph. The quantification of the arsenic species was conducted at m/z 75 (Liu et al. 2013). The residual arsenic concentration in the effluent was analyzed by ICP-MS (XSERIES 2, Thermo). The concentrations of dissolved Fe(II) and total dissolved iron were determined with a modified ferrozine method (Voelker and Sulzberger 1996). Total dissolved iron was quantified after adding hydroxylamine hydrochloride (final concentration = 60 mM). TOC was determined by LiquiTOC trace (Elementar, Germany). Before determination, all the samples were filtered with 0.22-μm glass fiber filters. Hydroxyl radical was tested with electron paramagnetic resonance (EPR) spectroscopy (ELEXSYS-II, Bruker) by choosing DMPO as a scavenger. The 2,5-dihydroxybenzoic acid (2,5-DHBA) concentrations (proportional to the produced hydroxyl radical in the Fenton reaction) were detected with high-performance liquid chromatography (Waters 2695) equipped with a florescence detector (Waters 2475) (Buxton 1988). X-ray diffraction (XRD) spectrum for SiO2-nZVI particles was obtained using Cu Ka radiation at 40 kV and 40 mA (MAC Science Co., M18XHF). X-ray photoelectron spectroscopy (XPS) data was recorded with the Al Ka line at 15 kV and 51 W (PHI X-tool, DE). The binding energies were determined by reference to the C1s component due to carbon being bound only to carbon or hydrogen, set at 284.8 eV. (The detailed information for the abovementioned method is provided in the Supporting Information.)

Statistical analysis

All experiments were performed in triplicate and the data were expressed as mean ± standard deviation. The statistical analyses were performed with one-way analysis of variance (ANOVA) using the Minitab 16 software. A multiple comparison Tukey test was applied to assay the differences among treatments.

Results and discussion

p-ASA degradation in the heterogeneous Fenton mediated by SiO2-nZVI and nZVI

To confirm the role of SiO2-nZVI in p-ASA remediation, the p-ASA degradation in the heterogeneous Fenton mediated by SiO2-nZVI and nZVI and the homogenous Fe2+/O2 system were studied (Fig. S1). In the control system (Fig. S1a), no evident changes in p-ASA concentration were observed, suggesting that the possible loss during the process was basically negligible. Similarly, no significant change of p-ASA was observed in the Fe2+/O2 system (Fig. S1a), which suggested that the homogenous Fe2+/O2 system could not generate the reactive oxygen species to degrade the pollutants in the absence of H2O2. Instead, the rapid decline of p-ASA was observed both in nZVI and in SiO2-nZVI systems, implying that the heterogeneous Fenton systems could effectively eliminate p-ASA (Fig. S1a). In addition, the p-ASA degradation percentage (99.6 ± 3.2%) in SiO2-nZVI system was much higher than that in nZVI system (88.3 ± 3.0%), indicating that the SiO2-nZVI exhibited higher activity than nZVI. The similar results also occurred to the apparent rate constant (Fig. S1a inset). According to our previous study, after SiO2 modification, lower aggregation, smaller size, and larger special surface area occurred to ZVI nanoparticles (Lv et al. 2016a, b), which could greatly improve their activities in the two-electron transfer process (Keenan and Sedlak 2008), thus resulting in the enhancement in the production of hydroxyl radicals (Fig. S1b). Furthermore, the SiO2-nZVI dosage also displayed significant influence on p-ASA degradation (Fig. S1c). With an increase of the SiO2-nZVI dosage, the p-ASA degradation greatly rose and reached 99.6% at 300 mg L−1 of Fe0 dosage. Therefore, the Fe0 dosage in SiO2-nZVI was set as 300 mg L−1 in the following experiments.

Effects of pH on p-ASA removal

Since the pH value plays a significant influence on the oxidization efficiency of Fenton process (Lv et al. 2016a, b), the performances of p-ASA removal at different pH values (from 2.0 to 7.0) were investigated and the results were shown in Fig. 1. As depicted in Fig. 1a, in the SiO2-nZVI-O2 system, a noteworthy decline of p-ASA over time was observed within 60 min, but the changes greatly differed under different pH values. The lowest removal efficiency (22.9%) of p-ASA occurred at pH 2.0, while the highest performance (99.6%) was found at pH 3.0. When the pH value exceeded 3.0, the final removal efficiency seemed to have no obvious changes (82.1–87.8%) with an increasing pH.

Fig. 1
figure 1

Effects of initial solution pH on p-ASA removal (a), degradation (b), and adsorption (c) in the heterogeneous Fenton system. d The amount and percentage of p-ASA removal via degradation and adsorption after 60 min, and the variation of pH value (e), oxidation-reduction potential (OPR) (f), the released Fe ions (g), and yielded 2,5-DHBA (h) during the whole heterogeneous Fenton reaction

Full size image

Actually, the removed p-ASA involved two portions, the p-ASA degraded by hydroxyl radical and the p-ASA adsorbed on SiO2-nZVI particles. As shown in Fig. 1b, the degradation of p-ASA greatly differed from p-ASA removal at varied pH values. Similarly, the highest (99.5%) and lowest (9.3%) degradation efficiencies also occurred at 3.0 and 2.0, respectively. However, once the pH was over 3.0, the degradation efficiency dramatically descended as the pH rises, which ranged from 2.72 to 7.93 mg L−1. Besides degradation, adsorption also played a vital position in p-ASA removal (Fig. 1c). Except the curve at pH 3.0, all the adsorption curves rose and then tended to be steady as time is over. After 60 min, the lowest equilibrium adsorption capacity (0.0015 mg) happened at pH 3.0 and the highest one (0.92 mg) was obtained when the pH was 7.0.

Based on the above results, the amount and percentage of p-ASA removal via degradation and adsorption at different pH values were illustrated in Fig. 1d. In the system at pH 2.0, the amount of p-ASA removal was lowest and the percentages of degradation and adsorption were 41% and 59%, respectively. Instead, at pH 3.0, the removal of p-ASA was fully ascribed into the degradation. At pH 4.0–7.0, the percentage of degradation gradually declined with an increase of pH values (from 83 to 25%).

Meanwhile, the close inspection of the variation of the pH values (Fig. 1e) in different systems manifested that the heterogeneous Fenton process induced by the SiO2-nZVI particles extremely depended on the hydrogen ion. The pH rose rapidly in the first 25 min and then tended to be stable. After 60 min, the final pH values of different systems were 2.64, 5.99, 6.03, 6.02, 6.28, and 7.19, respectively. In addition, the monitored ORP values (Fig. 1f) at different pHs also sharply ascended into the peak in the beginning 10 min and then gradually declined and finally became stable after 40 min. In which, the ORP values at pH 3.0 were much higher than those at other pHs, indicating the strongest oxidizing ability in the system. Accordingly, a distinct release of Fe ions was observed during the process (Fig. 1g). The Fe2+ concentration first increased and then declined, while the total dissolved iron ascended throughout the whole process. Due to the low pH (< 3.00) and the presence of nZVI particles, there was almost solely Fe2+ in the system with the initial pH of 2.0. In other systems, with the consumption of H+, Fe2+ was gradually oxidized to Fe3+, leading to a decline.

It has been reported in previous literatures that hydroxyl radicals could be generated via two-electron transfer during the oxidization of ZVI in the presence of oxygen (Keenan and Sedlak 2008; Li and Zhu 2014; Lv et al. 2016a, b). Herein, salicylic acid was chosen as the target to quantify the hydroxyl radicals at different pHs. The presence of 2,5-DHBA at 3.92 min (Fig. S2) and EPR results (Fig. S1b) convincingly testified the generation of hydroxyl radicals. Figure 1h clearly elucidated the production of 2,5-DHBA detected in the system throughout the reaction time. The yields of 2,5-DHBA increased rapidly in the first 20 min, suggesting high levels of hydroxyl radicals in the system. Subsequently, the yields gradually became steady after 60 min. At the end of the system at pH 3.0, the 2,5-DHBA concentration was 91.7 ± 1.6 μmol L−1, which was higher than that in other systems. Nevertheless, the actual amount of hydroxyl radicals was much more than that measured in the system because hydroxyl radicals also would react with the nZVI particles and Fe2+ during the process (Lv et al. 2016a, b).

All the results suggested that the initial pH value vastly acted on the removal of p-ASA, especially the degradation process, which was similar to Xie et al.’s findings (Xie et al. 2016). Appropriate acidity (pH 3–5) was beneficial to the degradation of p-ASA in the heterogeneous Fenton reaction, but in highly acidic (pH = 2.0) or weakly acidic (pH > 5.0) system, the adsorption might play a dominate role during the removal of p-ASA. In the nZVI-O2 process, the solution pH greatly affected the release of Fe2+, which was significant for the production of HO• (Keenan and Sedlak 2008). In the highly acidic system, the excess H+ quickly dissolved a large amount of nZVI particles, which led to a low generation of H2O2 via two-electron transfer reaction (Fe0 + O2 + 2H+ → H2O2 + Fe2+) (Keenan and Sedlak 2008), thereby resulting in less amount of HO• (Fig. 1h). In addition, much more HO• could be scavenged by H+ at the lower solution pH (Pignatello 1992), which also caused the low degradation efficiency of p-ASA. Furthermore, the low level of SiO2-nZVI particles caused low adsorption capacity of p-ASA. Under weak acidic condition, due to the low level of the released Fe2+ (Fig. 1g), the weak oxidation capacity resulted in a low degradation efficiency of p-ASA (Xie et al. 2016). However, considering the strong coordination between Fe and As (Jin et al. 2012; Tian et al. 2017), the high level of residual SiO2-nZVI particles exhibited excellent adsorption capacity on p-ASA (Fig. 1c).

Identification of the intermediates and the pathway of p-ASA during the process

The organic degradation intermediates and inorganic species, as well as the variation of TOC, were monitored during the heterogeneous Fenton process at pH 3.0 (Fig. 2). Despite the fact that almost all the p-ASA was eliminated after 60 min (Fig. 2b), the dissatisfactory mineralization of p-ASA was observed through the TOC test (21.9%), indicating the accumulation of some organic intermediates.

Fig. 2
figure 2

Organic and inorganic intermediates and their variation during p-ASA degradation: a the organic intermediates monitored with GC-MS after 20 min, b the variation of the released NH4+ and TOC during the reaction, and c the variation of the released As(V) and residual As(V) during the reaction

Full size image

Based on the GC-MS analysis, approximately eight types of possible byproducts were identified through the National Institute of Standards and Technology (NIST11) library. The detailed intermediates were presented in Fig. 2a, Fig. S3 and Table S1. As shown in Fig. 2a, peak 1 at retention time 8.96 min and peak 4 (13.69 min) were confirmed as aniline (molecular ion at m/z 93 and major fragment ions at m/z 66, 65, 39, and 28) and p-aminophenol (molecular ion at m/z 109 and major fragment ions were at m/z 80, 53, and 18), respectively, which were generated via the cleavage of C–As bond (Xu et al. 2007; Xie et al. 2016). Peak 2 at 9.15 min and peak 5 at 13.88 min were identified as phenol (molecular ion at m/z 94 and major fragment ions at m/z 66 and 65) and p-hydroquinone (molecular ion at m/z 110 and major fragment ions at m/z 81, 64, and 55), which were separately generated from aniline and p-aminophenol through the replacing of -NH2 by hydroxyl (Anotai et al. 2006; Fukushima et al. 2000). The p-benzoquinone (peak 3, with molecular ions at m/z 108) which appeared at 13.40 min might be attributed to the oxidation of p-hydroquinone by hydroxyl radical, which has been verified in our previous findings (Lv et al. 2016a, b). The peaks (6–8) with retention time at 18.20, 18.43, and 20.96 min were a series of ring-opening products, which were identified as fumaric acid, maleic acid, and trans,trans-2,4-hexadienedioic acid, respectively.

Besides the organic intermediates mentioned above, the release of As(V) and NH4+ was also observed along with the decline of p-ASA, suggesting the rupture of C–N and C–As bonds from the aromatic ring of p-ASA. Similar results were also observed during oxidation of aniline and organoarsenic compounds by Fenton process (Briviba et al. 1993; Czaplicka et al. 2014; Fukushima et al. 2000; Xie et al. 2016). Due to the high electron density on para-position of the aromatic ring, the -NH2 group of aniline was preferentially attacked by the reactive oxygen species such as HO• and 1O2, resulting in the generation of NH4+ and p-aminophenol (Fukushima et al. 2000). As depicted in Fig. 2b, no NO3− or NO2− was detected, while the concentration of NH4+ first increased and then stabilized at 0.62 ± 0.02 mg L−1 after 60 min, which was much lower than the theoretical NH4+ (0.83 mg L−1). This could be explained by two reasons: on the one hand, the accumulation of p-aminophenol in the system would cause the decline of NH4+. On the other hand, due to the negative charge on the surface of SiO2, the adsorption also could result in a decrease. Meanwhile, the obvious release of As species (adding H2SO4 to dissolve the nZVI particles) was detected with HPLC-ICP-MS (Figs. S4 and S5). During the whole process, only arsenate was monitored, manifesting that complete oxidation of As(III) under strongly oxidizing systems (Xie et al. 2016; Zhu et al. 2014). After 60 min, the concentration of As(V) was 3.32 ± 0.20 mg L−1, which was a little different from the theoretical As(V) (3.44 mg L−1). Actually, the residual As(V) in the effluent was much lower than the released As(V), and finally stabilized at 0.031 mg L−1, meeting both the drinking water standard of the World Health Organization (WHO) and the surface water standard of China (50 μg L−1), respectively. Due to the strong coordination between Fe and As (Gupta et al. 2009; Tian et al. 2017; Zhang et al. 2010), the iron particles in the system could effectively adsorb the arsenate ion released during the degradation of p-ASA.

Based on the intermediates and the general mechanisms involved in heterogeneous Fenton oxidation (Keenan and Sedlak 2008; Lv et al. 2016a, b), the degradation pathway was proposed in Fig. 3. The p-ASA oxidization by a heterogeneous Fenton process mainly involved three stages. The first step was the dearsenization from the aromatic ring of p-ASA. The hydroxyl radicals first attacked the C-As bond, leading to the formation of aniline and arsenite, and then immediately oxidized arsenite to arsenate. The second stage was deamination from the aromatic ring. During this stage, with the further assault by the hydroxyl radicals, the C–N bond broke down and formed phenolic compounds (phenol and p-hydroquinone) and NH4+. At the same time, the p-hydroquinone could be oxidized to p-benzoquinone, which has been verified in our previous investigations (Lv et al. 2016a, b). Finally, phenolic compounds were further converted to some small organic molecules through ring opening, such as fumaric acid, maleic acid, and trans,trans-2,4-yexadienedioic acid, and eventually mineralized to CO2 and H2O, which was confirmed by the decline of TOC (Fig. 2b).

Fig. 3
figure 3

Proposed degradation pathways of p-ASA in the SiO2-nZVI-O2 system

Full size image

Toxicity of p-ASA and its degradation products in treatment process

The toxicities of p-ASA and its degradation compounds in the effluent during heterogeneous Fenton process were evaluated by luminescent bacteria test and expressed in bioluminescence inhibition ratio. As shown in Fig. 4, in the control set, the inhibition to bioluminescence could be negligible. In the p-ASA groups, the bioluminescence inhibition ratio significantly increased from 17.3 ± 0.7 to 48.1 ± 2.5% along with an increase of the p-ASA concentration, indicating that the p-ASA exhibited evident toxicity to the bacteria. Due to the presence of the arsenic acid groups, the phenylarsonic acid could effectively inhibit the activities of the microorganisms in animal intestines when it was added into the fodder (Akhtar et al. 1992; Hanson et al. 1955; Howie 2005). Thus, the p-ASA could inhibit the bioluminescence of the bacteria. Nevertheless, in the effluent groups, the bioluminescence inhibition ratio decreased from 2.5 ± 1.2 to − 12.3 ± 3.0%, suggesting that no toxicity was detected in the effluent. According to the results of GC-MS, IC, and ICP-MS (Fig. 2), there are some ammonium ion and abundant organic compounds with low molecular weight in the effluent. These degradation compounds could be used as nitrogen and carbon nutrient supplementations (Lv et al. 2014) and then promote the bacteria’s growth, resulting in the enhancement of the bioluminescence. Therefore, the heterogeneous Fenton process mediated by SiO2-nZVI was feasible and meaningful for the treatment of p-ASA.

Fig. 4
figure 4

Acute toxicity to Vibrio qinghaiensis sp.-Q67 of the samples before and after the heterogeneous Fenton treatment

Full size image

Characterization of SiO2-nZVI particles after p-ASA oxidization

During p-ASA oxidization, it could be clearly observed that the black slurry gradually changed into yellowish-brown suspending liquid, indicating that some reaction occurred on the surface of the metallic surface. In order to analyze the components of the resultant SiO2-nZVI particles, the solid residues were characterized by XRD and XPS.

Figure 5a depicted the XRD patterns of the fresh SiO2-nZVI particles and the residual SiO2-nZVI composites after the heterogeneous Fenton reaction. After 60-min oxidation, it was visibly observed that the characteristic diffraction peaks of Fe0 (2θ = 44.7° (110), 65.0° (200), and 82.4° (211)) completely disappeared. Instead, a host of new diffraction peaks distinctly generated, indicating the generation of mixed iron oxide phases. The peaks at 2θ = 21.2°, 27.1°, 47.0°, and 64.1° were assigned to the characteristic diffraction peaks of lepidocrocite (L, JCPDS card no. 06-0696). Meanwhile, the magnetite (M, JCPDS card no. 19-0629) was also confirmed by the characteristic diffraction peaks at 2θ = 18.3°, 30.1°, 35.6°, 53.4°, 57.0°, and 75.0°. Therefore, the residual SiO2-nZVI composites after oxidation mainly consisted of lepidocrocite and magnetite. According to previous reports (Faria et al. 2014; Tian et al. 2017; Wang et al. 2016), both FeOOH (lepidocrocite) and Fe3O4 (magnetite) exhibited excellent capacity to adsorb the As species, such as arsenite, arsenate, and organoarsenic compounds. Thus, during the heterogeneous Fenton reaction, the released arsenate could effectively be removed by the iron oxides or hydroxides formed on the surface of SiO2-nZVI particles, resulting in the low level of As(V) in the effluent (Fig. 2b).

Fig. 5
figure 5

XRD patterns and XPS spectra of fresh SiO2-nZVI particles and the SiO2-nZVI oxide composites: a XRD patterns SiO2, fresh SiO2-nZVI particles and treated SiO2-nZVI particles after reaction, b Si2p and c As3d core levels of treated SiO2-nZVI particles, d Fe2p, e Fe3p, and f O1s core levels of the fresh and treated SiO2-nZVI particles

Full size image

In addition, X-ray photoelectron spectroscopy (XPS) was applied to further explore the surface chemical components of the treated SiO2-nZVI composites after the heterogeneous Fenton reaction. As displayed in the XPS survey scans (Fig. S6), the major elements on the surface of the SiO2-nZVI particles involved Fe, Si, C, and O, among which, the C was ascribed into the calibration of binding energies, while Si existed in the form of SiO2 considering the binding energy of high-resolution Si2p (103.4 eV, Fig. 5b). Besides those elements, a weak characteristic As peak (As3d) was evident in the XPS survey spectrum of SiO2-nZVI after reaction, which was attributed to the adsorption of released As species. Moreover, the high-resolution As3d (45.1 eV) spectrum (Fig. 5c) confirmed the presence of As(V) (Joshi et al. 2017), suggesting the generation of arsenate during the reaction, which agreed with the results in Fig. 2c.

Fe2p3/2 and Fe2p1/2 spectra of the Fe2p doublet of both samples (Fig. 5d) turned up at binding energies of 710.9 and 724.9 eV, with a slight shake-up satellite at 718.2 eV. The two feature peaks and the observation of satellite lines, which were the fingerprint of Fe3+ ions, suggested that some ferric oxides and hydroxides precipitated on the surfaces of the treated SiO2-nZVI particles. According to Segura et al.’s findings (Segura et al. 2015), the peak at ~ 711 eV manifested the potential presence of Fe2O3, Fe3O4, Fe (OH)3, or FeOOH, while the peak at ~ 725 eV suggested the presence of ferric oxides [Fe(III)]. Due to the extremely strong reducibility, the nZVI particles were coated by a layer of iron oxides once exposed to the air atmosphere, which generally resulted in a weak peak of zero-valent iron (706.5 eV of binding energy) (Liu et al. 2014). To further determine the iron species on the surface of both samples, the high-resolution Fe3p core levels (Fig. 5e) were deconvoluted into the Fe3+ (bind energy of 55.7 eV) and Fe2+ (bind energy of 53.9 eV) peaks, manifesting that besides ferric oxides and hydroxides, there are still some ferrous oxides (FeO or Fe3O4) on the surface of the samples (Yamashita and Hayes 2008).

The oxygen species surrounding iron ions were obtained from the high-resolution O1s scans. As shown in Fig. 5f, the high-resolution O1s scans of fresh SiO2-nZVI exhibited the only one fitted peak at 529.6 eV, which was assigned to O2− (Segura et al. 2015). Nevertheless, besides O2− peak, another peak, assigned to typical -OH species turned up at 531.3 eV, which suggested that both O2− and -OH species were present surrounding the iron ions after reaction. Therefore, according to the results of XRD patterns and XPS spectra, after the heterogeneous Fenton process, the residual SiO2-nZVI composites were mainly composed of FeOOH, Fe3O4, and Fe2O3. The generated the iron oxides and hydroxides on the surface of SiO2-nZVI severely hindered the Fenton reaction and then led to a decline of the degradation efficiency of pollutants (Keenan and Sedlak 2008; Lv et al. 2016a, b), but played a significant role in the adsorption of the released arsenate species via coordination interactions (Gupta et al. 2009; Hu et al. 2015; Liu et al. 2016; Wu et al. 2017; Zhang et al. 2010), which could effectively avoid the severe secondary pollution and environmental risk caused by As species.

Based on the XRD and XPS results, it could be seen that after reaction, some iron oxides and hydroxides were coated on the surface of the particles. At this time, the SiO2-nZVI particles were passivized after reaction. Thus, we concluded that the treated SiO2-nZVI particles could not in situ mediate the heterogeneous Fenton process and believed that the treated SiO2-nZVI particles could not be reused. To confirm the effects of the coated iron oxides and hydroxides on the reuse of SiO2-nZVI particles, the reuse experiments were carried out and the results were showed in Fig. S7. Compared with the fresh SiO2-nZVI particles, the recycled SiO2-nZVI particles exhibited rather low performance on p-ASA degradation. The p-ASA degradation efficiencies for the first reuse and second reuse were 48.7 ± 2.8% and 26.4 ± 3.4%, respectively, which were only about one-half and one-fourth larger than those of fresh SiO2-nZVI particles. However, when the recycled SiO2-nZVI particles were treated with acid (1 mol L−1 of HCl solution), the p-ASA degradation efficiencies increased evidently (74.7 ± 2.4% and 45.6 ± 3.7% for the first and second reuse), meaning that the surface passivation could severely inactivate the reaction of the SiO2-nZVI particles. Meanwhile, the recycled SiO2-nZVI particles treated with acid still showed much lower p-ASA degradation efficiency than fresh ones, which was attributed to the low amount of Fe0 in the samples. Based on these results, after reaction, the passivation on the surface and the consumption of the Fe0 both would significantly reduce the activity of the recycled SiO2-nZVI particles, resulting in the remarkable decline on the p-ASA degradation efficiency.

Performance of heterogeneous Fenton process in simulated natural waters

Effects of the anions on the p-ASA degradation

Considering the widespread presence of some anions in the real wastewater, their effects on p-ASA degradation were investigated. As depicted in Fig. 6, compared with control set, the anions (sulfates, chlorides, nitrates, carbonates, and phosphates) displayed a different influence on p-ASA degradation. Given that no evident change in p-ASA degradation efficiency was observed in the groups of sulfates, nitrates, and phosphates, it could be concluded that the sulfates, nitrates, and phosphates had no impacts on p-ASA degradation. Instead, the chlorides and carbonates exhibited significant inhibition to the p-ASA degradation and the inhibition enhanced along with an increase of the concentration. The p-ASA degradation efficiencies in the chloride groups (76.4 ± 2.4% and 45.2 ± 1.8% for 10 and 20 mg L−1, respectively) were higher than those in the carbonate groups (67.6 ± 2.7% and 31.4 ± 3.7% for 10 and 20 mg L−1, respectively), meaning that the carbonates showed a stronger inhibition to p-ASA degradation than chlorides. Due to the reducibility, the chloride ion was reported to inhibit the decomposition of dichlorvos (Lu et al. 1997) and 2,4-dichlorophenol (Tang and Huang 1996) in the Fenton process. Similarly, the bicarbonate ions also played positive role in scavenging the hydroxyl radicals (Liao et al. 2001). Furthermore, according to Grebel et al.’s (2010) study, in the UV/H2O2 system, the carbonates exhibited greater impact on organic contaminant destruction rate suppression than chloride ion.

Fig. 6
figure 6

Effects of the anions on the p-ASA degradation. Experimental conditions: [Fe0]0 = 300 mg L−1, [p-ASA]0 = 10 mg L−1, and pHinitial = 3.0

Full size image

Effects of DOM on the p-ASA degradation

DOM in natural waters is an excellent radical scavenger and can consume the reactive oxygen species (ROS) generated in advanced oxidation processes (AOPs), leading to a decline of the oxidation efficiency of the targeted pollutants (Matilainen and Sillanpää 2010). Meanwhile, DOM can also occupy the adsorption sites on the surface of the sorbents for the targeted pollutants, resulting in a decreasing removal of the targeted contaminants (Grafe et al. 2001, 2002; Li et al. 2015; Redman et al. 2002). Therefore, the effects of DOM on p-ASA degradation and arsenic adsorption were investigated and the results were depicted in Fig. 7. At low levels of DOM (< 1.0 mg C L−1), the degradation efficiencies of p-ASA (96.6% and 96.6% for 0.5 and 1.0 mg C L−1, respectively) were comparable to that in pure water, and the residual As concentrations (0.031 to 0.0049 mg L−1) in these systems were also below the limit of the drinking water standard of the WHO (0.05 mg L−1), suggesting that DOM in low level played a negligible role in the treatment of p-ASA in the SiO2-nZVI-O2 system. Instead, once the DOM level exceeded 2 mg C L−1, it could be observed that the presence of high-level humic acid significantly reduced the degradation of p-ASA and the removal of arsenic. The degradation efficiency of p-ASA descended from 99.5% in pure water to 44.6% in the system with 5 mg C L−1 of DOM, while the residual arsenic concentration in the effluent ascended from 0.031 to 0.712 mg L−1. To eliminate the inhibition of high DOM on p-ASA oxidation, the dosage of SiO2-nZVI was increased and extra H2O2 was added. The results (Fig. S8) showed that even at high level of DOM (5 mg C L−1), high efficiencies in p-ASA degradation (> 98%) and arsenic removal (< 0.05 mg L−1) in the simulated natural waters could still be obtained with the dosages of Fe0 increased from 300 to 500 mg L−1 or adding 20 mmol L−1 of H2O2. Similarly, Xie et al. (2016) also found that high-level DOM greatly impeded the oxidation of p-ASA in the Fenton reaction, but either increasing the dosage of Fenton’s reagent or extending the reaction time could effectively overcome the interference from high-level DOM. Thus, in this heterogeneous Fenton process, despite the fact that high DOM exhibited inhibition on p-ASA oxidation and arsenic removal, the interference could be eliminated by increasing the dosage of SiO2-nZVI or adding extra oxidants such as H2O2, O3, and peroxysulfate.

Fig. 7
figure 7

Effects of DOM concentrations (0, 0.5, 1, 2, 3, and 5 mg C L−1) on the degradation of p-ASA and the removal of arsenic during the heterogeneous Fenton process after 60 min. Experimental conditions: [Fe0]0 = 300 mg L−1, [p-ASA]0 = 10 mg L−1, and pHinitial = 3.0

Full size image

Conclusions

The heterogeneous Fenton process mediated by SiO2-nZVI could be a promising treatment for complete degradation of the emerging organoarsenic contaminants with simultaneous arsenic removal in the pure water and natural water. With the attack of hydroxyl radical, p-ASA could be oxidized to arsenate, ammonium ion, and plentiful phenolic compounds via cleavage of C–As and C–N bonds. Meanwhile, the released arsenate could be effectively removed through the adsorption on the formed iron oxides and hydroxides and the residual arsenic met both the drinking water standard of the World Health Organization (WHO) and the surface water standard of China. The results consistently indicated that heterogeneous Fenton process is promising as a simple and effective treatment for the remediation of p-ASA-containing wastewaters and reducing the risk of organoarsenic feed additives in swine wastewater.