Introduction

In the last decades, the development of new catalytic oxidation technologies for the removal of organic pollutants has received a great attention. The removal of organic compounds, especially phenolic ones, from wastewaters cannot be achieved by conventional treatments. So, the catalytic wet air oxidation has been assigned an outstanding priority over other wastewater treatment technologies, e.g. photocatalysis (Kanakaraju et al. 2014), electrocoagulation and electro-Fenton (Daghrir and Drogui 2013). It is an attractive method for the treatment of pollutants, which are too diluted to incinerate and too toxic to treat biologically due to their ability to reach mineralization (Cybulski 2007). The use of catalyst makes the process more attractive by achieving high degradation of pollutants at lower temperature and pressure. Moreover, the solid catalyst offers a further advantage, compared to homogeneous catalysis, as the catalyst is easily regenerated and reused (Song and Lu 2015; Bhargava et al. 2007). Noble metals such as Ru and Pt supported on Ti and Zr oxides have been found to be active and stable in catalytic wet air oxidation (Pham et al. 2007; Wang et al. 2008; Mikulová et al. 2007). Previous studies (Oliviero et al. 2001; Renard et al. 2005) have demonstrated that the nature of the material influences the catalytic wet air oxidation of organic compounds. This technology is used as a pretreatment of olive oil mill wastewaters, which are toxic and hardly biodegradable due to the presence of phenolic pollutants (Pintar et al. 2004; Manole et al. 2007). We used p-hydroxybenzoic acid, typically found in olive oil mill wastewaters as model molecule which its physicochemical properties are given in Table 1.

Table 1 Physicochemical properties of p-hydroxybenzoic acid
Full size table

In our previous work, the catalytic wet air oxidation of p-hydroxybenzoic acid was conducted in a batch reactor in the presence of ruthenium catalysts supported by CeO2 mixed oxides, i.e. CeO2–Al2O3 (Triki et al. 2011) and CeO2–TiO2 (Triki et al. 2009). These mixed oxides were elaborated via the sol–gel method combined with supercritical drying. This method offers not only a good homogeneity of the different precursors but also the possibility of better controlling of the metal particle size and porosity of the catalyst (Mishra et al. 2004). We have found that doping CeO2 with an optimum ratio (for the Ce–Al(20) support, 20 is a weight ratio (CeO2/(CeO2 + Al2O3)) and for Ce–Ti(1/3), 1/3 is a molar ratio (Ce/Ti)) leads to good activity of the supported Ru catalysts in the catalytic wet air oxidation of p-hydroxybenzoic acid (Triki et al. 2011). Our results showed that decreasing the Ce/Ti molar ratio from 3 to 1/3 in CeO2–TiO2 materials improves the activity of Ru-supported catalysts in the catalytic wet air oxidation of p-hydroxybenzoic acid. Nevertheless, CeO2–TiO2 seems to be an interesting support worth further exploration. The aim of this work is to study the catalytic properties of Ru/CeO2–TiO2 in the catalytic wet air oxidation of p-hydroxybenzoic acid batch reactor at 140 °C and 50 bar of air. We thoroughly investigated the effect of the Ce/Ti molar ratio, specific surface area of the supports and the dispersion of metal on the catalytic properties of Ru catalysts.

Materials and methods

Catalyst preparation

CeO2 was prepared by dissolving Ce(NO3)3·6H2O in absolute ethanol. After stirring, droplets of NH4OH were added to obtain a precipitate. CeO2–TiO2 mixed oxides were prepared by the sol–gel method as follows: Ce(NO3)3·6H2O was dissolved in absolute ethanol. After the addition of ethylacetoacetate, as chelating agent, titanium(IV) isopropoxide was added in the desired Ce/Ti molar ratio (1/6, 1/5 or 1/4). Hydrolysis was performed with HNO3 aqueous solution (0.1 M). The obtained gels were dried in supercritical conditions of ethanol (T = 243 °C, P = 63 bar) and then calcined under oxygen at 500 °C for 3 h. TiO2 was prepared in the same way but without the addition of cerium nitrate. The 3 wt% Ru catalysts were prepared by impregnation of the support with an ethanolic solution of Ru(NO)(NO3)3. The mixture was dried at 90 °C for 24 h and then reduced under H2 at 300 °C for 2 h. The catalysts were labelled as Ru/CeTi(x), where x corresponds to the Ce/Ti molar ratio.

Catalyst characterization

N2 physisorption at 77 K was carried out in an automatic ASAP 2020 apparatus from Micromeritics. The specific surface area was determined using Brunauer–Emmett–Teller (BET) method, and the average pore diameter was calculated with Barrett–Joyner–Halenda (BJH) method. X-ray diffraction (XRD) measurements were taken using a Philips PW1050 diffractometer with CuKα radiation at λ = 1.5418 Å. The phases were identified with reference to Joint Committee on Powder Diffraction Standards (JCPDS) cards. The crystallite size of the samples was calculated using the Debye–Scherrer equation:

$$ d_{\text{XRD}} = \frac{K \lambda }{\beta \cos \theta } $$
(1)

where d XRD is the crystallite size (nm), K is the shape factor (0.90), λ is the wavelength (0.154 nm), β is the corrected full width at half maximum (radian), and θ is the Bragg angle (radian). H2 chemisorption studies were conducted in a Micromeritics ASAP 2010. Before the analysis, the catalyst was reduced in situ (H2, 300 °C, 2 h). After reduction, the H2 on the catalyst surface was removed by He (300 °C, 2 h). The sample was subsequently cooled to 36 °C under the same He flow, and hydrogen chemisorption was then performed at the same temperature. Ruthenium dispersion and particle size were determined assuming an H/Ru stoichiometry of 1. Metal dispersion was calculated based on the monolayer coverage of H2 chemisorption on Ru:

$$ D\left( \% \right) = \frac{{(V_{\text{mon}} /22414)}}{{{\text{metal}}\;{\text{weight}} \left( \% \right)}} \times S \times M \times 100 $$
(2)

where V mon is the monolayer coverage of H2 on Ru (cm3/g STP), 22,414 is the molar volume (mL/mol, STP), S is the stoichiometric factor of H/Ru (S = 1), and M is the atomic mass of Ru (101.07 g/mol). The Ru crystallite size was calculated from the dispersion values assuming the spherical model. Transmission electron microscopy (TEM) images were recorded with a JEOL JEM-1011 electron microscope at 200 kV. The samples were suspended and dispersed by ultrasonic treatment in ethanol. A drop of the fine suspension was placed on a copper TEM grid, which was then loaded into the microscope.

Catalytic wet air oxidation experiments

Catalytic tests were conducted in a Hastelloy C22 batch reactor at 140 °C and a total air pressure of 50 bar with a stirring rate of 1500 rpm. These experiment conditions were previously optimized by Pham et al. (2007). The reactor was loaded with 150 mL of p-hydroxybenzoic acid (CAS N° 99-96-7, Sigma-Aldrich, 99 %) aqueous solution (10 mmol/L; total organic carbon (TOC) = 840 mg/L) and 500 mg of catalyst. After flushing with N2 to remove air, the reactor was heated to 140 °C under stirring. Then, the stirrer was stopped and pure oxygen was introduced into the reactor until the total pressure reached up to 50 bar. The reaction started when the stirring was switched on. This point was taken as “zero time”. Liquid samples were periodically withdrawn from the reactor every 20 min for the first hour and then every hour, and the reaction time was 7 h for all runs. Samples were analysed by means of high-performance liquid chromatography (HPLC, Thermo Finnigan) using a C18 column (Hypersil Gold column 4.6 mm × 250 mm) heated at 40 °C and an ultraviolet detector at 254 nm. The mobile phase was composed by H2O (Milli-Q), acetonitrile and H2SO4 (70/20/10) with a flow rate of 0.7 mL/min. Identifications and quantifications of the aromatic compounds and organic acids were achieved by comparing with standard. Total organic carbon (TOC) was measured with a Shimadzu Model 5050 TOC analyser. Ru leaching and Ce leaching of the catalysts were determined by inductively coupled plasma (ICP) performed on a PerkinElmer Optima 2000 DV ICP apparatus.

Results and discussion

Catalyst characterization

Textural analysis showed that both specific surface areas and average pore diameter of CeO2–TiO2 are higher than those of pure TiO2 and CeO2. Compared to unmixed TiO2 and CeO2, specific surface area increases from 82 to 183 m2/g for the CeO2–TiO2 samples. This result could be due to the modification of the hydrolysis–condensation reaction rates by cerium nitrate (Lopez et al. 2004). Moreover, the textural properties of the CeO2–TiO2 mixed oxides depend on the Ce/Ti molar ratio. In fact, the specific surface area of CeTi(1/5), CeTi(1/6) and CeTi(1/4) was 183, 146 and 125 m2/g, respectively. The presence of cerium during the formation of the TiO2 gel had a marked effect on the specific surface area of the mixed solid. This effect has also been observed during the preparation of doped TiO2 by means of the sol–gel depending on the nature of the doping agent (Gomez et al. 2003). Thus, it is clear that the addition of Ce to TiO2 causes significant textural changes in the CeO2–TiO2 mixed oxides. On the other hand, a slight decrease in the specific surface area was observed after the addition of ruthenium. This could be attributed to some blockage of the porous structure of the support by Ru particles.

X-ray diffraction analysis revealed that for CeO2, the peaks at 2θ = 28.6, 33.1, 47.5 and 56.3° (JCPDS 34-0394) were attributed to cubic CeO2. For TiO2, only the anatase phase represented by peaks at 2θ = 25.2, 37.8, 48.05 and 53.8° (JCPDS 21-1272) was detected. The CeTi(1/5) sample presents patterns with peaks of both anatase and cubic CeO2 but broader and weaker than those of pure oxides, indicating a good dispersion of its crystalline phases. In addition, we found that with the introduction of Ce, the crystallite size decreased from 49 to 10, 8 and 40 nm for CeTi(1/6), CeTi(1/5) and CeTi(1/4), respectively, indicating that the Ce species inhibited the growth of anatase crystallites. No diffraction peak of Ru was detected in the patterns of the catalysts, which could indicate the presence of very small particles and/or indicate that the Ru particles are amorphous (Okal et al. 2011).

From the H2 chemisorption studies, the results showed that the Ru/TiO2 and Ru/CeTi(x) catalysts presented a Ru dispersion of 9.5–32 % and therefore particle sizes of about 3.1–9.4 nm. The Ru dispersion of Ru/CeTi(1/5) was the highest (32 %), which could be due to the high specific surface area of the CeTi(1/5) support.

Transmission electron microscopy images of Ru catalysts are presented in Fig. 1. The images show small Ru particles with uniform dispersion. The Ru/TiO2 and Ru/CeTi(1/5) catalysts presented small crystallite sizes. These results are in agreement with those obtained by H2 chemisorption and X-ray diffraction.

Fig. 1
figure 1

Transmission electron microscopy images of Ru-supported catalysts. a Ru/CeTi(1/5); b Ru/TiO2; and c Ru/CeO2. The figure shows small Ru particles with uniform dispersion

Full size image

Catalytic tests

Figure 2 shows the evolution of the p-hydroxybenzoic acid conversion over time in the presence of Ru catalysts. In the absence of the catalyst or in the presence of the support alone, conversion of the p-hydroxybenzoic acid remained below 7 % after 5 h of reaction and then was nearly constant under the operating conditions. With the addition of Ru, the conversion of p-hydroxybenzoic acid was significantly enhanced and high catalytic activity was observed. Our results showed an increase in p-hydroxybenzoic acid conversion over Ru-supported CeO2–TiO2 mixed oxides comparing to pure oxides. In fact, after 5 h of reaction, the Ru/CeTi(1/6) and Ru/CeTi(1/4) catalysts present higher conversions of p-hydroxybenzoic acid (84 and 79 %, respectively) compared to Ru/TiO2 and Ru/CeO2 (63 and 40 %, respectively). Moreover, total conversion of p-hydroxybenzoic acid and 93 % of total organic carbon abatement after 7 h of reaction were obtained with the Ru/CeTi(1/5) catalyst.

Fig. 2
figure 2

p-hydroxybenzoic acid conversion in the presence of Ru-supported catalysts (triangle symbol) Ru/CeTi(1/5); (filled square symbol) Ru/CeTi(1/6); (times symbol) Ru/CeTi(1/4); (circle symbol) Ru/TiO2; (diamond symbol) Ru/CeO2; and (opened square symbol) without catalyst. Note the increase in p-hydroxybenzoic acid conversion over Ru-supported CeO2–TiO2 mixed oxides comparing to pure oxides. The highest conversion was obtained with an optimum Ce/Ti molar ratio of (1/5)

Full size image

The activities of the supported Ru catalysts increased as follows: Ru/CeTi(1/5) > Ru/CeTi(1/6) > Ru/CeTi(1/4) > Ru/TiO2 > Ru/CeO2. This order of activity is similar to that observed for the specific surface areas of these different catalysts. Ru/CeTi(1/5) (175 m2/g) was more active than Ru/CeTi(1/6) (124 m2/g), which in turn is more active than Ru/CeTi(1/4) (110 m2/g). This point is even more evident when correlating the initial reaction rate for the catalytic wet air oxidation reaction based on the specific surface area. Our results showed that the initial activities of the supported Ru catalysts are quite high particularly for Ru/CeTi(1/5) which reaches an initial reaction rate of 0.31 mmol/gmin. The Ru/CeTi(1/5) catalyst is about three times more active than Ru/CeTi(1/4) (0.13 mmol/gmin) and Ru/CeTi(1/6) (0.15 mmol/gmin). The Ru/CeTi(1/4) catalyst, although less active, still exhibits higher activity than the catalyst supported into pure oxide (The initial reaction rates of Ru/TiO2 and Ru/CeO2 were 0.07 and 0.03 mmol/gmin, respectively). A direct correlation can be observed between the enhancement of the textural properties of the mixed oxide support and catalytic wet air oxidation activity. The increase in catalytic wet air oxidation activities could be ascribed to an increase in the specific surface area of the CeO2–TiO2 support, which would enhance the adsorption of oxygen on the catalyst surface. The Ce–Ti system, prepared via the sol–gel method, plays a significant role compared to the corresponding single-oxide system because it permits good CeO2 dispersion, promotes its electronic properties and so improves the dispersion of the metal (Epron et al. 2002). On the other hand, it seems that the reduction behaviour of Ru on the Ce–Ti system plays a crucial role in the improvement of catalytic activity because the Ru–O–Ce bonds in the well-dispersed Ru particles are highly fragile. The most important intermediates observed during the reaction were phenol, hydroquinone, maleic acid and oxalic acid as previously reported (Triki et al. 2009). Figure 3 presents a simplified reaction scheme describing the formation of the main intermediate compounds during p-hydroxybenzoic acid degradation. This degradation proceeds through decarboxylation to phenol and further hydroxylation of the aromatic ring to yield mainly hydroquinone. The breakdown of the aromatic ring results in the formation of a wide range of cleavage compounds of these; it was possible to detect maleic acid and fumaric acid (by isomerization), CO2 and water (Pham et al. 2007).

Fig. 3
figure 3

Simplified reaction scheme of p-hydroxybenzoic acid degradation. The oxidation of p-hydroxybenzoic acid proceeds through decarboxylation to phenol and further hydroxylation of the aromatic ring to yield mainly hydroquinone. Breakdown of the aromatic ring results in the formation of cleavage compounds of these. The formation of maleic and fumaric acids (by isomerization), as well as CO2 and water, was detected

Full size image

During catalytic wet air oxidation, the active component leaches out from the catalyst, which could cause additional pollution. To investigate the catalyst stability in terms of metal leaching, the concentration of Ru, Ti and Ce ions in the solution after 7 h of reaction was analysed using inductively coupled plasma. We found that no metal leaching of Ce, Ti or Ru (<0.02 mg/L) was detected in any of the treated aqueous solutions, indicating a good stability of Ru catalysts during the catalytic wet air oxidation of p-hydroxybenzoic acid.

Conclusion

Ru catalysts supported on CeO2–TiO2 aerogel mixed oxides exhibit good activity in the catalytic wet air oxidation of p-hydroxybenzoic acid. Ru/CeTi(1/5) shows the highest activity (100 % of p-hydroxybenzoic acid conversion after 7 h). This molar ratio makes it possible to obtain the highest specific surface area for the support and therefore a better Ru dispersion with small particle sizes. The high catalytic activity is due not only to the high dispersion of Ru particles, but also to the synergy effect between CeO2 and TiO2 affecting the structural properties of the CeO2–TiO2 support and so permitting an enhancement of O2 availability to the active sites. Further studies are, however, required to characterize the catalyst after reaction and to optimize the operating conditions under realistic conditions.