Introduction

Carbon monoxide (CO) is a very toxic gas for humans and animals, produced by the incomplete combustion of any carbon substance. CO is in fact a colorless, tasteless, hazardous and odorless gas and for this reason it is the most common type of gaseous poisoning and every year several people die for accidental exposure to this gas. For this reason, there is an urgent need of devices able to detect and signal a CO alarm [1–7]. Huge amounts of this gas are emitted in the world, mainly from the living process, volcanic activity or bushfires. The condition of the environment and the human health are directly related to the anthropogenic emission. There are numerous and different sources of CO formation, e.g. transport, energy production, agriculture, chemical and steel industry [8, 9]. One of the most important solutions for its elimination is the catalytic oxidation. The catalytic oxidation of CO has been of considerable interest recently due to its relevance in many industrial applications such as gas purification in CO2 lasers, CO gas sensors, air-purification devices for respiratory protection, and pollution control devices for reducing industrial and environmental emission [10–12].

In the recent few years, a lot of catalysts were used in this reaction. Pillared clay (PILC) possesses several interesting properties, such as high surface area, large interlayer spacing, high thermal stability, special surface acidity. Due to their properties, PILC are very suitable to be used as adsorbents, catalysts and catalyst supports [13–21].

In order to get a faster and more effective method to prepare PILCs, the ultrasound irradiation was used. In the past few years, several works have reported the benefits of ultrasound for the reduction of the pillarization time. Thus for example, Pérez et al. [22] concluded that ultrasounds considerably reduce the modification time for solids, in fact in just 10 min they get the same effect to those obtained with prolonged synthesis method (24 h). Olaya et al. [23] reported the synthesis of PILC with and without ultrasound. They found that using ultrasound has the advantage of decreasing the consumption of water between 90 and 95 % as well as the processing time. Similar benefits were obtained by Tamoul [24] for the synthesis of zirconium-pillared bentonite.

The use of both noble metals [25–27] and transition metal oxides [28–30] supported on clays for the CO oxidation reaction has been well documented. In general, noble metals are highly active for the oxidation of CO, whereas they are expensive due to rareness. In contrast, transition metal oxides are less costly, and their catalytic activities can be optimized by tailoring their crystal structures, morphologies and surface properties during their synthesis [31–33].

Manganese oxides have been considered as one of the most promising transition metal oxide catalysts for oxidation reactions owing to their strong oxygen storage/release ability and excellent redox properties. The polyvalent property of manganese shows its highest catalytic activity in the complete oxidation of air pollutants [31, 34, 35].

The aim of our present study was to prepare environmentally friendly and cheaper catalysts based on manganese oxides supported on Al-pillared bentonite, which is an important clay mineral source in Tunisia. The characterization of catalysts was performed by X-ray powder diffraction (XRD), N2 adsorption at 77 K and H2 temperature programmed reduction (H2-TPR) techniques. The catalytic activities of the resulting solids were investigated in the oxidation of CO.

Experimental

Materials

Natural bentonite from the Haidoudi deposit in the region of Gabes in southern Tunisia was used as a starting material in this work. The chemical composition of the natural bentonite is: SiO2 63.30, Al2O3 22.35, Fe2O3 7.38, Na2O 1.34, MgO 1.50, CaO 2.26, and K2O 1.3, expressed in the oxide from 100 g−1 of calcined sample. The specific surface area SBET = 72 m2/g, pore volume Vp = 0.131 cm3/g, d001 = 15.37 Å and the cation exchange capacity (CEC) is 65.1 meq/100 g.

In order to prepare the pillared materials, crude bentonite was purified and transformed into its sodium form by treatment with 1 mol/L solution of sodium chloride (NaCl). The Na-saturated bentonite was washed with deionized water until excess chloride ions were removed (silver nitrate test), and dried at 80 °C. Fig.  1 shows the diffractograms of Na-clay (normal untreated, heated at 550 °C and glycolated. The basal distance d001 of the normal Na-clay was 12.75 Å. After treatment with ethylene glycol, the d002 of smectite appeared at 8.89 Å, which is attributable to an interstratified illite smectite [36]. In addition, according to the chemical analysis, the average elemental structural formula of the Na-bentonite (denoted H) purified clay is [Si7.312Al0.688]IV [Al3.014Fe0.656Mg0.357]VI O20 (OH)4Na0.714 Ca0.037 K0.174. Their specific surface area and total pore volume are about 118 m2/g and 0.150 cm3/g and its cation exchange capacity (CEC) is equal to 78 meq/100 g.

Fig. 1
figure 1

X-ray diffraction patterns of the oriented Na-clay (fraction <2 µm) N normal untreated, H heated, G glycolated)

Full size image

Synthesis of pillared clay

The pillaring solution was prepared by slow adding of a 0.2 M NaOH solution into a 0.1 M Al (NO3)3·9H2O solution until the OH/Al (mole ratio) was equal to 2.4. The obtained mixture was aged at room temperature for 24 h under constant stirring. The pillaring solution was then added dropwise under stirring at room temperature to a bentonite suspension of 2 wt% prepared in deionized water and stirred for 2 h, maintaining a ratio of 10 mmol of total metal per gram of clay. The final clay suspension was later aged under two different conditions: (1) at room temperature for 24 h under constant stirring and (2) for 20 min in an ultrasonic bath (Bandelin Sonorex, operating frequency 35 kHz). The resulting product was separated by centrifugation and washed with distilled water by dialysis membranes (5 times) and dried at 80 °C, The sample was ground, sieved (106 meshes) and calcined at 500 °C for 2 h at a heating rate of 10 °C min−1. The pillared bentonites prepared using the conventional treatment were labeled (Al–H) and the pillared bentonites prepared using the ultrasonic treatment were labeled (Al–H)*.

Catalyst preparation

The catalysts were prepared by the incipient wetness impregnation of the solids resulting from calcination at 500 °C for 2 h of the Al-PILC. The required volume of an aqueous solution of Mn(NO3)2·4H2O was slowly added to the PILC to give solids with a content of about 2 and 10 wt% amount of manganese. The resulting material was subsequently dried at 80 °C and heated at 500 °C for 2 h. These catalysts were labeled as 2 % Mn (Al–H) or 2 % Mn (Al–H)* and 10 % Mn (Al–H) or 10 % Mn (Al–H)*.

Characterization

X-ray diffraction (XRD) patterns were recorded over powder solids using a PANalytical XPert-V diffractometer, with a scanning speed of 5°/min, employing Cu Kα filtered radiation (λ = 1.5418 Å).

The chemical analysis of the starting material and modified samples was performed on an Analytik Jena model Vario 6 flame atomic absorption spectrophotometer (FAAS).

Surface area (SBET) values were obtained from the nitrogen adsorption/desorption isotherms at 77 K, using a fully automated model ASAP 2020 gas adsorption system (Micromeritics Instruments). Prior to the analysis, the samples were degassed for 2 h at 200 °C under a nitrogen atmosphere. The specific surface area (SBET) was calculated using the BET method from the adsorption data over a relative pressure range of 0.05–0.30. The specific external surface area (Sext) and the micropore volume (Vµ) were estimated by the t-plot method. The mesopore volumes (Vmes) are estimated by the Barrett–Joyner–Halenda (BJH) method [37].

The reducibility of the catalysts was investigated by temperature-programmed reduction (TPR), using a quadrupole mass spectrometer (Pfeiffer Vacuum Balzers, model QMS 200) equipped with a single Faraday detector, and controlled via a computer using the software Quadstar 422. Typically, 80 mg of sample were loaded in a U-shaped quartz microreactor and treated during 1 h with heat 500 °C, to remove water adsorbed on the sample surface. After cooling down to room temperature, the reduction gas (5 % H2/Ar) was introduced at a flow rate of 60 STP cm3/min. The system was stabilized for 2 h, and then the temperature programmed to heat with a ramp of 10 °C/min up to 950 °C.

Catalytic activity

The catalytic activity of the natural material and the modified clays was evaluated in the reaction of CO oxidation. This reaction was carried out in a conventional continuous flow U-shaped quartz microreactor. The catalyst is placed over quartz wool and carborundum (Prolabo). A thermocouple in contact with the sample assures the right measure of temperature. The feed mixtures were prepared using mass flow controllers. The reaction products were analyzed by mass spectrometry, using a Pfeiffer Vacuum Balzers, model QMS 200 mass spectrometer with capabilities for quantitative analysis.

The light-off curves of CO oxidation (700, 10 °C/min) were obtained with a mixture (1 % CO, 20 % O2 and 79 % He) at a total flow rate of 60 ml/min.

The catalysts were pre-activated in situ during 1 h at 500 °C with a mixture of 21 % O2 in helium at a total flow rate of 60 ml/min and then stabilized at room temperature before the light-off curve started.

Results and discussion

XRD

Fig. 2 shows the XRD patterns of un-pillared, Al-pillared bentonite prepared by conventional and ultrasound methods and 2; 10 % manganese oxide impregnated catalysts. The XRD results shows a clear shift of the basal spacing (d001) obtained for the PILC (17 and 18 Å for the samples prepared with and without ultrasound, respectively) with respect to the value obtained for the starting clay (12.75 Å), thus indicating that the clay was successfully pillared [22, 38–40].

Fig. 2
figure 2

X-ray diffraction patterns of bentonite, Al-pillared clay with and without ultrasound calcined at 500 °C and impregnated with two manganese loading 2 and 10 wt%. a (a) H; (b) (H–Al); (c) 2 % Mn (H–Al); (d) 10 % Mn (H–Al). b (a) H; (b) (H–Al)*; (c) 2 % Mn (H–Al)*; (d) 10 % Mn (H–Al)* The symbols correspond to MnO2 (closed diamond) and Mn2O3 (closed heart)

Full size image

The use of ultrasounds induces a slight increase in the basal spacing of the PILCs. As observed in Fig. 2, this sample shows a narrower and more defined signal characterized by a value of width at medium height clearly lower (FWHM = 0.5°, 2θ) than the one observed in the sample modified in the absence of ultrasounds (FWHM = 0.93°, 2θ).

The introduction of manganese does not modify the basal spacing of both pillared bentonite synthesised by the conventional and ultrasonic treatment. However, the line 001 becomes less intense and broader. These observations could indicate that impregnation produced a significant decrease in the ordering of the solids, but does not affect the layered structure of the supports.

Compared with Al-PILCs samples, no additional peaks characteristic of the manganese oxide phases were observed for PILCs modified with a 2 wt% of manganese. This may not require the absence of manganese oxides due to the fact that manganese is not present in sufficient quantities to be detected by XRD [33, 41]. However, in the XRD patterns of the catalysts impregnated with (10 wt% of Mn), new diffraction peaks appear. According to ASTM cards number 812261 and 780390, reflections at 28.68°, 37.2°, 42.82°, 56.65° and 59.37° correspond to manganese oxides in the form of MnO2 while those at 32.9°, 55.1° and 65.1° reveal the presence of Mn2O3 [29, 30, 42]. Similar results were obtained by Kazin et al. [43] and Craciun [44] after calcinations in air of a Mn/SiO2 catalyst at 500 °C.

Chemical analysis

The chemical compositions of all samples, as obtained from atomic absorption, are presented in Table 1. After pillaring via both conventional and ultrasonic treatment, an increase in the amounts of aluminum with a corresponding decrease in the amount of exchangeable cations highlights the success of Al incorporation in the solids. For the PILCs in the presence of ultrasound get similar results, with a little increase in the amount of Al compared with those obtained by conventional method. Meanwhile, this demonstrates the positive effect of ultrasound towards a greater introduction of the pillaring agent in the clay structure.

Table 1 Chemical analysis (wt%) of parent and modified interstratified illite–smectite, referred to ignited solids (0 % water)
Full size table

The impregnation of the manganese was also verified by chemical analysis. The increase in the MnO content in the Al-pillared bentonites after manganese impregnation shows that manganese was successfully inserted into the PILC structure. The incorporated amounts correspond to the employed amount, showing the effectiveness of the impregnation method [45].

Textural properties

The nitrogen adsorption/desorption isotherms of Al-PILC with and without ultrasound and the manganese oxide impregnated catalyst are shown in Fig. 3, and their textural parameters calculated from these isotherms are given in Table 2.

Fig. 3
figure 3

Nitrogen adsorption/desorption isotherms of different samples. a Support prepared with conventional method and b support prepared with ultrasound method

Full size image
Table 2 Textural properties of untreated, pillared and impregnated pillared samples
Full size table

From the nitrogen the adsorption–desorption isotherms, it can be inferred that all our samples are micro and mesoporous materials. Indeed, nitrogen isotherms belong to a mixed type of International Union of Pure and Applied Chemistry (IUPAC) classification  [46]. The initial part of the isotherms is of type I with uptake of nitrogen at very low relative pressure, which corresponds to adsorption in micropores. At intermediate and high relative pressures, the isotherms are of type IV with a narrow hysteresis loops of type H3 associated with capillary condensation in mesopores of plate-like aggregates with open slit-shaped capillaries [47–50].

For the PILCs we observed an increase in the BET surface areas and in the porosity, mainly in microporosity, comparing with the purified clay. This increase confirms the successful pillaring of the bentonite.

It is important to emphasize that the use of ultrasound during the intercalation induces an improvement of textural properties compared to the conventional method. However, in the impregnated material, the specific surface area as well as the micropores volume decreased comparing to the Al-PILC prepared with both methods. This effect can be related to the partial blockage of the interlayer space of the clay by the impregnated metal oxides, which prevents the access of nitrogen [26, 48].

Temperature-programmed reduction (TPR)

The catalysts were characterized by temperature programmed reduction in order to obtain information on the reducibility and oxidation state of the metals in the impregnated solids.

The H2-TPR profiles of the starting clay, PILC with and without ultrasound (supports) and the impregnated clays with 2 and 10 wt% Mn are presented in Fig. 4. The unpillared interstratified illite–smectite reduction profile shows two reduction peaks at around 610 and 870 °C, with a shoulder at 400 °C. The peak at 610 °C can be assigned to the reduction of the magnetite to wustite (Fe3O4 → FeO), and the high temperature peak (870 °C) may result from the reduction of wustite to metallic iron (FeO → Fe). The shoulder at 400 °C has been previously associated to the reduction of the superficial iron oxides in the structure (hematite to magnetite; Fe2O3 → Fe3O4) [40, 51]. The reduction profile of the pillared sample (H–Al) shows only one peak at 610 °C. This peak probably accounts for the total reduction of Fe species in this sample, much more accessible to reduction gas after pillaring. The assignment of these peaks should be done with caution as they are well above the temperature used in the previous calcination of the samples. According to Fig. 4b, the use of ultrasound does not significantly affect the reduction behavior of this sample.

Fig. 4
figure 4

Profiles H2 reduction (H2-TPR) of various samples a (support in absence of ultrasound): (a) H; (b) (H–Al); (c) 2 % Mn (H–Al); (d) 10 % Mn (H–Al). b (support in presence of ultrasound): (a) H; (b) (H–Al)*; (c) 2 % Mn (H–Al)*; (d) 10 % Mn (H–Al)*

Full size image

The incorporation of a 2 wt% of Mn leads to a broadening of the reduction peak at around 600 °C, which now accounts for the global reduction process of Fe and Mn species. In the case of the 10 % loaded sample, new maxima associated to MnOx reduction are observed. According to the normal reduction scheme proposed for MnOx [30, 42, 52], the peak at 380 °C involves the reduction of MnO2 and Mn2O3 to Mn3O4, and the maxima at 455 °C could be assigned to the formation of MnO. Given that the peak which appears at 455 °C is significantly more intense than the one at 380 °C, we can conclude that the oxide Mn2O3 is a dominant form. As indicated for the manganese-free clays, it seems that the reducibility of the samples is not influenced by the use of ultrasounds.

Quantitative analyses on the H2-TPR profiles can give the H2 consumption of the samples, the total amount of hydrogen consumed of the reduction peaks estimated by integrating these profiles and the temperature of first peaks were compiled in Table 3.

Table 3 Quantitative results of H2-TPR
Full size table

In the case of the supports, a slight increase from 5.37 to 5.98 µ mol/g of hydrogen consumption in comparison to starting clay was verified, suggesting that when the clay is pillared, the newly formed porous structure promotes the reduction of iron oxides that are not initially accessible in the starting bentonite. It is important to emphasize that the increase of Mn content greatly enhances the hydrogen consumption. The catalyst obtained with 10 % Mn has the lowest reduction temperature and the highest hydrogen consumption, possibly providing a better catalytic system in the oxidation reaction.

Catalytic activity

The catalytic oxidation of CO by O2 over unPILC and different synthesised catalysts was investigated at temperatures within the range of 25–700 °C. Figs.  5a and 5b show the light-off curves for the purified bentonite and manganese oxides with 2 and 10 wt% supported on Al-pillared bentonite synthesized by using both, the conventional method and ultrasonic treatment. As it is shown in Fig. 5, the fresh and Al-PILCs presented the low catalytic activity for CO oxidation. Similar results were found by Carriazo et al. [25] and Pérez et al. [22] for bentonite clays from Colombia or Turkey [53]. In order to implement the structural and textural properties of Al-PILCs with activity for this reaction, they have been used as support for preparing MnOx-supported catalysts. The results corresponding to the catalytic performance of two different manganese weight loadings (2 and 10 %) catalysts for the oxidation of CO are also included in Fig. 5 and Table 4. The influence of the addition of manganese is clearly deduced from these results. Indeed, the 2 % Mn sample shows a light-off temperature (T50) of 600 °C, with a maximum conversion value of 60 % obtained at the higher temperature investigated (700 °C). Moreover, full CO conversion is obtained with the 10 % Mn catalyst, which also shows a significant decrease in the light-off temperature (375 °C) with respect to the lower loading sample 2 % Mn (T50 = 600 °C).

Fig. 5
figure 5

CO conversion of parent clay, support prepared with (b) and without (a) ultrasound and impregnated catalysts, 1 % CO + 20 % O2 in helium; catalyst weight = 0.25 g, total flow rate = 60 ml/min

Full size image
Table 4 Maximum conversion and temperature at 50 % conversion
Full size table

The activity of manganese oxides supported on bentonite based catalysts have been previously investigated for oxidation reactions such as the combustion of volatile organic compounds [54]. However, to the best of our knowledge, these are the first published results showing the activity of manganese oxides impregnated on bentonite PILCs for CO oxidation. In term of light-off temperature, our results are rather close to those reported by Zaki et al. [55] who obtained a CO conversion of 50 % at 340 °C using 10 % wt Mn2O3 supported on TiO2 with Al2O3 and by Venkataswamy et al. [56] who obtained a light-off temperature of 374 °C for the manganese oxide supported on alumina (MnOx/Al2O3).

Conclusion

As a summary of the results presented in this work, manganese oxide supported over inexpensive support PILCs were prepared and used as catalysts for oxidation of CO. XRD results show an increase of the basal spacing after pillaring and the conservation of the layered structure after Mn-impregnation. The results of N2 adsorption reveal that the textural proprieties of PILC are enhanced. The TPR patterns showed a two-step reduction profile of manganese oxide and a three reduction peaks of iron. Activity tests in CO oxidation show that supported manganese oxide catalysts are found to be very suitable for this reaction. The higher loading catalysts (with 10 wt% of Mn) showed the highest catalytic activity for CO oxidation compared to the lower loading samples (with 2 wt% of Mn), Al-PILC and starting clay, over which the CO conversion is up to 50 % at a temperature of 375 °C. It is important to mention that the use of ultrasound lead to an enhancement of the cristallinity and considerably reduces the pillaring duration, since in only 20 min treatment we observe similar catalytic performances with the catalysts prepared during 24 h by conventional method.