Deactivation and regeneration of wet air oxidation catalysts

Abstract

Available data concerning Wet Air Oxidation heterogeneous catalyst deactivation were collected and discussed in this review article. Deactivation mainly takes place through catalyst oxidation (overoxidation, superficial oxidation of metal particles, leaching), sintering and fouling. Only some of these phenomena appear to be reversible and strategies were developed in order to limit or prevent them from occurring. Compared to other industrial sectors, deactivation of WAO catalysts still remains insufficiently known. In particular, future research should be directed to the study of catalyst resistance to mechanical damages and to poisoning.

---

Introduction

With the intention of improving their standard of living, human beings have continuously developed new technologies, which have frequently affected the environment. Nowadays, air, soil and water pollution has to be dealt with. Many industrial sectors (petrochemical, pharmaceutical, foodetc.) and everyday life activities produce liquid effluents containing toxic and/or refractory compounds, which have to be treated before discharge or recycling in an industrial process. In addition, mass production, agriculture, urbanization and population growth are so many factors leading to an increase in the amount of contaminated waters. As a consequence, the availability of drinking water has dramatically dropped during the last decades. The will to limit this phenomenon led to the establishment of increasingly strict water pollution regulations.

The selection of an appropriate pollution removal process depends on many factors such as the nature (toxicity, biodegradability etc.) and the concentration of pollutants. In addition, the flow rate of the effluent to be treated has to be taken into account. Biological treatment, which relies on digestive properties of certain microorganisms, like bacteria or fungi, as well as incineration, which consists of a total combustion at temperatures above 1000 °C, are the most frequently employed. Nevertheless, biological treatment takes a long time (from 25 to 60 h), requires large installations and cannot be considered for the treatment of non biodegradable, too concentrated or toxic compounds, which could be deadly to the employed microorganisms. It also produces sludge, which has to be recovered and treated. For its part, incineration generates ashes, solid wastes and toxic fumes containing NO_x, SO_x and dioxins. Moreover, since it is particularly expensive, its use is limited to highly concentrated effluents, which cannot be treated otherwise.

The alternatives to these reference processes are physical and chemical treatments. Physical processes, like stripping, adsorption or membrane processes, are generally intended for low concentrated wastewaters and aim at reducing the pollution below acceptable levels. A wide panel of chemical treatments is also available. Chemical oxidation can be carried out at room temperature through the use of classical chemicals such as chorine, chlorine dioxide or potassium permanganate but these reactants are very expensive and environment-unfriendly. Advanced oxidation processes (AOPs) rely on the use of very active radicals, produced from UV-light activated oxygen (photocatalysis process), ozone (ozonation process) or hydrogen peroxide (wet peroxide oxidation and Fenton processes). Nevertheless, the high costs of these reactants limit the use of such processes to the treatment of low concentrated effluents (typically having a chemical oxygen demand below 5 g L⁻¹), which do not need significant amounts of the oxidizing agent.¹

Wet Air Oxidation (WAO) processes aim at totally degrading aqueous pollutants in the presence of pressurized oxygen. They are particularly adapted to wastewaters that are too dilute to be incinerated and too toxic to be biologically treated. Known since the early 20th century,^2,3 these processes can be applied to spent activated carbon regeneration and to the treatment of various wastes such as sewage sludge, distillery wastes and many effluents resulting from paper, textile, food or chemical industries.^4,5

Thermal Wet Air Oxidation (TWAO) is performed at elevated temperatures (125–450 °C) and high pressures (5–200 bar).^4–6 These harsh conditions are necessary to obtain acceptable performances. Typical conversion values range between 80 and 99% after 10 to 120 min of reaction. However, complete mineralization of the pollutants cannot be achieved owing to the formation of certain low molecular weight intermediate compounds, mainly acetic and propionic acids, which are resistant to oxidation^7,8 and accumulate in the reaction medium.⁹ TWAO is recognized as a highly efficient and viable water pollution removal technique. Nevertheless, its main drawback is its high cost, which directly results from its severe operating conditions. Catalytic Wet Air Oxidation (CWAO) was considered in order to moderate these conditions.

Having recourse to a catalytic system significantly improves the overall efficiency of the process, in particular, it makes the elimination of refractory compounds much easier. The operating pressure and temperature can typically be lowered below 30 bar and 200 °C, respectively, which results in a decrease of both capital and operating costs. Numerous homogeneous and heterogeneous catalysts were tested.

Soluble metallic salts usually are very efficient since they are in direct contact with the reactants. The most active amongst them are copper and iron salts.¹⁰ However, these metallic species need to be precipitated and separated from the treated effluent, which increases both the duration and the cost of the process and restricts the commercial applications of such catalysts.

On the contrary, heterogeneous catalysts can easily be recovered from the aqueous medium. They can be divided into three categories.

(i) Many transition metals (Co, Cr, Cu, Fe, Mn, Ti, V, Znetc.) were tested as catalysts, under the form of supported, bulk, simple or mixed oxides. These catalysts are not very selective and quickly deactivate by leaching or formation of a carbonaceous overlayer over their surface.^5,11 The CuO simple oxide¹² and MnO₂–CeO₂ composite oxide¹³ are probably the most active transition metal-based catalysts employed in CWAO.

(ii) Carbon-based catalysts are resistant to the acidic conditions met in CWAO. Furthermore, they are active even in the absence of deposited metal species. Activated carbons are probably the most promising carbon-based catalysts, in consequence of their significant adsorption properties.^14,15

(iii) Noble metal-based catalysts are classically supported on γ-Al₂O₃, CeO₂, TiO₂, ZrO₂, SiO₂ or activated carbon. The metal contents typically range from 0.1 to 5.0 wt%. Although they are much more expensive than the oxides of transition metals, they are more stable¹⁶ and more active. In particular, they make the degradation of refractory intermediates much easier. Amongst the frequently studied noble metals (Ir, Pd, Pt, Rh, Ru), platinum, ruthenium and palladium reveal to be the most efficient ones,^16,17 but it appears to be difficult to obtain a ranking of these metals since their activity greatly depends on the support phase and on the catalyst synthesis protocols.^18–20 Their performances can exceed those of copper salts,¹⁸ in typical operating conditions the use of noble metal-based catalysts can lead to a total conversion of the pollutants and to mineralization rates reaching 85–98%.

As stated by Besson and Gallezot, in their review about the deactivation of metal catalysts, although the literature concerning liquid phase reactions is abundant, very few studies were focused on deactivation, ageing, poisoning and regeneration of employed catalysts.²¹ The authors suggested that this lack of information was related to the fact that these reactions are, for the most part, carried out in batch reactors, which are not adapted to deactivation studies. Indeed, catalyst recovering and recycling procedures are complicated because of the working mode of such reactors and the risk of an artificial decrease in activity, caused by an uncontrolled loss of catalyst, can never be ruled out. Additionally, conducting such studies turn out to be arduous owing to the significant number of phenomena potentially involved in deactivation (leaching, poisoning, thermal degradationetc.).

The present paper aims to review several aspects concerning the deactivation and reactivation of Wet Air Oxidation catalysts.

Deactivation is more rarely met in the field of homogeneous catalysis than in the case of heterogeneous catalysis. It generally occurs when the metal active species are made unavailable, due to precipitation or complexation, for example. This particular case will not be described in this article.

The five main causes for heterogeneous catalyst deactivation are sintering of the support or active phase, leaching of the active species, fouling of the surface by polymeric species, poisoning of the catalytic sites by strongly adsorbed species and mechanical damages. Poisoning is usually related to the presence of halogen-, sulphur- or phosphorus-containing compounds. Unfortunately, very little information can be found concerning this deactivation mode in the domain of CWAO. Despite its practical interest, the mechanical damaging of WAO catalysts is also rarely taken into account, probably because most of the fundamental studies are carried out in the presence of powder catalysts. For these reasons, both poisoning and mechanical damages will not be discussed further in this paper. After a description of the most frequent phenomena involved in WAO solid catalyst deactivation, available regeneration procedures will be described and discussed. It is to be noted that, from an industrial point of view, deactivated catalysts are generally preferred to be ex situ regenerated rather than in situ regenerated for several reasons, including time, safety considerations and better recovery of activity.²²

Oxidation of catalyst components

1. Oxidation of support phase

Oxidative conditions met during CWAO can be responsible for a degradation of carbon-based catalysts¹⁵ and for an alteration of their surface properties (specific surface area, porosity or surface functional group modifications).^19,23 These phenomena can be observed at a temperature as low as 120 °C¹⁹ and classically result in a decrease in catalytic activity. This deactivation pathway may be limited by the use of milder operating conditions but cannot be totally avoided. Of course, this phenomenon is irreversible since the catalyst itself is consumed. Shih et al. also observed that the oxidation of a Pt/C catalyst surface during CWAO was responsible for a decrease in the affinity between Pt and the carbon-based support, which led to migration and aggregation of Pt clusters into large particles. This support-oxidation induced sintering phenomenon could be limited by grafting TiO₂ onto the carbon surface in order to anchor the Pt particles.²⁴ The extensive use of carbon-based catalysts seems to be compromised as a result of their instability in CWAO conditions.

2. Oxidation of metallic active phase

Speaking of the negative influence of oxidative medium, a too-high catalyst surface coverage by adsorbed oxygen can be detrimental to activity, insofar as adsorption of pollutants becomes limited and electronic transfer from the catalyst to the adsorbed molecule is made more difficult. This phenomenon, referred to as overoxidation, is likely to occur in CWAO conditions, considering the high oxygen pressures employed.²⁵ It mainly depends on the nature, the structure and the texture of the catalyst. Thereby, high redox potential metals, such as platinum, palladium or gold, are less prone to this type of deactivation.²¹ Furthermore, small particles (<2 nm) deactivate more readily owing to their higher affinity with oxygen.

A demonstration of this deactivation pathway can be found in the works of Béziat et al. concerning the WAO of succinic acid by a 2.8% Ru/TiO₂ catalyst (Fig. 1). These authors noted a reversible decrease in activity at 150 and 160 °C²⁶ as well as a significant influence of reducing or oxidizing pretreatments to which the catalyst was subjected.²⁷ They attributed these behaviours to the coverage of the surface of ruthenium particles by oxygen. In their working conditions (T = 150 °C, [Phenol] = 0.02 mol L⁻¹), Masende et al. showed that the inhibitory effect observed at stoichiometric oxygen excesses above 80% were due to the loss of activity of the employed 5% Pt/Graphite catalyst as a result of overoxidation.²⁸

Fig. 1 Deactivation by overoxidation of a 2.8% Ru/TiO₂ catalyst during CWAO of succinic acid at 150 °C in a trickle-bed reactor.²⁶

Overoxidation of the catalyst is a reversible phenomenon, which, thereby, is not an actual problem. This deactivation mode can thus be easily reversed by operating at lower oxygen concentrations, which can be achieved by regulating the flow rate of this reactant, by diluting it into an inert gas (usually nitrogen) or by working at conditions in which oxygen transfer is diffusion limited. The redox potential of the molecules present within the reaction medium can also play a role and highly reducing species can limit overoxidation.²¹

Metal oxide can also be formed, on the surface or in the bulk phase of the metal particles, depending on the redox potential of the metal. Indeed, as shown in Fig. 2 for standard conditions (T = 298 K, P = 1 bar), metal oxidation state depends on both the acidity and the potential of the medium.²⁹ For example, it is well-known that the surface layer of the platinum-group metals is oxidized upon contact with air.²¹ Larachi et al. characterized, by XPS, spent 1% Pt/Al₂O₃ and 1% Pt/MnO₂–CeO₂ catalysts recovered from phenol WAO experiments.³⁰ In all cases, analyses revealed the presence of PtO and PtO₂oxides but no metallic Pt⁰ surface species could be detected. Several authors reported that the oxidation of the metal phase could be responsible for a loss of activity of the catalysts employed in the WAO of certain compounds, such as ammonia³¹ or acetic acid.³² Delanoë studied the influence of reducing and oxidizing treatments upon the activity of a 5% Ru/CeO₂ catalyst tested in the oxidation of acetic acid and confirmed that better performances could be achieved when ruthenium was under its reduced state.³³ This author also observed that, during the reaction, ruthenium particles were oxidized on their surface as RuO₂³⁴ and that a reducing treatment could make it possible to recover more than 90% of the initial catalytic activity.³³ The negative consequences of an oxidizing treatment on activity were also confirmed by Béziat et al. in the WAO of succinic acid by a 2.8% Ru/TiO₂ catalyst.²⁷ Pintar et al. estimated that 22 to 30% of deposited ruthenium was partially oxidized as RuO₂ during WAO process. This phenomenon was responsible for a loss of activity in the case of acetic acid oxidation while it did not seem to affect the performances in the case of formic acid and was not the prevailing factor during the degradation of phenol.³² This phenomenon is not as easily reversed as simple strong oxygen adsorption but the initial activity can be recovered by subjecting the catalyst to a reducing treatment. Finally, it is also to be noted that a hot and acidic medium does not only affect supported particles but can also lead to support phase transformation. Massa et al. studied the CWAO of phenol aqueous solutions over several 1% Ru/% CeO₂–Al₂O₃ catalysts and concluded that deactivation was partly due to the transformation of the alumina support into the boehmite phase.³⁵

Fig. 2 Changes in the chemical state of various metals at T = 25 °C, P = 1 bar, pH = 3, depending on solution potential. The solid line (E = 0.5 V/SHE) corresponds to the potential of Ru measured in a 0.083 mol L⁻¹acetic acid solution in the presence of oxygen.³⁴

3. Leaching

Leaching of the active components of a heterogeneous catalyst is the major cause for deactivation met in liquid phase reactions. It routinely and significantly takes place and its consequences are both the deactivation of the catalyst and the generation of a secondary pollution. Too often, authors have claimed promising results for heterogeneous catalysts while the observed activity was, in fact, essentially due to the active species dissolved in the reaction medium.

As already mentioned, it is well-known that both the acidity and the potential of the reaction medium affect the oxidation state and the form under which one element exists. These latter can be predicted through the consideration of Pourbaix diagrams and, in WAO conditions, it appears that many elements tend to leach. Indeed, during the oxidation of organic substances carboxylic acids are formed and are responsible for a decrease in the pH of the medium. Highly oxidizing and acidic conditions, as well as the elevated temperatures met in Wet Air Oxidation processes, clearly favour the leaching of catalyst components. Zhang and Chuang studied the influence of pH on the stability of a Pd/Al₂O₃ catalyst and palladium and aluminium leaching was revealed to be particularly significant in highly acidic (pH = 2) and highly basic (pH = 11) media.³⁶ Hamoudi et al. observed that in the presence of a MnO₂–CeO₂ catalyst, the higher the temperature, the more important the amount of dissolved manganese.³⁷

Leaching is the most common cause for the deactivation of oxides of transition metals.^11,16,38 For example, copper concentrations of 75 ppm,¹¹ or exceeding 150 ppm,^39,40 were reported during the use of copper-based catalysts. Likewise, manganese is not very resistant to leaching and concentrations ranging from 5³⁷ to 348 ppm⁴⁰ can be found in the literature. Recently, Kouraichi et al. observed that Mn leaching was higher for pre-reduced MnO_x–CeO₂ catalysts, owing to the poor stability of the MnO oxide under acidic conditions.⁴¹ In fact, Pourbaix diagrams clearly show that amongst metal oxides ZrO₂, TiO₂ and CeO₂ are the most resistant to leaching in WAO conditions and, as such, constitute the best candidates for catalytic supports. Indeed, neither titanium^26,27,42 nor zirconium⁴² could be detected in solution, when employing the corresponding oxides. On the contrary, other supports, like alumina, tend to dissolve into the reaction medium.⁴³ Concerning Ce, although some studies concluded that it was totally stable,^43,44 it was, in a few cases, detected in the liquid phase, at very low concentrations (from 0.04⁴⁵ to 0.18 ppm).³⁷

Noble metal-based catalysts are considerably more resistant to this deactivation mode.^37,46 In the rare cases when a loss of metal could be observed, it was essentially due to the dissolution of the support phase. In fact, the stability of the noble metals is such a recognized fact that very few authors considered necessary to verify it. Available papers mention a perfect stability of Ru.^{15,25–27,42} Wang et al. reported a ruthenium maximal concentration of 0.01 ppm during experiments performed with ruthenium-based catalysts supported on ceria-zirconia pellets in a continuous packed-bubble column reactor.⁴⁵ The relevance of this information is debatable (in particular, what are the detection limits of the equipment used?). Moreover, all the noble metals do not show the same resistance to leaching. Zhang and Chuang, who tested a Pt–Pd–Ce/Al₂O₃ catalyst, observed that only palladium dissolved in the reaction medium.⁴³ Particularly interesting results were recently presented by Grosjean et al.⁴⁷ These authors considered the WAO of N,N-dimethylformamide in the presence of noble metal-based (Pt, Pd, Ru) heterogeneous catalysts supported on TiO₂ or ZrO₂. They reported a fast dissolution of the active components, ranging from 20 to 100% of deposited metals and attributed this phenomenon to the formation of dimethylamine and methylamine. These compounds, which possess lone pair electrons on their nitrogen atom, are able to complex the metals, thus leading to leaching.

Leaching of the active species generally leads to an irreversible alteration of heterogeneous catalysts. Thereby, the best attitude to take regarding this deactivation pathway simply consists of avoiding it. For this purpose, numerous works have been aimed at reducing the severity of the operating conditions, at modifying the synthesis protocols of the catalysts, as well as the pretreatments they were subjected to, or at searching for more suitable catalytic supports.

During WAO experiments, the main modifiable operating parameters are temperature, pressure, concentration and pH. It appears that, in the case of copper, temperature does not significantly affect leaching while the acidity of the medium plays a predominant role in this phenomenon.⁴⁸Leaching notably occurs when the pH value is below 4 and it is possible to limit, but not to totally eliminate, this phenomenon by adjusting the pH value between 10 and 12.^46,49 However, this implies a pH adjustment as an additional step and, as a consequence, an increase in the treating costs and durations. Álvarez et al. noted that a fresh copper oxide-based catalyst deposited on activated carbon could remain stable for phenol concentrations lower than 1.5 g L⁻¹. Nevertheless, the same authors observed copper leaching during the reuse of this catalyst.⁴⁹ For their part, Fortuny et al. also concluded that a basic medium made it possible to limit the dissolution of the active components of transition metal-based oxide catalysts and, therefore, to obtain better performances in a trickle-bed reactor.⁴⁶

Several strategies were developed in an attempt to stabilize the active components (Co, Cu, Fe, Mnetc.) of the heterogeneous catalysts typically employed in redox reactions. For example, Massa et al. recently coated a CuO/γ-Al₂O₃ catalyst with hydrophobic layers, which significantly minimized the solubilisation of the copper species.⁵⁰ As reviewed by Arends and Sheldon, other strategies include framework-substituted molecular sieves, amorphous mixed oxides by grafting or sol–gel methods, grafting or tethering to the inner walls of mesoporous molecular sieves, encapsulation by ship-in-a-bottle or other techniques and ion exchange in layered double hydroxides.⁵¹ Nevertheless, as concluded by the authors, many of these systems are, in reality, unstable towards leaching. In fact, as already mentioned, potential-pH diagrams clearly reveal that these elements will, sooner or later, leach. Consequently, catalyst structure modifications can at the most delay this phenomenon. The use of noble metal-based catalysts thus remains, despite the high cost of the latter, the best way to prevent leaching from occurring.

Sintering

Sintering corresponds to a loss of the active surface consecutively to the growth of the support or metal particles. Catalysis being a surface phenomenon, sintering usually leads to deactivation. It can easily be followed by X-ray diffraction and transmission electron microscopy, as well as H₂ or CO chemisorption. It is rarely observed below 500 °C in gas phase reactions and is not very likely to occur during liquid phase reactions, which are usually performed at lower temperatures. However, in the case of WAO, the presence of hot water can be an aggravating factor. It is indeed well-known that water vapour is able to accelerate some alterations and this is why it is commonly employed in order to artificially age catalysts.^52,53 Besides, in aqueous media, particle growth can even occur at room temperature.^54–56

Sintering can easily be comprehended in the case of supported metal catalysts, if considering that the relief of the support is not perfectly smooth but is constituted of “valleys” and “tops”. The atoms located on “top” positions will tend to migrate towards more stable “valley” positions. As reminded by Moulijn et al.,⁵⁷ this phenomenon results from the increase of atom or particle mobility with temperature and is closely connected to melting points. The so-called Hüttig (eqn (1)) and Tammann (eqn (2)) temperatures are given by the following formulas, in which the temperatures are expressed as Kelvin:

T_Hüttig = 0.3·T_Melting (1)

T_Tammann = 0.5·T_Melting (2)

It appears logical that, as temperature increases and gets closer to the melting point, atom mobility increases and, consequently, sintering gets more important. When the Hüttig temperature is reached, the mobility of the atoms located on structure defects increases. At the Tammann temperature, the bulk atoms themselves become mobile. It is also interesting to note that, in the case of small particles, sintering can occur at temperatures lower than those predicted by T_Hüttig and T_Tammann. The melting, Tammann and Hüttig temperatures are given for some metals and oxides in Table 1.

Table 1 The melting, Tammann and Hüttig temperatures for various metals and oxides

Compound	T Melting (°C)	T Tammann (°C)	T Hüttig (°C)
Au	1065	396	128
Cu	1085	406	134
CuO	1200	464	169
Pt	1755	741	335
Ru	2450	1089	544
CeO2	2400	1064	529
TiO2	1845	786	362
ZrO2	2900	1314	679

---

These data show that reduced platinum and ruthenium, as well as common WAO catalytic supports such as CeO₂, TiO₂ and ZrO₂, should be resistant to sintering in typical WAO conditions. On the contrary, gold, copper and copper oxide appear to be more sensitive to this deactivation mode. In the case of WAO, it is also to be mentioned that the oxidizing medium may play a role in sintering. Several authors have, indeed, reported an aggravation of the sintering of platinum in an oxidizing atmosphere^58,59 and the implication, in this process, of the metal oxide species was proposed in 1976 by Fiedorow et al.⁶⁰ The existence of strong metal–support interactions may make catalysts more stable towards this deactivation mode.⁶¹

Experimentally, the resistance of WAO catalysts to sintering is only rarely verified. Pintar et al. as well as Béziat et al. did not detect any variation in the dispersion of the ruthenium-based catalysts employed at 190 °C in the WAO of an industrial effluent⁴² and of succinic acid solutions,²⁷ respectively. To our knowledge, Besson et al. are the only authors who reported deactivation of WAO catalysts by sintering, they concluded that the decrease in activity of a 2.2% Au/TiO₂ catalyst tested in the WAO of succinic acid was caused by the growth of gold particles, for which the average size increased from 1.7 nm to 4.2 nm during the reaction.⁶² This observation is in agreement with the sensitivity of gold particles towards sintering, mentioned earlier.

The increase in particle size is, a priori, responsible for a permanent deactivation. Nevertheless, supported metal phases can be redispersed by means of specific strategies. This is, for example, the case for γ-Al₂O₃ supported Pt catalysts employed in reforming. Because of their significant price and of their widespread use, a regeneration process is frequently carried out. It consists of treating the deactivated catalyst in the presence of a highly oxidizing gas, at elevated temperature (400–700 °C).⁶³ Oxidized noble metal species become mobile, since they are partially vaporized in the gas phase and since they can migrate on the catalytic surface. Re-adsorption of these species on the support active sites makes it possible to improve the metal dispersion.

Unfortunately, such procedures require time and energy and, consequently, constitute a shortfall for manufacturers. The best strategy therefore consists in limiting particle sintering. This can be achieved by lowering the severity of the operating conditions. The addition of appropriate promoters in the catalytic formulation can also be considered as an efficient tactic. The addition of Zr can, for example, increase the thermal stability of ceria.⁶⁴

Fouling

The second major cause of deactivation of WAO solid catalysts, after leaching, is the formation of a carbonaceous deposit limiting surface accessibility. The formation of this deposit occurs simultaneously as the mineralization of organic pollutants takes place and is highly favoured in the presence of unsaturated compounds, which can easily polymerize (in particular, aromatic compounds, such as phenol). Indeed, Béziat el al., who tested a 2.8% Ru/TiO₂ catalyst in the Wet Air Oxidation of succinic and acetic acids, between 180 and 200 °C, did not notice the formation of such a deposit.²⁷ On the contrary, Belkacemi et al. attributed the deactivation of several catalysts employed in the wet oxidation of high-strength alcohol-distillery waste liquors to coke laydown.⁶⁵ Renard et al. also observed the formation of carbonaceous deposits in the presence of stearic acid.⁶⁶

In the case of the WAO of phenol, although the formation of this inhibiting adsorbed phase is frequently mentioned,^67–69 very few studies have been aimed at determining its structure and its composition. A few authors verified that the catalyst actually deactivated by performing successive reaction runs with the same sample,^11,40,44,70 but such an approach was rarely systematic. Moreover, even when such experiments were carried out, the influence of polymeric species on the catalyst behaviour was rarely discussed in detail. It is also to be noticed that, in some particular cases, the formation of an adsorbed polymeric phase did not necessarily lead to deactivation. Quintanilla et al. attributed the non deactivation of a fouled Fe-based catalyst supported on activated carbon to the fact that the deposit was blocking the access to micropores, while the reaction was mainly taking place in meso and macropores.²³ Lee et al. suggested that, in the case of Al₂O₃- or CeO₂-supported Pt catalysts, carbonaceous deposits were primarily formed on the surface of the Pt particles and continuously migrated onto the support surface in the vicinity of the Pt particles.⁷¹

Deactivation by fouling affects both the oxides of transition metals and supported noble metals.^16,20,35,40 In some cases, the presence of the adsorbed species is clearly visible and manifests itself by a darkening of the catalyst.^72,73 Delgado et al.⁷⁴, as well as Hamoudi et al.,⁴⁴ respectively, observed this carbonaceous deposit by transmission electron microscopy and scanning electron microscopy. It appears that this deposit is in the form of a film, which uniformly covers the catalyst grains. TEM images obtained by Keav et al. for fresh and used 2.5% Pt/CeO₂ catalysts are given in Fig. 3.⁷⁵ Chen et al. compared EDS spectra obtained from the same catalyst before and after phenol WAO experiments and noticed the apparition of a C element corresponding peak.⁴⁰ According to Pintar and Levec, the adsorbed species are not soluble into conventional organic solvents.¹¹ They make the catalyst surface more irregular.³⁰ Kim and Ihm also noted that the deposit was not compact and had its own micropores.⁷³

Fig. 3 TEM images of fresh (a) and spent (b) 2.5% Pt/CeO₂ catalysts.⁷⁵

The extent to which polymeric species are deposited can sometimes be significant. For example, the remarkable performance demonstrated by certain catalysts in the elimination of phenol may, for the most part, be due to their ability to convert this organic pollutant into adsorbed compounds, depending on the experimental conditions. This could be observed by Delgado et al.⁷⁴ as well as Hamoudi et al.⁷⁶ in the case of MnO₂–CeO₂ oxide-based catalysts. Pintar and Levec determined that, after 60 min of reaction, in the presence of a Cu-, Zn- and Al-based oxide catalyst, 47% of the initial carbon was in the form of an adsorbed phase.¹¹ Besides, these authors concluded that a fraction of the polymerization products was adsorbed on the reactor walls and could not be quantified. Hamoudi et al. measured by XPS that, after reaction, the amounts of surface carbon of Pt/γ-Al₂O₃ and MnO₂–CeO₂ catalysts, respectively, increased from 0.14 to 43% and from 2.5 to 50%.³⁷ These authors also concluded that the carbonaceous overlayer was mainly located on the support phase and that, by covering the active sites, it was responsible for the decrease in activity. They determined that, after reaction, the amount of surface platinum decreased by a factor of 5, for the first catalyst and that the surface Mn and Ce amounts, respectively, decreased from 35 to 1.7% and from 9 to 0.5%, for the second one.

Since phenol oxidation occurs through a free-radical mechanism,^11,72 the formation of a high molecular weight species, to which such mechanisms easily lead, is not surprising. Keav et al. investigated the nature of these adsorbed organic molecules. As shown in Fig. 4, they extracted and identified aromatic polycyclic compounds from the chromenone, xanthone and fluorenone families.⁷⁵ Although of different structures, aromatic dimers and trimers such as hydroxylated biphenyls, diphenyl ethers, dibenzofurans and dibenzodioxins were identified during the phenol liquid-phase elimination by (i) molecular oxygen in supercritical conditions,⁷⁷ (ii) Fenton reaction⁷⁸ and (iii) adsorption over activated carbon in an oxidizing medium.⁷⁹ The formation of these species was attributed to oxidative coupling reactions. According to S-SIMS analyzes carried out by Hamoudi et al., the carbonaceous deposit would be constituted by low polycondensation products, evaluated to approximately 4 condensed aromatic rings.⁸⁰ Oliviero et al. also observed the formation of condensation products during the CWAO of aniline.⁸¹

Fig. 4 Gas chromatogram of the extracted adsorbed compounds from a spent phenol WAO 1.25% Ru/CeO₂ catalyst.⁷⁵

The amount of carbonaceous deposit varies with reaction time.^80,82 Its composition is also dependent on the nature of the catalyst⁷³ and on operating conditions.^74,80 The formation of this deposit appears to be favoured at temperatures below 150 °C and at high pollutant concentrations.^5,37,74 Hamoudi et al. observed that, in the case of phenol oxidation, the amount of carbon adsorbed on a MnO₂–CeO₂ catalyst increased with temperature, in the 80–130 °C interval.⁸² In other articles dealing with a similar catalyst, the same authors concluded from XPS analyzes that the deposit preferentially formed on the surface cerium atoms rather than on manganese, thus confirming the importance of the catalyst components in this deactivation pathway.^44,80 The adsorbed phase is constituted, on the one hand, by strongly adsorbed intermediate species issued from phenol degradation³² and, on the other hand, by high molecular weight compounds resulting from polymerization reactions.^11,28 According to the NMR ¹³C CPMAS analyzes carried out by Pintar and Levec on a Cu-, Zn- and Al-based oxide catalyst tested in the WAO of phenol, and as presented in Fig. 5, the adsorbed phase is formed by two reactions carried out in the liquid phase: (i) stepwise addition polymerization of glyoxal to phenol and (ii) polymerization of glyoxal.¹¹ Furthermore, the involvement in the polymerization reaction of intermediate species, formed during phenol degradation, is supported by the formation of a carbonaceous deposit even when pure p-benzoquinone or glyoxal are oxidized.

Fig. 5 Phenol oxidation pathway suggested by Pintar and Levec.¹¹

The changes in activity induced by catalyst oxidation, sintering and leaching are hardly predictable. On the contrary, the loss of activity associated with surface coverage by adsorbed species can be kinetically modelled. Hamoudi et al. suggested a reaction scheme taking into account the adsorption and desorption steps over catalytic sites (*) as well as the deposition of the polymeric species (Fig. 6).^76,82. This network takes into account 4 categories of compounds: (A) the initial reactant, (B) organic by-products, (C) carbon dioxide and (W) polymeric species. This model is based on several assumptions: (i) surface reactions are rate controlling, (ii) all active sites are identical, (iii) deactivation occurs via parallel fouling reaction and an identical deposit is formed by the two main reactions of the network, (iv) the foulant adsorbs irreversibly on the catalyst and its rate of desorption is negligible, (v) the deactivation function is defined as the fraction of sites remaining active.

Fig. 6 WAO reaction scheme suggested by Hamoudi et al.^76,82 (*) Adsorption sites, (A) Initial reactant, (B) Organic by-products, (C) CO₂, (W) Polymeric species.

Carbon deposits are not necessarily of an organic nature. Mikulová et al. concluded that during the WAO of acetic acid by noble metal catalysts supported on cerium-, zirconium- and/or praseodymium-based oxides, the formation of Ce(CO₃)(OH) and Pr(CO₃)(OH) hydroxycarbonates species was responsible for the catalyst deactivation.^83–86 More recently, Liet al. observed an identical behaviour of similar catalysts tested in the WAO of 2-chlorophenol.⁸⁷ According to Nousir et al., in the case of phenol oxidation, the deactivation of Pt catalysts supported on cerium-based oxides is due to different forms of carbon deposits, namely carbonate and polymeric species.⁸⁸

Indeed, it is known that the formation of carbonate species can take place on noble metal catalysts supported on ceria and doped ceria, during processes other than WAO. For example, during the Water Gas Shift reaction, several authors have mentioned that the formation of formates,⁸⁹carbonates^89–91 or hydroxycarbonates^91,92 was brought about by the presence of CO and/or CO₂ and that these species covered both the support active sites and the metal particles. Therefore, they were suspected to be responsible for the catalyst deactivation, although this theory remains a controversial issue. In some cases, a correlation between the amount of carbonate species and the deactivation level⁹⁰ could be established and the initial activity could be totally recovered by calcinations.^89,90,93 Other authors could not find such correlations⁹¹ or simply did not detect carbonate species. In those cases, the catalyst deactivation was attributed to ceria over-reduction⁹⁴ or to the sintering of metal⁹¹ or support particles.⁹⁵

Numerous strategies can be employed for the elimination of the adsorbed species. The degradation of the deposit can be achieved by oxidation in the presence of an oxidizing agent (oxygen,⁹⁶ozone⁹⁷ or nitrous oxide⁹⁸), hydrocracking under hydrogen pressure⁹⁹ and extraction by liquid solvents⁹⁷ or supercritical fluids.¹⁰⁰ Several examples of regeneration treatments, as well as their efficiency, are shown in Table 2. In order to facilitate the comparison between these different processes, the reactant flow rates were expressed as volumic flow rates Q (in mL min⁻¹). In the cases where data given by the authors were insufficient to calculate Q, Weight Hourly Space Velocities (WHSV, in grams of reactant per gram of catalyst per hour) were indicated.

Table 2 Examples of procedures for the elimination of the carbonaceous deposit

Ref.	Catalyst	%C (%)	Conditions of regeneration	Carbon deposit removal	Activity recovery
Combustion processes
98	10% WO3/γ-Al2O3	18.0	O2, WHSV = 0.4 h^−1, T = 400 °C, t = 4 h	99% for the 1st regeneration; much lower for the 2nd one	Almost total for the 1st regeneration; incomplete for the 2nd one
			N2O, WHSV = 0.4 h^−1, T = 400 °C, t = 10 h
101	HZY zeolite	15.0–20.0	O2, Q = 500 mL min^−1, T = 450 °C, t = 6 h	Incomplete	Almost total over several regeneration cycles
			5% O3/O2, Q = 320 mL min^−1, T = 200 °C, t = 4 h	Incomplete	Limited
96	8.4% Cu/γ-Al2O3	2.3	21% O2/CO2, Q = 2800 mL min^−1, T = 800 °C, t = 10 min	100%	Total for stainless steel reactor; Between 55 and 95% for alumina reactor
			5% O2/CO2, Q = 2800 mL min^−1, T = 800 °C, t = 40 min
			Air, Q = 2800 mL min^−1, T = 800 °C, t = 10 min
102	HZSM-5 zeolite	3.5	1.7% O2/He, Q = 60 mL min^−1, T = 450 °C, t = 10 h	57%	92%
			2% N2O/He, Q = 60 mL min^−1, T = 450 °C, t = 10 h	71%	Total
97	2.5% Pd-20% H3PW12O40/SiO2	Unknown	Air, Q = 50 mL min^−1, T = 350 °C, t = 2 h	Not measured	Almost total; better performances if combustion followed by a reduction step
	20% H3PW12O40/SiO2	≈5.3	6% O3/O2, Q = 320 mL min^−1, T = 150 °C, t = 5 h	90.5%	Almost total
104	HZSM-5 zeolite	≈7.3	Air, atmospheric pressure (static conditions), T = 600 °C, t = 12 h	Not measured	Almost total over the 1st three regeneration cycles; incomplete for the following ones
103	NiB/SiO2	Unknown	5% Air/N2, Q = 200 mL min^−1, T = 350 °C, until no more CO2 is produced	Not measured	Total if combustion followed by a reduction step; nonexistent otherwise
Reduction processes
99	0.4% Pt/La-Y	3.0	H2, Q = 40 mL min^−1, P = 15 bar, T = 300 °C, t = 7.75 h	93%	Total over two regeneration cycles
103	NiB/SiO2	Unknown	H2, N2 or 10% H2/N2, Q = 200 mL min^−1, T = 250 °C, t = 2 h	Very low	Between 10 and 20%
Extraction processes
97	20% H3PW12O40/SiO2	≈5.3	Refluxing dichloromethane, t = 3 h	19%	Non considered
103	NiB/SiO2	Unknown	Alkaline ethanol solution (KOH, 0.01 mol L^−1), Q = 20 mL min^−1, room temperature, until rinsing solution turns colourless	Not measured	Total over three regeneration cycles
100	USY zeolite	Unknown	Supercritical fluids (propane, n-butane, i-butane, n-pentane and i-pentane), Q = 2.2–5.3 mL min^−1, P = 56–206 bar T = 130–210 °C, t = 2 h	Not measured	Between 10 and 82%, depending on the employed fluid

---

Elimination of the carbonaceous deposit is, in most cases, carried out by combustion. Various oxidizing agents can be used. Ozone and nitrous oxide are stronger, but also more expensive, oxidizers than oxygen, which, generally, make it possible to lower the regeneration temperature and duration. Nevertheless, unlike oxygen, O₃ and N₂O do not always lead to the total elimination of the adsorbed compounds and some contradictory results suggest that the efficiency of the oxidizing agent also depends on the chemical structure of the deposit, as well as on the catalyst nature and morphology.^98,101,102 Reactivation by oxidation in the presence of oxygen is recognized for its efficiency, its easiness to perform and its speed.^96,98 However, because of the exothermic nature of the combustion reaction, particular attention must be given to regeneration conditions: the oxygen must be diluted with an inert gas while the temperature must be carefully controlled, in order to prevent overheating and any subsequent catalyst thermal degradation. Indeed, Ammendola et al. concluded that the heat capacity of the regenerating fluid, as well as the thermal conductivity of the reactor employed for reactivation, were key factors to be taken into account in order to make the heat exchanges easier and to avoid overheating.⁹⁶ Finally, depending on the nature of the catalyst, a reduction step may be desirable,⁹⁷ if not indispensable,¹⁰³ to activity recovery.

Reactivation by hydrocracking has to be performed under high hydrogen pressure,⁹⁹ otherwise it is almost inefficient.¹⁰³ Besides, hydrogen is more expensive than oxygen or air and its use leads to an increase in operating costs. Eliminating the adsorbed phase through extraction by an organic solvent implies that the deposit has to be totally soluble in the latter. Since extraction yields rarely reach 100%, reactivation is often incomplete.^97,100 Additionally, WAO being a pollution removal process, regeneration procedures involving the use of potentially environment-unfriendly organic solvents, do not constitute the best option.

The total elimination of the adsorbed carbon is not always indispensable to the complete recovery of the initial activity.^101,102 This can be explained by the existence of several types of adsorption sites and/or several types of adsorbed species, among which only some are involved in the deactivation phenomenon.

Particular attention must be given to the catalyst long-term resistance towards reactivation conditions. Although, in some cases, catalytic performances can be maintained throughout several reaction-reactivation cycles,^99,103 the activity recovery can sometimes dramatically drop after successive regenerations.^98,104

Few studies have been focused on WAO spent catalyst reactivation. Chen et al. experimented with the regeneration of a spent copper-promoted CeO₂/γ-Al₂O₃ catalystviaacetone, HCl or HNO₃ solution rinsing.⁴⁰ Only acetone rinsing was revealed to be efficient but two consecutive reaction-reactivation cycles led to a much smaller activity. The authors attributed these modest performances to residual carbon deposits and to the extended leaching of the metal species in the second run. Liu and Sun concluded that the activity of a fouled Fe₂O₃–CeO₂–TiO₂/γ-Al₂O₃ catalyst used in the CWAO of an azo dye, methyl orange, could be efficiently restored by HCl rinsing followed by calcination.¹⁰⁵ Other regeneration procedures consisting of rinsing by alkaline or acid solutions have been patented.^106,107 Pintar and Levec verified that the catalyst could be regenerated by burning out the polymeric products, but the study of CWAO catalyst regeneration was not the primary objective of their work and no further details were given.¹¹ Several authors performed temperature-programmed oxidation analyzes of phenol WAO spent catalysts. The temperature intervals in which the adsorbed carbonaceous materials are converted into carbon dioxide are listed in Table 3. Combustion takes place between 150 and 550 °C, depending on the catalyst nature and a complete regeneration can therefore be envisaged within this temperature range. Massa et al. confirmed that such a regeneration process could also be carried out under air at 400 °C.³⁵ Keav et al. demonstrated that it was possible to totally reactivate a spent 2.5% Pt/CeO₂ catalyst by an ex situ combustion followed by a reduction step and that this treatment could be repeated 3 times without any loss of activity.⁷⁰ It goes without saying that, by virtue of their nature, carbon-based catalysts cannot be reactivated by such oxidizing treatments.

Table 3 Combustion temperatures (T_Comb) of carbonaceous deposits formed on various catalysts

References	Catalyst	T Comb (°C)
30,37,44,74,80	MnO2–CeO2	200–300
30,44	1% Pt/MnO2–CeO2	150–300
76	PtxAg1−x/MnO2–CeO2	200–300
108	K–Mn–Ce–O	250–400
45	2% Ru/ZrO2–CeO2	100–350
30,37	1% Pt/Al2O3	200–550
73	5% Cu/γ-Al2O3	200–500
	5% Fe/γ-Al2O3	250–550
	5% Ni/γ-Al2O3	150–500
	5% Co/γ-Al2O3	300–550
	5% Mn/γ-Al2O3	250–550

---

As in the cases of leaching and sintering, an alternative strategy consists of avoiding the formation of the adsorbed species, rather than in degrading them, by modifying either the operating conditions or the nature and the structure of the catalyst.

The type of reactor employed to carry out the WAO process has an influence on the formation of polymers. The latter appears to be favoured in batch and semi-batch reactors. Indeed, Pintar and Levec,¹⁰⁹ Alejandre et al.¹¹⁰ as well as Fortuny et al.¹¹¹ observed that no polymeric deposit was formed during phenol WAO in a trickle-bed reactor. The first authors explained this phenomenon based on the hypothesis that the polymerization reaction mainly takes place in the aqueous phase. The higher “solid catalyst-to-liquid” ratios met in trickle-bed reactors improve the accessibility of the catalytic surface and make the oxidation of intermediate species easier by limiting their participation in polymerization reactions. Finally, Pintar et al. recently observed, during Ru/TiO₂ catalyzed WAO of phenol in a trickle-bed reactor, that the accumulation of benzoquinone and hydroquinone intermediates could also be responsible for deactivation and that the use of temperatures above 200 °C could prevent this phenomenon from happening.³²

Hussain et al. synthesized a potassium-doped MnO₂–CeO₂ catalyst resistant to the accumulation of adsorbed species. After WAO of phenol, its measured carbon content was 14 times smaller than that of a non-promoted MnO₂–CeO₂ catalyst.^108,112 It seems that K-doping inhibits the formation of carbon deposits and therefore prevents the fouling of active sites, by favouring CO₂ selectivity. The catalytic activity remained constant during two successive WAO runs without regeneration. The authors explained these performances by the electro-donating ability of K₂O, which favours the activation of oxygen and its conversion into peroxides. These very active species promote the deep oxidation of organic molecules, which explains why less carbonaceous deposits are formed. Potassium leaching detected during this study was very low (≤1.2 ppm), which can be explained by the low temperatures (110 °C) as well as the short reaction times (10–20 min) employed in those experiments.

Conclusion

Wet Air Oxidation is a promising and efficient water pollution removal process. It can be employed to treat liquid effluents from various origins, with Chemical Oxygen Demands typically between 5 and 100 g L⁻¹. Several types of catalysts are available to increase the performances of the process.

Homogeneous catalysts are dissolved into the reaction medium and their use therefore implies an additional precipitation and recovery step. Heterogeneous catalysts are active and can easily be separated from the treated effluent. However, these catalysts can be deactivated in WAO operating conditions.

The deactivation of WAO solid catalysts is mainly associated with four phenomena. The oxidizing medium met in this process can be involved in various catalyst structure modifications. Leaching, directly or indirectly, leads to a loss of catalytic components while sintering results in a decrease in active surface. Both of these are generally responsible for irreversible alterations. Fouling limits the accessibility of active sites, but activity can be recovered since the adsorbed phase can be eliminated.

Currently, most of the studies are focused on activity and selectivity issues. Although capital, these qualities are not sufficient to make a catalytic process viable and too few studies have been devoted to the deactivation phenomena and to reactivation possibilities.

This observation is explained by the difficulty to study deactivation. Indeed, several phenomena can simultaneously be involved and may even be interlinked. For example, Masende et al. concluded that there was a connection between the superficial oxidation of the supported metal particle and the formation of the polymers responsible for fouling.²⁸ Also, as already mentioned, leaching of the support phase can lead to a loss of supported particles. The second difficulty met in the study of deactivation phenomena is that the latter depend on numerous factors and particularly on the acidity of the medium, the temperature, oxygen partial pressure and pollutant concentration (Fig. 7). Of course, catalyst nature and concentration should also be considered. The reactant flow rates must be taken into account in the case of continuous reactors. The acidity of the effluent can be modified through the addition of acidic or basic reactants. However, this implies additional pH-adjustment steps for the incoming and treated effluents and, as a consequence, increases in operating costs and durations. Although this parameter can be modified in fundamental studies, since the organic pollutant concentration is characteristic of the incoming effluent, it is difficult to change in the case of real effluents. Nevertheless, the dilution of the polluted effluent by the treated one may be envisaged when concentrations are too high. It results from these considerations that amongst the operating parameters, only temperature, pressure and catalyst concentration can easily be modified.

Fig. 7 Summary of deactivation modes and involved operating parameters.

More studies should aim at identifying the WAO catalyst deactivation modes and mechanisms in order to determine the optimal operating conditions limiting activity losses. The selection of these conditions must be a compromise between activity and long-term stability. In the case where the optimal operating conditions do not make it possible to totally avoid deactivation, efficient regeneration procedures must be developed.

As a conclusion, very few data are available concerning the stability of real catalysts employed in the treatment of real effluents in real WAO process conditions. Future research should be directed to the study of shaped catalyst mechanical resistance and of catalyst poisoning by halogen-, sulphur- or phosphorus-containing compounds, which are potentially present as traces in the incoming effluents.

References

1. R. Andreozzi, V. Caprio, A. Insola and R. Marotta, Catal. Today, 1999, 53, 51.
2. R.W. Strehlenert and H. Kiaer & Co. Ltd., Swedish Patent SE 34 941, 1911.
3. R.W. Strehlenert and H. Kiaer & Co. Ltd., American Patent US 1 149 420, 1915.
4. F. Luck, Catal. Today, 1999, 53, 81.
5. V. S. Mishra, V. V. Mahajani and J. B. Joshi, Ind. Eng. Chem. Res., 1995, 34, 2.
6. A. Cybulski, Ind. Eng. Chem. Res., 2007, 46, 4007.
7. L. Li, P. Chen and E. F. Gloyna, AIChE J., 1991, 37, 1687.
8. S. Imamura, Y. Tonomura, M. Terada and T. Kitao, Mizu Shori Gigutsu, 1979, 20, 317.
9. W. J. Fisher, Water Res., 1971, 5, 187.
10. S. Imamura, Ind. Eng. Chem. Res., 1999, 38, 1743.
11. A. Pintar and J. Levec, J. Catal., 1992, 135, 345.
12. H. Ohta, S. Goto and H. Teshima, Ind. Eng. Chem. Fundam., 1980, 19, 180.
13. S. Imamura, M. Nakamura, N. Kawabata, J. Yoshida and S. Ishida, Ind. Eng. Chem. Prod. Res. Dev., 1986, 25, 34.
14. P. M. Álvarez, D. McLurgh and P. Plucinski, Ind. Eng. Chem. Res., 2002, 41, 2147.
15. D. Duprez, F. Delanoë, J. Barbier Jr., P. Isnard and G. Blanchard, Catal. Today, 1996, 29, 317.
16. F. Stüber, J. Font, A. Fortuny, C. Bengoa, A. Eftaxias and A. Fabregat, Top. Catal., 2005, 33, 3.
17. A. Cybulski and J. Trawczynski, Appl. Catal., B, 2004, 47, 1.
18. S. Imamura, I. Fukuda and S. Ishida, Ind. Eng. Chem. Res., 1988, 27, 718.
19. J. Trawczynski, Carbon, 2003, 41, 1515.
20. J. Barbier Jr., L. Oliviero, B. Renard and D. Duprez, Top. Catal., 2005, 33, 77.
21. M. Besson and P. Gallezot, Catal. Today, 2003, 81, 547.
22. P. Dufresne, Appl. Catal., A, 2007, 322, 67.
23. A. Quintanilla, N. Menéndez, J. Tornero, J. A. Casas and J. J. Rodríguez, Appl. Catal., B, 2008, 81, 105.
24. C. C. Shih and J. R. Chang, J. Catal., 2006, 240, 137.
25. D. P. Minh, P. Gallezot and M. Besson, Appl. Catal., B, 2007, 75, 71.
26. J. C. Béziat, M. Besson, P. Gallezot and S. Durécu, Ind. Eng. Chem. Res., 1999, 38, 1310.
27. J. C. Béziat, M. Besson, P. Gallezot and S. Durécu, J. Catal., 1999, 182, 129.
28. Z. P. G. Masende, B. F. M. Kuster, K. J. Ptasinski, F. J. J. G. Janssen, J. H. Y. Katima and J. C. Schouten, Catal. Today, 2003, 79–80, 357.
29. M. Pourbaix, Atlas d’équilibres électrochimiques à 25 °C, Gauthier-Villars & Cie Editeur, Paris, 1963.
30. F. Larachi, K. Belkacemi, S. Hamoudi and A. Sayari, Catal. Today, 2001, 64, 163.
31. H. Takayama, Q. Jiang-Yan, K. Inazu and K. Aika, Chem. Lett., 1999, 28, 377.
32. A. Pintar, J. Batista and T. Tišler, Appl. Catal., B, 2008, 84, 30.
33. F. Delanoë, Thesis of the University of Poitiers, 1996.
34. J. Barbier Jr., F. Delanoë, F. Jabouille, D. Duprez, G. Blanchard and P. Isnard, J. Catal., 1998, 177, 378.
35. P. Massa, F. Ivorra, P. Haure, F. M. Cabello and R. Fenoglio, Catal. Commun., 2007, 8, 424.
36. Q. Zhang and K. T. Chuang, Can. J. Chem. Eng., 1999, 77, 399.
37. S. Hamoudi, F. Larachi and A. Sayari, J. Catal., 1998, 177, 247.
38. S. S. Lin, C. L. Chen, D. J. Chang and C. C. Chen, Water Res., 2002, 36, 3009.
39. A. Santos, P. Yustos, A. Quintanilla and F. García-Ochoa, Appl. Catal., B, 2004, 53, 181.
40. I. P. Chen, S. S. Lin, C. H. Wang and S. H. Chang, Chemosphere, 2007, 66, 172.
41. R. Kouraichi, J. J. Delgado, J. D. López-Castro, M. Stitou, J. M. Rodríguez-Izquierdo and M. A. Cauqui, Catal. Today, 2010, 154, 195.
42. A. Pintar, M. Besson and P. Gallezot, Appl. Catal., B, 2001, 30, 123.
43. Q. Zhang and K. T. Chuang, Environ. Sci. Technol., 1999, 33, 3641.
44. S. Hamoudi and F. Larachi, Ind. Eng. Chem. Res., 1998, 37, 3561.
45. J. Wang, W. Zhu, S. Yang, W. Wang and Y. Zhou, Appl. Catal., B, 2008, 78, 30.
46. A. Fortuny, C. Bengoa, J. Font and A. Fabregat, J. Hazard. Mater., 1999, 64, 181.
47. N. Grosjean, C. Descorme and M. Besson, Appl. Catal., B, 2010, 97, 276.
48. A. Santos, P. Yustos, A. Quintanilla, G. Ruiz and F. Garcia-Ochoa, Appl. Catal., B, 2005, 61, 323.
49. P. M. Álvarez, D. McLurgh and P. Plucinski, Ind. Eng. Chem. Res., 2002, 41, 2153.
50. P. Massa, F. Ivorra, P. Haure and R. Fenoglio, Catal. Commun., 2009, 10, 1706.
51. I. W. C. E. Arends and R. A. Shelton, Appl. Catal., A, 2001, 212, 175.
52. S. J. Schmieg and D. N. Belton, Appl. Catal., B, 1995, 6, 127.
53. S. Balcon, S. Mary, C. Kappenstein and E. Gengembre, Appl. Catal., A, 2000, 196, 179.
54. A. Douidah, P. Marécot, S. Labruquère and J. Barbier, Appl. Catal., A, 2001, 210, 111.
55. A. Garron, K. Lázár and F. Epron, Appl. Catal., B, 2005, 59, 57.
56. I. Dodouche, D. Pereira Barbosa, M. do Carmo Rangel and F. Epron, Appl. Catal., B, 2009, 93, 50.
57. J. A. Moulijn, A. E. Van Diepen and F. Kapteijn, Appl. Catal., A, 2001, 212, 3.
58. F. H. Huang and C. H. Li, Scr. Metall., 1973, 7, 1239.
59. P. Wynblatt and N. A. Gjostein, Prog. Solid State Chem., 1975, 9, 21.
60. R. M. J. Fiedorow and S. E. Wanke, J. Catal., 1976, 43, 34.
61. Y. Nagai, T. Hirabayashi, K. Dohmae, N. Takagi, T. Minami, H. Shinjoh and S. Matsumoto, J. Catal., 2006, 242, 103.
62. M. Besson, A. Kallel, P. Gallezot, R. Zanella and C. Louis, Catal. Commun., 2003, 4, 471.
63. M. J. d’Aniello Jr., D. R. Monroe, C. J. Carr and M. H. Krueger, J. Catal., 1988, 109, 407.
64. M. Ozawa, M. Kimura and A. Isogai, J. Alloys Compd., 1993, 193, 73.
65. K. Belkacemi, F. Larachi, S. Hamoudi and A. Sayari, Appl. Catal., A, 2000, 199, 199.
66. B. Renard, J. Barbier Jr., D. Duprez and S. Durécu, Appl. Catal., B, 2005, 55, 1.
67. J. Guo and M. Al-Dahhan, Appl. Catal., A, 2006, 299, 175.
68. D. Goi, C. de Leitenburg, G. Dolcetti and A. Trovarelli, J. Alloys Compd., 2006, 408–412, 1136.
69. K. H. Kim, J. R. Kim and S. K. Ihm, J. Hazard. Mater., 2009, 167, 1158.
70. S. Keav, A. Martin, J. Barbier Jr. and D. Duprez, Catal. Today, 2010, 151, 143.
71. D. K. Lee, D. S. Kim, T. H. Kim, Y. K. Lee, S. E. Jeong, N. T. Le, M. J. Cho and S. N. Henan, Catal. Today, 2010, 154, 244.
72. A. Sadana and J. R. Katzer, Ind. Eng. Chem. Fundam., 1974, 13, 127.
73. S. K. Kim and S. K. Ihm, Top. Catal., 2005, 33, 171.
74. J. J. Delgado, J. A. Pérez-Omil, J. M. Rodríguez-Izquierdo and M. A. Cauqui, Catal. Commun., 2006, 7, 639.
75. S. Keav, A. Martin, J. Barbier Jr. and D. Duprez, C. R. Chim., 2010, 13, 508.
76. S. Hamoudi, K. Belkacemi, L. Bonneviot and F. Larachi, Catal. Today, 2000, 62, 379.
77. T. D. Thornton and P. E. Savage, J. Supercrit. Fluids, 1990, 3, 240.
78. J. Poerschmann and U. Trommler, J. Chromatogr., A, 2009, 1216, 5570.
79. E. Y. Osei-Twum, N. S. Abuzaid and G. Nahkla, Bull. Environ. Contam. Toxicol., 1996, 56, 513.
80. S. Hamoudi, F. Larachi, A. Adnot and A. Sayari, J. Catal., 1999, 185, 333.
81. L. Oliviero, J. Barbier Jr. and D. Duprez, Appl. Catal., B, 2003, 40, 163.
82. S. Hamoudi, K. Belkacemi and F. Larachi, Chem. Eng. Sci., 1999, 54, 3569.
83. J. Mikulová, S. Rossignol, J. Barbier Jr., D. Mesnard, C. Kappenstein and D. Duprez, Appl. Catal., B, 2007, 72, 1.
84. J. Mikulová, S. Rossignol, J. Barbier Jr., D. Duprez and C. Kappenstein, Catal. Today, 2007, 124, 185.
85. J. Mikulová, J. Barbier Jr., S. Rossignol, D. Mesnard, D. Duprez and C. Kappenstein, J. Catal., 2007, 251, 172.
86. J. Gáalová, J. Barbier Jr. and S. Rossignol, J. Hazard. Mater., 2010, 181, 633.
87. N. Li, C. Descorme and M. Besson, Appl. Catal., B, 2008, 80, 237.
88. S. Nousir, S. Keav, J. Barbier Jr., M. Bensitel, R. Brahmi and D. Duprez, Appl. Catal., B, 2008, 84, 723.
89. C. H. Kim and L. T. Thompson, J. Catal., 2005, 230, 66.
90. X. Liu, W. Ruettinger, X. Xu and R. Farrauto, Appl. Catal., B, 2005, 56, 69.
91. W. Ruettinger, X. Liu and R. J. Farrauto, Appl. Catal., B, 2006, 65, 135.
92. Q. Fu, W. Deng, H. Saltsburg and M. Flytzani-Stephanopoulos, Appl. Catal., B, 2005, 56, 57.
93. A. Goguet, F. Meunier, J. P. Breen, R. Burch, M. I. Petch and A. Faur Ghenciu, J. Catal., 2004, 226, 382.
94. J. M. Zalc, V. Sokolovskii and D. G. Löffler, J. Catal., 2002, 206, 169.
95. T. Bunluesin, R. J. Gorte and G. W. Graham, Appl. Catal., B, 1998, 15, 107.
96. P. Ammendola, R. Chirone, G. Ruoppolo and G. Russo, Exp. Therm. Fluid Sci., 2010, 34, 262.
97. I. V. Kozhevnikov, S. Holmes and M. R. H. Siddiqui, Appl. Catal., A, 2001, 214, 47.
98. G. J. Hutchings, R. Hunter and L. Jansen Van Rensburg, Appl. Catal., 1988, 41, 253.
99. R. Klingmann, R. Josl, Y. Traa, R. Gläser and J. Weitkamp, Appl. Catal., A, 2005, 281, 215.
100. D. M. Ginosar, D. N. Thompson and K. C. Burch, Appl. Catal., A, 2004, 262, 223.
101. G. J. Hutchings, R. G. Copperthwaite, T. Themistocleous, G. A. Foulds, A. S. Bielovitch, B. J. Loots, G. Nowitz and P. Van Eck, Appl. Catal., 1987, 34, 153.
102. D. P. Ivanov, V. I. Sobolev and G. I. Panov, Appl. Catal., A, 2003, 241, 113.
103. W. J. Wang, H. X. Li, S. H. Xie, Y. J. Li and J. F. Deng, Appl. Catal., A, 1999, 184, 33.
104. X. Guo, Y. Zheng, B. Zhang and J. Chen, Biomass Bioenergy, 2009, 33, 1469.
105. Y. Liu and D. Sun, Appl. Catal., B, 2007, 72, 205.
106. S. Maeda, M. Ikeda, T. Ishii and K. Mitsui, Nippon Catalytic Chem. Ind., Japanese Patent JP 9 010 602 A, 1997.
107. Y. Harada, M. Futagawa and K. Yamazaki, Osaka Gas Co. Ltd., Japanese Patent JP 2 000 117 109 A, 2000.
108. S. T. Hussain, A. Sayari and F. Larachi, J. Catal., 2001, 201, 153.
109. A. Pintar and J. Levec, Ind. Eng. Chem. Res., 1994, 33, 3070.
110. A. Alejandre, F. Medina, P. Salagre, A. Fabregat and J. E. Sueiras, Appl. Catal., B, 1998, 18, 307.
111. A. Fortuny, C. Bengoa, J. Font, F. Castells and A. Fabregat, Catal. Today, 1999, 53, 107.
112. S. T. Hussain, A. Sayari and F. Larachi, Appl. Catal., B, 2001, 34, 1.

---