2.1 Catalysts preparations

The raw materials used in the preparation of the catalysts were iron nitrate nonahydrate (FeNO₃·9H₂O 99.5%) and palladium (II) nitrate dihydrate (Pd(NO₃)₂·2H₂O, 99%) from Panreac, used as precursors of the iron and palladium oxide active phases and low cost residues from rice production (Rice bran (RB) and rice husk ash (RHA). RB used as a Pore Generating Agent (PGA) when it is burnt out of the structured bodies on heat treatment at 500 °C or higher and RHA employed to modify the pore size distribution of the materials to improve their catalytic activities by increasing their total pore volumes and pore widths but avoiding loss in compressive strength. Also sepiolite Pansil (a low cost magnesium silicate from Tolsa S.A.) was employed as the catalyst support due to its ease of extrusion and ability to form ceramic bodies when calcined above 350 °C.

The catalysts were conformed as extrudates to avoid the pressure drop limitations of powder catalysts in high volume effluent gas treatments. The materials were extruded by mixing different amounts of the precursor solids (Sepiolite, RB and RHA) in deionized water with aqueous solutions of the iron and palladium nitrates in an orbital mixer EIRICH model R-02E, with the required volume (ca. 50 wt% water) to attain adequate rheological properties for the subsequent extrusion of the paste with a Bonnot single screw extruder, as solid cylinders with a 3.5 mm external diameter. These extruded materials were allowed to dry at room temperature for 72 h then dried at 150 °C and subsequently heat treated in air at 500 or 700 °C for 4 h, increasing the temperature at 5 °C/min up to the final value in a muffle furnace. These temperatures were chosen from previous experience of working with sepiolite, which undergoes several well documented morphological changes with heat treatment (Yates et al., 2003). The resulting catalysts were denoted as weight compositions of Sepiolite, iron nitrate, RB and RHA (Sep/Fe nitrate/RB/RHA). For comparison purposes the best iron catalyst was impregnated with palladium (II) nitrate dihydrate and further calcined at 500 °C to decompose the precursor salt.

2.2 Catalysts characterisation

The thermal behaviours of both the raw materials and the conformed composites were analysed to study their stabilities in a TG-DTA Stanton model STA 781, heating them from room temperature to 900 °C at a heating rate of 10 °C/min, in an air flow of 20 ml/min. Based on these TG-DTA analyses, each precursor dough was treated at three temperatures: 150, 500 and 700 °C, to dry, decompose and modify the activities porosities, surface areas and compressive strengths of the final materials.

The elemental analyses of the solids were carried out employing inductively coupled plasma (ICP) in a Perkin-Elmer, model Optima 3300 DV. The crystalline characteristics of the materials were analysed by X-ray diffraction in a Seifert model XRD 3000P in the range 2θ = 5–75°, with a 0.02°/step and acquisition time of 2 sec/step.

Transmission electron microscopy was carried out in a JEOL model FXII electron microscope operating at 200 kV, on solids deposited on carbon covered copper grids.

The textural characteristics of the materials were determined using nitrogen adsorption at 77 K and mercury intrusion porosimetry. The ads-desorption isotherms were carried out in a Micromeritics ASAP 2020 apparatus, on samples that had been previously outgassed to a vacuum of 10^-2 Pa during 16 h, to ensure clean dry surfaces that were free from any loosely bound adsorbed species. For the materials dried at 150 °C the outgassing procedure was carried out at this temperature. The materials heated to 500 °C or higher were outgassed at 300 °C. The volumes and pore size distributions of pores with diameters higher than 7.5 nm were determined by mercury intrusion porosimetry (MIP) on a Fisons Instruments Pascal 140/240 apparatus, using ca. 0.2 g of material that had been previously dried at 150 °C for 16 h. The pressure/intrusion data were analysed by the Washburn equation, taking the recommended values for the mercury contact angle (141°) and surface tension (484 mN/m).

The compressive strengths of the composites were measured on a dynamometer Chatillon model LTMC, by means of applying a pressure on the external surface of the extrudates until breakage, with a test head of 1.32 cm². The rupture pressure (kg cm⁻¹) was taken as the average value of 10 measurements.

Basicity analyses were performed to study the amount and strength of the basic centres on the composites by the desorption/decomposition of acetic acid previously adsorbed onto the materials, using a TG/DTA Stanton 781 coupled to a mass spectrometer Thermostar QMS200 M3 and analysing both the amount and temperature at which CO₂ was produced by the decomposition of the acetic acid on increasing the temperature from room temperature up to 900 °C at 10 °C min⁻¹.

2.3 Toluene oxidation

The catalytic activity for the total oxidation of toluene was carried out in a stainless steel fixed bed reactor (internal diameter 3.5 cm length 85 cm), with silica wool in the inside at half height and a quartz tube of 12 mm internal diameter and 14 mm external diameter. The free space between the quartz tube and the reactor wall was filled with silicon carbide of particle sizes lower than 0.84 mm, to fix the quartz tube in place. The temperature was controlled by means of two thermocouples at different heights inside the reactor. The catalytic activity was measured using the cylindrical extrudes of 1 cm in length. Approximately 50 ml/min air flow was bubbled through toluene maintained at −12 °C in a cryostat, that was then diluted with more air in order to attain a total air flow of 3500 ml/min with 100 ppm of toluene, at a constant gas hour space velocity (GHSV) of 35000 h⁻¹ and a linear velocity of 0.52 m/s for all experiments. The toluene concentrations at the inlet and outlet of the reactor were analysed with a Flame ionization detector (FID, Horiba FIA – 510) and the CO₂ produced on total oxidation was measured in a Horiba VIA-510 Analyser by infrared spectroscopy (IR).

3.1 Characterisation of precursors and catalysts

The catalysts nomenclature is based in amounts by weight of Sep/Fe/RB/RHA, the palladium containing catalyst is denoted as Pd 17/2/2/2. TG-DTA analyses of the calcination of the precursors indicated that the decomposition of rice bran (RB) (Fig. 1a) gave rise to two exotherms at 250 °C and 430 °C due to the combustion of carbohydrates (Martin-Luengo et al., 2011), with a total weight loss close to 90%. Chemical analyses of the remnant material indicated it was silica. The thermal analysis on rice husk ash (RHA) (not shown) did not lead to any significant weight loss, as expected, since it had been prepared by incinerating rice husk waste from rice production at 850 °C in the factorýs cogeneration plant.

Close

Fig. 1

TG-DTA thermograms of catalyst precursors (Sep/FE/RB/RHA): a) 0/0/RB/0b) Fe(NO₃)₂ 6H₂Oc) 17/1/0/0d) 17/0/0/0e) 17/2/2/2f) 17/1/2/2.

Export to PPT

Iron nitrate nonahydrate decomposes below 200 °C, with an endotherm at 80 °C due to loss of physisorbed water, another at ca. 115 °C due to the loss of seven molecules of water of hydration, with the rest of the water molecules and decomposition of nitrates occurring at 195 °C, with a total weight loss of ca. 73% (Fig. 1b).

The thermal behaviours of extruded Fe/Sep (17/1/0/0, Fig. 1c) and sepiolite (17/0/0/0 Fig. 1d) were similar, with an endotherm at 130 °C due to loss of adsorbed water, another at 290 °C due to loss of zeolitic water that occupies the sepiolite channels, at 475 °C loss of crystallization water attached to silicon and magnesium terminal ions of the octahedral layer of sepiolite, causing a collapse of its internal structure, another endothermic peak at T > 500 °C due to loss of constitutional water, which forms part of the structure as hydroxyls, and an exotherm at about 827 °C due to phase transformation of sepiolite to enstatite (Yates et al., 2003). From previous results it is known that treatment of the extrudates at 500 °C leads to the transformation of sepiolite into a stable ceramic material that cannot be rehydrated due to the loss of constitutional water and the folding of the structure at a temperature of approximately 330 °C. With heat treatment at 700 °C the structure gains compressive strength but suffers a reduction in the surface area due to sintering of the material.

The thermograms of the catalysts precursors were a mixture of the decomposition of the various components, as expected, materials 17/2/2/2 (e) and 17/1/2/2 (f), are chosen as examples. According to the TG-DTA results a minimum temperature of 500 °C in air was enough to decompose the various precursors and produce the conformed ceramic catalyst and 700 °C was also chosen in order to check the effect of a higher calcination temperature on the catalytic activities. Although heat-treatment at higher temperatures increases the compressive strengths, this is achieved at the expense of a notable reduction in the specific surface area of the material.

The compositions of the catalysts precursors and percent composition of the catalysts after calcination were analysed by Inductively Coupled Plasma (ICP) (Table 1). Due to its natural origin, sepiolite contains 28% silicon, 12% magnesium, 2.5% calcium, 0.9% potassium, 0.7% iron, 0.7% phosphorous and less than 0.5% for aluminium and sodium. The content of alkaline and alkaline earth cations and iron gives sepiolite its basic character.

X-ray diffraction analysis of the crystallinities of precursors and catalyst 17/2/2/2 treated at different temperatures, chosen as examples are shown in Fig. 2. As expected, catalyst 17/2/2/2 treated at 150 °C gave rise to the characteristic peak for sepiolite at 2θ = 7.2°, related to the microporous channels, which on heating above 350 °C disappeared due to the collapse and folding of the structure. The characteristic main reflexion peaks for iron oxides were not present in the XRD patterns, being for Fe₂O₃ at 35.7° (JCPDS No. 39–1346) and Fe₃O₄ at 34.8° (JCPDS No. 75–0033). Furthermore, the XRD patterns of catalysts 17/2/2/2 and Pd 17/2/2/2 are identical and no PdO oxide is detected at 33.8° (JCPDS No. 41–1107). The results support the good dispersions achieved with this production technique, since crystals with sizes of less than ca. 4–5 nm cannot be detected by XRD. The peaks at 26.5° and 22° were due to quartz and cristobalite present in the sepiolite.

Close

Fig. 2

XRD patterns of Sep/Fe/RB/RHA (17/2/2/2) (a) 150 °C, (b) 500 °C, (c) 700 °C, (d) 850 °C, (e) sepiolite (17/0/0/0), (f) enstatite.

Export to PPT

Analyses of textural properties by nitrogen adsorption–desorption and mercury intrusion porosimetry indicated (Table 1 and Fig. 3) that as the heat treatment temperature was increased there was a reduction in the overall pore volume (V_t) and higher temperatures caused an increase of compressive strengths (see Table 1). When iron was added (17/1/0/0) the overall pore size distributions observed were similar to those of sepiolite (17/0/0/0) extrudates although slightly shifted to wider pores and giving rise to an increase in the total pore volume. For the material that incorporated both iron and RB (17/1/2/0) there was a substantial increase in both the overall pore volume and pore widths, with the formation of a more open porous structure in this incorporated catalyst, that would lead to a reduction in any diffusion limitations and positively affect the catalytic activity (Mazaheri et al., 2018). As shown in Table 1, the palladium containing catalyst (Pd 17/2/2/2) did not show differences in its textural characteristics versus its non-palladium containing counterpart (17/2/2/2).

Close

Fig. 3

MIP of catalysts (Sep/Fe/RB/RHA) calcined at 500 °C.

Export to PPT

The catalyst that contained sepiolite, iron and RHA (17/1/0/2) gave rise to a distinct bimodal cumulative pore volume curve with the usual porosity in pores below about 200 nm due to the sepiolite but an additional porosity in pores close to 1 µm due to the interparticulate porosity between the primary particles of the RHA. The presence of these much wider pores greatly reduces any diffusional limitations of these incorporated catalysts, thus leading to a marked increase in the catalytic activity. The catalysts prepared by calcining at 500 °C, had porosities that benefitted from the advantages of both using the PGA and the RHA: increased overall pore volumes and the presence of much wider macropores, close to 1 µm (Fig. 3).

The effect of the compositions and heat treatments of the structured catalysts on their compressive strengths are shown in Table 1. From these results it was clear that the catalysts with higher compressive strengths were those treated at 700 °C and the weakest those treated at 500 °C. This was to be expected since initial heat treatment at 500 °C causes the degradation of any PGA and a significant increase in the total pore volume of the composite, whilst heat treatments at higher temperatures then leads to sintering of the ceramic material and phase changes which reduces the total pore volumes and improves the compressive strengths (Yates et al., 2003).

Regarding the dimensional stability of the materials, as the heat treatment temperature was increased the exterior dimensions of the conformed materials were reduced due to the corresponding increase in their densities. From these results, high heat treatment temperatures would be the best choice to achieve better compressive strengths and abrasion resistances, but this must be balanced with the loss in specific surface area and consequent loss in catalytic activity.

Analyses of basic oxidation sites were carried out using TG-MS, monitoring the desorption/decomposition of acetic acid previously adsorbed onto the catalysts surfaces forming carbonate, bidentate or bridged acetate species. By recording the intensities and positions of the peaks corresponding to the decomposition of these entities to CO₂ (m/e 44) both the amount and strength of basic sites, closely related to their catalytic oxidation activities were determined. The areas and intensities indicate the amounts of basic centres and their strengths are related to the temperatures at which the bands are observed (Fig. 4) (Yates et al., 2017a).

Close

Fig. 4

Basicity analysis of catalysts after acetic acid adsorption (m/e 44 vs Temperature).

Export to PPT

From previous studies it has been shown that the desorption/decomposition of acetic acid at temperatures below 250 °C were related to the amount of low strength basic sites, while those at 350–550 °C were due to medium and high strength basic centres: Mg-O (and Mg-O-H) and Fe-O (and Fe-O-H) (Yates et al., 2017). Fig. 4 indicates that sepiolite has small amounts of basic centres of low, medium and high strengths, as expected due to its content of alkaline and alkaline earth cations (12% magnesium, 2.5% calcium, 0.9% potassium) and 0.7% iron and it can also be seen that for the catalysts calcined at 500 °C, 17/1/0/0 has similar amount of basic sites to its counterparts that contain RB and RHA (17/1/2/0, 17/1/0/2 and 17/1/2/2). The catalyst with a higher amount of iron (17/2/2/2) has larger amounts of strong basic centres and catalyst 17/1/2/2 showed lower amount of strong basic sites than 17/2/2/2 due to sintering of the active phase. It should also be noted that on increasing the calcination temperature the amount of high strength basic sites decreases, in agreement with the loss in surface area. The palladium containing catalyst (Pd 17/2/2/2) shows similar amount of basic sites as catalyst 17/2/2/2, but it has a greater amount of high strength basic sites (peak at ca 439 °C) due to the displacement of the temperature of strong basic sites to higher values (see Fig. 4).

Transmission electron micrographs of the catalysts were used to recover information on the morphology and particle size of iron oxide catalysts (Fig. 5) showing lower dispersion of the iron phase in the catalyst with higher amount of iron, in agreement with the lower amounts of basic sites found, due to sintering of the iron oxide species and with the lower activities found for the catalyst with higher amount of iron added. The importance of this parameter is related to the oxidation activities (see below).

Close

Fig. 5

Transmission electron micrographs and particle size distributions of catalysts.

Export to PPT

3.2 Toluene oxidation

The catalytic activity tests showed no reaction products other than H₂O and CO₂ under the conditions used in this work. Repeated catalytic runs were reproducible for at least three consecutive reactions, without any appreciable loss in activity, indicating the robustness of the conformed catalysts.

In this work the amounts of strong basic sites measured by acetic acid adsorption–desorption/decomposition (Fig. 4) are related to the order of activities (see Fig. 6 and Table 2), being related to loss of active surface area by agglomeration of the iron oxide particles, by increasing the temperature of calcination or the amount of iron with the consequent loss of activity (Aguayo-Villareal, 2017; Nogueira et al., 2011).

Close

Fig. 6

Catalytic oxidation of toluene (Conversion (%) vs Temperature (°C)).

Export to PPT

Moreover, in agreement with the proposed mechanism, the addition of RB and/or RHA result in improved catalytic activities where the textural analyses indicate that there are increases in both pore width and pore volume of the catalysts on adding these components, i.e. the presence of RHA in the catalysts produces pores of ca. 900 nm, and on adding RB the pore volumes are increased and therefore the molecules of toluene will reach more easily the active phase (Szegedi et al., 2013). In agreement with these, the activities for the catalysts calcined at 500 °C were higher than those of their counterparts calcined at 700 °C (Yates et al., 2017), since, although an increase in calcination temperature improves the compressive strength, the decrease in surface area and pore volume on calcining at 700 °C with the consequent loss of the active basic sites causes a reduction in their catalytic activities.

From the characterisation and activity results, the best catalytic activity for total oxidation of toluene of the iron containing catalysts is due to the favourable combination of the surface properties, and the better dispersion of the active oxide phases (Boycheva et al., 2019; Nogueira et al., 2011).

Judicious choice of components and thermal treatments allow us to achieve catalysts that have activities similar to those found for similar processes in published work, although often those are based on synthetic and/or expensive materials, or using potentially dangerous compounds and are therefore less environmentally acceptable than the green catalysts developed in this work (Markova-Velichkova et al., 2013).

Addition of palladium to the best iron containing catalyst, improves the oxidation of toluene with values similar to other published palladium containing catalysts (Kima et al., 2015) (Table 2 and Fig. 6), due to its content of higher amounts of strong basic sites compared to its non-palladium containing counterpart (17/2/2/2). However, palladium and its compounds are known to be highly toxic, especially compared to those of iron (Kaushik and Moores, 2017; Egorova et al., 2019) and therefore its presence in particulates that are produced during use has to be taken into account. Lower energy and therefore costs have to be considered versus the problems caused by toxics contained in the catalysts which are released into the atmosphere due to abrasion during use.