The catalytic ozonation of diclofenac (DCF) with iron silicate-loaded pumice (FSO/PMC) in aqueous solution was investigated. FSO/PMC was synthesized by a co-precipitation–impregnation method and characterized using scanning electron microscope, N2 adsorption–desorption, X-ray fluorescence, and pHpzc measurements. Results showed that the FSO/PMC/O3 process obviously improved total organic carbon (TOC) removal efficiency from 32.3% (using sole ozonation) to 73.3% in 60 min. DCF mineralization in various oxidation processes was found to follow a two-stage pseudo-first-order kinetics. The presence of FSO/PMC effectively improved the mass transfer of ozone from gas to liquid phase and increased the efficiency of ozone decomposition, which results in the formation of •OH radicals. The ozonation of DCF generated large amounts of the ozone-refractory carboxylic acids, and these compounds were found to be continuously removed in the FSO/PMC/O3 process due to the catalytic activity of FSO/PMC. The synergetic effect between ozonation and FSO/PMC adsorption indicated that FSO/PMC is a promising catalyst for the ozonation process.
Pharmaceutical products are widely used for the treatment or prevention of human and animal diseases. However, most of these compounds are excreted through urine and feces, and easily enter into the sewage collection system after consumption. The traditional wastewater treatment plants are in many cases unable to effectively remove these compounds, so natural water bodies (i.e. surface water, ground water and sediment) are frequently detected with these compounds around the world at concentrations up to ng/L or even μg/L levels (Vieno et al. 2007). Among these pharmaceutical compounds, diclofenac (DCF) is one of the most frequently detected pharmaceuticals in aquatic environments. It is usually used to reduce inflammation in arthritis and treat painful diseases of rheumatic origin as a non-steroidal anti-inflammatory drug (Sari et al. 2014). Recent publications have demonstrated that the presence of DCF has some adverse effects on aquatic and terrestrial life (Oaks et al. 2004; Guiloski et al. 2015).
At present, many approaches have been proposed to remove DCF in water, including chemical oxidation, adsorption, membrane separation and advanced oxidation processes (Krajišnik et al. 2013; Ziylan & Ince 2013; Sarasidis et al. 2014). Among these technologies, ozonation and heterogeneous catalytic ozonation are highly recommended for better performance in removing this compound. Ozonation can rapidly decompose DCF due to the high reaction rate constant, but it is difficult for DCF to be completely converted to carbon dioxide and water, because many ozone-refractory intermediates form during the ozonation process. Heterogeneous catalytic ozonation generally accelerates the transformation of ozone into the •OH radical, which is a highly reactive oxidant with the reaction rate constants of 106 to 109 M−1S−1 with most of the organic compounds. In addition, the presence of a solid catalyst can also increase the utilization efficiency of ozone. DCF mineralization in catalytic ozonation was reported to achieve a significant enhancement (Beltrán et al. 2009). In order to improve the generation rate of •OH radicals in heterogeneous catalytic ozonation, various metal oxide-loaded porous materials were developed as heterogeneous catalysts for the ozonation process (Qi et al. 2012).
Pumice (PMC) is a kind of porous, light-weight material with well-developed internal pore structures, which is formed in the cooling process of molten magma, and can be easily obtained. Pumice possesses a large surface area, high strength, and good acid/alkali resistance. In the past, pumice was mostly used as an adsorption and supporting material in water treatment (Safari et al. 2014; Öztel et al. 2015). Recently, its catalytic activity and supporting character for the ozonation of low strength organics have been reported (Yuan et al. 2012). In addition, the metal silicate oxide in pumice was reported to be the heterogeneous catalyst; it could effectively improve the removal efficiency of refractory organic compounds, possibly due to its large surface area with surficial hydroxyl groups, and its metal-ion leaching was generally low (Liu et al. 2011).
In this work, to further improve the catalytic activity of metal silicate oxide in pumice, for the first time, iron silicate was loaded on the pumice by a co-precipitation–impregnation method. DCF was chosen as the model compound to evaluate the catalytic activity of iron silicate-loaded pumice (FSO/PMC). The objectives were to compare the characteristics of PMC and FSO/PMC, to evaluate the catalytic activity of FSO/PMC for ozonation of DCF in aqueous solution, and to explore the mechanism of DCF mineralization in the FSO/PMC/O3 process.
MATERIALS AND METHODS
Materials and reagents
Pumice was purchased from Zhangjiakou city (Hebei Province, China). DCF, oxamic acid and malonic acid were obtained from Tokyo Chemical Industry Co., Ltd (Tokyo, Japan). Oxalic acid was obtained from Bodi Chemical Co., Ltd (Tianjin, China). Methanol and acetic acid were high performance liquid chromatography (HPLC) grade. All the other chemicals and solvents were analytical grade and used without further purification. All solutions were prepared with distilled–deionized water (≥18.0 MΩ cm) produced from a Banstead NANO pure water treatment system (Thermo Fisher Scientific Inc., USA). The pH of the solution was adjusted by nitric acid and sodium hydroxide.
Preparation of FSO/PMC
Before the test, the raw pumice was crushed and sieved to a size of 0.2–0.3 mm particles, which were then soaked in 2 M nitric acid for 2 h to remove the impurities attached on the surface of the pumice. After that, the pumice was washed several times with ultra-pure water to remove the excess acid, then dried at 80°C for 16 h and stored for later use.
Iron silicate-loaded pumice (FSO/PMC) was prepared by adding 50 ml of 1.0 M Na2SiO3 slowly into 60 ml of 2 M HNO3 under magnetic stirring until the pH of the solution reached 2–3, then 75 ml of 1.0 M Fe(NO3)3 was added into the solution and stirred for another 30 min, followed by the addition of 20.0 g pretreated pumice and mixing for 30 min. After that, 2 M ammonium hydroxide was added dropwise into the solution to allow the iron–silicon oxide to be polymerized and deposited well on the surface of the pretreated pumice at room temperature under mechanical stirring. The pH of the suspension was adjusted to 8–9 and incubated at 60°C for 24 h, then the polymerized precipitates were sieved again and washed several times with ultra-pure water until the pH and conductivity of the supernatant remained at constant. Finally, the obtained solid was dried at 80°C for 24 h and stored for further use.
All tests were performed in a 500 ml flat-bottomed flask, which was placed in a thermostatic bath maintained at a desirable temperature. Ozone was produced from an ozone generator (OZAT-CFS-1A, Switzerland) using dry pure oxygen as the source. A 500 ml solution of 29.6 mg L−1 DCF and 400 mg of catalyst were added into the reactor under magnetic stirring, then the O3 gas was continuously bubbled into the solution with a silica diffuser. The excess ozone in the outlet gas was trapped by a 20% KI solution. Water samples were taken at regular intervals and simultaneously filtered with a 0.45 μm polytetrafluoroethylene (PTFE) membrane, and 0.1 M of Na2SO3 solution was used to quench the reaction of catalytic ozonation in the water samples. The adsorption process was performed under the same experimental conditions with the input of the oxygen gas. The evolution of aqueous ozone concentration in various oxidation processes was determined in the same reactor, in which water samples were withdrawn at regular intervals and immediately added into the indigo solution without filtering out the catalyst. An additional test confirmed that the presence of the catalyst showed no effect on the absorbance of the indigo solution. All the experimental processes were conducted in triplicate, and data shown in the figures are the average of the tests, and the standard deviations have been presented as error bars.
DCF was determined with an HPLC (Waters Instrument, USA) using a Zorbax extend-C18 column (5 μm, 0.46 × 25 cm, Agilent Technologies). The mobile phase during the detection of DCF was a mixture of methanol and acetic acid (1‰) at 90:10 (v:v), the flow rate was 0.8 ml min−1, the column temperature was 30°C, and the detection wavelength was 280 nm. Total organic carbon (TOC) was measured with a Shimadzu TOC 5000 analyzer. The aqueous ozone concentration was determined by the indigo method (Bader & Hoigné 1981). The gaseous ozone concentration was analyzed with the iodometric method (Birdsall et al. 1952). The surface morphology and dispersion of solid samples were carried out using a scanning electron microscope (SEM, Hitachi SU810, Japan). The N2 adsorption–desorption isotherms were obtained at 77 K on a surface area analyzer (Micromeritics ASAP 2020, USA). The surface area was determined by the Brunauer–Emmett–Teller (BET) method. The pore volume, pore size distribution and pore diameter were estimated using the Barrett–Joyner–Halenda (BJH) method. The element composition of solid samples was analyzed with X-ray fluorescence (XRF, Axios PW4400, The Netherlands). The carboxylic acids were measured with a Dionex ICS-2100 IC using an AS-11 column (4 × 250 mm), the mobile phase was 10 mM KOH with a flow of 1.0 ml min−1. The pHpzc of solid samples was evaluated with a mass titration method (Noh & Schwarz 1990), and the pHpzc of PMC and FSO/PMC was determined to be 6.48 and 7.21, respectively.
RESULTS AND DISCUSSION
Characterization of PMC and FSO/PMC
The chemical compositions of PMC and FSO/PMC were analyzed with XRF, and the results are listed in Table 1. It can be observed that both materials have similar composition. The oxides of Si, Al and Fe were the major constituents, while the other oxides were relatively minor. Modification of the pumice surface with loaded iron silicate obviously affected the contents of SiO2, Al2O3 and Fe2O3, in which the SiO2 and Al2O3 contents were reduced to 49.836% and 14.686%, respectively, while the Fe2O3 content was increased to 18.285%. Combining the results of SEM, N2 adsorption–desorption and XRF, it can be deduced that the co-precipitation–impregnation method is a good approach to loaded iron silicate on pumice.
|Component .||% (w/w) .||% (w/w) .|
|Component .||% (w/w) .||% (w/w) .|
Removal of DCF and TOC between different processes
The kinetic data of TOC removal were further analyzed among the three processes. The results showed that the TOC removal followed a two-stage pseudo-first-order kinetics. In general, a fast initial stage was followed by a slower final stage, which was particularly obvious for the ozonation and PMC/O3 processes. This suggests that TOC removal in the sole-ozonation and PMC/O3 processes is limited by the continuous supply of ozone, especially for the final stage, when the deficiency of ozone becomes more and more critical due to the competition of oxidants by the intermediates. However, it was interesting to note that the rate difference of the two stages almost vanished in the FSO/PMC/O3 process. Since all the three processes have the same ozone dosage, this observation indicated that the fast mineralization in the FSO/PMC/O3 process was less dependent on the ozone supply, but was likely contributed from a combined effect of adsorption and catalytic oxidation. The former performs as a reservoir to temporarily store and/or trap the intermediates nearby the catalyst and allows the latter to quickly and efficiently oxidize the adsorbed intermediates. This is supported by the huge surface area of FSO/PMC. A fast and consistent mineralization was therefore observed in the FSO/PMC/O3 process.
Discussion on the reaction mechanism of the catalytic ozonation process
Evolution of aqueous ozone concentration between different oxidation processes
In the O3/FSO/PMC process, the aqueous [O3] increased rapidly to 1.57 mg L−1 in 2 min and reached an equilibrium level at 2.41 mg L−1 in about 10 min. In comparison to that without the addition of the catalyst, the balanced aqueous [O3] of the former increased 17% within a shorter purging period. This indicated that the presence of the catalyst not only improves the mass transfer of ozone, but also acts as a reservoir to temporarily store the soluble ozone (and even traps some gaseous ozone) as an additional ozone source for future use. In the O3/FSO/PMC/DCF process, the aqueous [O3]-time profile was initially similar to that of the O3/DCF process, but the former reached a higher [O3] plateau than the latter, suggesting the presence of a second ozone source as proposed earlier.
The presence of DCF always led to a lower [O3] plateau than the corresponding case without DCF apparently due to ozone consumption. The ozonation of DCF resulted in the formation of less-ozone-sensitive/refractory intermediates, which could be adsorbed on the FSO/PMC surface and reacted with the stronger oxidative radical species (generated from the adsorbed/decomposed ozone on the catalyst surface). This could be the reason that TOC removal was significantly improved in the FSO/PMC/O3 process (Figure 3). Therefore, the presence of FSO/PMC can effectively improve the ozone decomposition on its surface and likely generate the more powerful oxidative radical species.
Generation of •OH radicals in the presence of FSO/PMC
Intermediates adsorption/degradation on catalyst surface
As mentioned before, the ozonation of DCF generated large amounts of ozone-refractory intermediates. The adsorption of these refractory intermediates on the solid catalyst may also be a crucial step for its mineralization (Nawrocki & Kasprzyk-Hordern 2010). To assess the adsorption of intermediates during the process, a two-step sequential test was performed. Firstly, the O3 gas was continuously bubbled into the DCF solution for 1 h, then the remaining dissolved ozone was removed by the continuous input of oxygen gas for 10 min. Secondly, FSO/PMC was added into the solution followed by waiting (while stirring) for 20 min. Results showed that the addition of FSO/PMC led to a TOC reduction by 17.8%, which was much higher than that of DCF adsorption on FSO/PMC. It indicated that the intermediates generated from DCF decay exhibited much higher affinity toward FSO/PMC relative to DCF.
Based on the above discussion, the elimination of DCF and generated intermediates during the FSO/PMC/O3 process mainly involved three pathways: (1) the direct oxidation of ozone molecules, (2) the generated intermediates during the ozonation of DCF being adsorbed on the surface of the FSO/PMC and undergoing a surface reaction, (3) the presence of FSO/PMC significantly increasing the •OH radical yield, leading to the elimination of the generated refractory intermediates.
Stability of FSO/PMC
The reusability of the solid catalyst has to be taken into account before its real application in water treatment (Yan et al. 2013). In this work, the catalytic activity of FSO/PMC was examined in five successive ozonation experiments. After each experiment, the used catalyst was filtered by a 0.45 μm cellulose acetate membrane and dried at 80°C for the next cycle. As observed in Table 2, there was a slight variation in the TOC removal efficiency after five successive ozonation experiments. This phenomenon indicated that FSO/PMC has high catalytic activity and stability during the ozonation of DCF. In addition, the leaching of iron ions during the FSO/PMC/O3 process was also investigated. The results indicated that the leaching of iron ions for each experiment could be ignored relative to the catalyst dosage. The maximum leaching concentration of iron ions was 0.0087 mg L−1, which further demonstrated the high stability of FSO/PMC during the ozonation process.
|Repeat times .||TOC removal (%) .||Fe leaching (mg L−1) .|
|Repeat times .||TOC removal (%) .||Fe leaching (mg L−1) .|
Iron silicate-loaded pumice (FSO/PMC) was synthesized by a co-precipitation–impregnation method. FSO/PMC exhibited excellent catalytic activity for DCF mineralization during ozonation. Under the reaction conditions, TOC removal reached 73.3% with the FSO/PMC/O3 process in 60 min, while only reaching 46.4% and 32.3% with the PMC/O3 and sole-ozonation processes, respectively. The presence of FSO/PMC obviously improved the mass transfer of ozone and promoted the transformation of ozone into •OH radicals. Meanwhile, FSO/PMC exhibited to some extent the adsorption of intermediates such as ozone-refractory carboxylic acids, and further mineralized them. FSO/PMC kept good reusability and stability during the ozonation process. The synergetic effect between FSO/PMC and O3 illustrated that FSO/PMC is a promising metal oxide-loaded porous catalyst in a catalytic ozonation process.
The authors are grateful for the financial support from the open funding project of HKPolyU-RISUD Joint PhD Supervision Scheme (1-ZVCS) and State Key Laboratory of Urban Water Resource and Environment of Harbin Institute of Technology of China (Contract No. QAK201307).