Introduction

The removal of sulfur from petroleum cuts is a priority in oil processing due to the exigencies imposed by environmental protection regulations. Among the sulfur-containing compounds, thiols are among the less desirable since they cause foul odor, deterioration of additives in finished products, poisoning of the catalysts used in oil processing and corrosion. Therefore, it is necessary to remove them, either by extraction or by their subsequent transforming to inoffensive alkyl disulfides. In the petroleum industry, such processes are usually called “sweetening”. The most widely used sweetening process is Merox developed by UOP [1–3], which is based on the aerobic oxidation of alkane thiols to dialkyl disulfides at 25–40 °C, in alkaline environment (NaOH aqueous solution 10%wt.) in the presence of a soluble Co(II) complex catalyst with oxygen carrier properties (e.g. disodium Co(II)phtalocyanine-disulfonate), according to the reactions:

$$ 2{\rm RSH} + 2{\rm NaOH} \Leftrightarrow 2{\rm RSNa} + 2{\rm H}_{2} {\rm O} $$
(1)
$$ 2{\text{RSNa}} + {\text{H}}_{ 2} {\text{O}} + [{\text{O}}] \to {\text R-\text S-\text S-\text R}+ 2{\text{NaOH}} $$
(2)
$$ 2{\text{RSH}} + [{\text{O}}] \to {\text R-\text S- \text S-\text R }+ {\text{H}}_{ 2} {\text{O}} $$
(3)

The addition of a base is necessary (pH = 9–9.5) because in the initial step a mercaptide ion has to be formed [4]. The sweetening process is highly efficient in demercaptanization since in the case of alkyl radicals R with more than 3 carbon atoms the disulfides which are not soluble in gasoline precipitate thus facilitating their complete removal from gasoline. However, the presence of the base solution raises serious problems due to its caustic character. Therefore, a special attention has been granted to the replacement of homogeneous sweetening catalysts with solid catalysts, such as those obtained by impregnation of an activated carbon with the Merox catalyst and the base compound. Several studies reported also the catalytic activity of metal complexes immobilized into layered double hydroxides (LDH) supports, especially Co(II)phtalocyanine-tetrasulfonate [Co(PcTS]4− for the auto-oxidation of thiols with O2 [5–7]. The immobilization of the transition metal complex into LDH has three main effects on the catalytic behavior. First, the activity of Co complex is improved by heterogenization; second, the catalyst is more stable in time than the homogeneous homologue, and third, the LDH or the mixed oxides obtained from LDHs precursors not only act as a carrier but also as a solid base [8, 9], favoring the dissociation of thiols while playing a role similar to the one of the alkaline solutions used in homogeneous conditions.

The first report concerning the oxidation of thiols by molecular oxygen in the presence of a molybdenum compound entrapped in a LDH host (e.g. Zn–Al) was published by Corma et al. [10]. Layered double hydroxides defined by the general formula: [M z+1−x M 3+x (OH)2]b+[A n−b/n ] m H2O, (M = metal, A = interlayer anion, b = x or 2x−1, for z = 2 or 1), are anionic clays with laminar structure consisting of positively charged brucite-type metal hydroxide layers with balancing anions and water molecules in the interlayer space. The most representative naturally occurring solid of this type is the mineral hydrotalcite (HT-CO3 2−) having the formula Mg3Al(OH)8(CO3)0.5·2H2O [11–16]. This type of solid is known to be a useful solid base catalyst. Its composition may be modified in order to obtain catalysts with both redox and base properties, yielding the so-called hydrotalcite-like compounds (HTlc). One of the methods used for the obtaining of HTlc is the replacement of Mg2+ or Al3+ cations from the brucite-type layer with other bivalent or trivalent transition metal cations having similar cationic radius and coordination preference. Another method implies the exchange of the carbonate anions from the structure with oxometallate or polyoxometallate anions. This method is applied for the modification with metals that are not compatible with the octahedral sites of the brucite-type sheet, such as Mo, or W [16–20]. HTlcs containing interlayer oxometallate or polyoxometallate anions can be used as precursors for obtaining new catalysts with outstanding properties due to the high dispersion and the possibility of migration and rearrangement of the active species [14, 16, 18]. Several Mo-containing HTlcs were proved to be active catalysts for different oxidation reactions such as selective olefin oxidation [21–23] or oxidative dehydrogenation of propane [24]. Most of the studies were directed towards HTlcs where Mo was present as heptamolybdate (Mo7O24 6−) species acting as pillars that expanded the distance between the brucite-type sheets [16, 17, 20–25]. Generally, these solids were prepared at low pH values in order to stabilize the Mo7O24 6− anions.

Recently, several methods have been developed for producing hydrotalcite-like materials containing molybdate (MoO4 2−) anions (Mo-HT), which are good catalysts for selective oxidations under mild conditions [26–28]. The dispersion, the accessibility and the coordination state of Mo in the active sites as well as its redox behavior are among the factors that influence strongly the catalytic activity. These characteristics are mainly determined by the preparation conditions (composition of the reaction mixture, pH, temperature, ageing treatment) used to obtain the catalyst.

Based on the above-mentioned facts, the aim of this contribution is to investigate the catalytic performances of Mo-containing HTlc (Mo-HT) and their derived mixed oxides (Mo-CHT) in the oxidation of tert-butanethiol (t-BuSH) with air. This test reaction has been selected since t-BuSH is one of the impurities difficult to be removed by catalytic sweetening from naphta gasoline. The influence of the preparation method as well as of the molybdenum source on the structural and physico-chemical properties of the solids is going to be analyzed in correlation with their catalytic activity.

Experimental

Preparation of the catalysts

Three Mo-containing HTlcs samples (Mo-HT) were obtained at constant pH 10 following two different procedures: ionic exchange (IE) without the presence of an organic swelling agent, or coprecipitation at high supersaturation using Mg(NO3)2, Al(NO3)3 and two different molybdenum sources, Na2MoO4, and (NH4)6Mo7O24·NaOH and HNO3 aqueous solutions (1.5 M) were used for pH adjustment. All reagents were purchased at high chemical purities (>98%) from Merck. Decarbonated distilled water was used for the preparation of the aqueous solutions. Both the ionic exchange and the direct synthesis were performed under inert atmosphere in order to avoid the contamination with CO2.

For the preparation by ionic exchange according to the method previously developed by Kwon [29], an amount of Na2MoO4 solution (7.5×10−3 M) containing a number of MoO4 2− anions equal to the half of the anionic exchange capacity (AEC) of a parent hydrotalcite Mg3Al(OH)8(NO3)·2H2O (HT-NO3) synthesized by coprecipitation at constant pH = 10 and low supersaturation [11, 26] had been used. The pH was adjusted to 10 using a 1.5 M aqueous solution of NaOH.

The preparation by co-precipitation at pH = 10 and high supersaturation [11, 28], was performed by adding a Mg and Al-nitrates solution (1.5 M (Mg+Al), ratio Mg/Al = 3 initial pH = 4.5) with a constant flow rate of 1mL/min to a base solution (initial pH = 10.5) containing the appropriate amount of a molybdenum salt (e.g. Na2MoO4, or (NH4)6Mo7O24) aiming at a composition of Mg:Al:Mo = 3:1:0.25. The concentration of Mo in the aqueous base solution was 0.1 M. Aqueous NaOH solution (1.5 M) was used for the pH adjustment in the preparation with Na2MoO4 [28], while a NH4OH solution (25%wt. NH3) was used for the pH adjustment in the preparation with (NH4)6Mo7O24.

All the samples were submitted to an aging treatment at 65–70 °C during 18 h, followed by filtration, washing with decarbonated distilled water (up to pH = 7), and drying 18 h at 90 °C. Depending on the preparation procedure (PP) and the molybdenum source (S) used for obtaining the precursors, their names had been abbreviated as Mo-HTPP-S: (i) Mo-HTIE-Na2MoO4 (IE=ionic exchange at pH 10); (ii) Mo-HTHS-Na2MoO4 (HS = direct synthesis at constant pH = 10 and high supersaturation and (iii) Mo-HTHS-(NH4)6Mo7O24.

The mixed oxide catalysts (Mo-CHT) were obtained following the calcinations of the above-mentioned Mo-HT precursors under air-flow at 450 °C during 24 h. The names of these samples are defined by replacing the suffix HT with CHT in the name of the precursor, e.g. Mo-CHTIE-Na2MoO4, Mo-CHTHS-Na2MoO4 and Mo-CHTHS-(NH4)6Mo7O24, respectively.

Characterization

The physico-chemical properties of the prepared solids were determined by chemical and TG-DTA analyses, XRD, BET, SEM–EDS, FTIR, DR-UV–Vis and Raman spectroscopy.

The metal content in the samples was determined by atomic absorption spectrometry (AAS) using a VARIAN AAS apparatus, while the carbon and nitrogen content of the samples was determined by elemental analysis performed with Carlo Erba automatic analyzer. The obtained results are presented in Table 1.

Table 1 The chemical compositions of the catalysts
Full size table

TG-DTA analyses were performed by heating the catalyst samples (50 mg) from 50 °C up to 800 °C (1 °C/min) under N2 flow (50 mL/min) using SETARAM TG-DTA apparatus for recording TG and DTG curves and α-Al2O3 as standard.

The X-ray diffraction patterns of the samples were recorded using a DRON 3 diffractometer equipped with a graphite monochromator (λCuKα = 1.5418 Ǻ). The samples were scanned from 7 to 70° (2θ) in steps of 0.05° with an acquisition time of 2 s at each point. The profile fitting calculations were performed using Jandel Scientific computer software and Voigt functions.

The catalysts were characterized by diffuse reflectance UV–VIS spectroscopy in the range 220–800 nm using SPECORD 80 UV–VIS spectrometer with an integration sphere coated with MgO taken as reference.

FTIR spectra in the range 400–4000 cm−1 were recorded on BioRad FTS 135 spectrometer using the KBr pellets technique. The concentration of the solid sample in the KBr matrix was 0.1 wt%.

Micro Raman spectra of the catalysts in the range 1424–196 cm−1 were recorded with 1 scan/8 s using SPEX 1877 triplemate spectrophotometer and the laser line 514.5360 nm as exciting radiation.

The specific surface area of the solids was determined by BET method from nitrogen adsorption–desorption isotherms at −196 °C using a Carlo-Erba instrument. Prior to the determination, the samples were degassed at 90 °C during 24 h under a pressure of 0.1 Pa.

The homogeneity of the samples was characterized by SEM–EDX using a HITACHI S2400 scanning electron microscope equipped with a RONTEC energy dispersive X-Ray high vacuum detector (W filament, E = 25 kV). The samples were analyzed from 15 mm working distance, with a resolution of 12 nm.

The surface base sites of the catalysts were determined on the basis of the irreversible adsorption of organic acids, such as acrylic acid, pKa = 4.2 and phenol pKa = 9.9, using a method that is suitable for both dried and calcined samples which was described in detail in a previous publication [30]. The amount of adsorbed phenol is related to the number of strong surface base sites while the amount of adsorbed acrylic acid is related to the total number of base sites. The number of weak base sites is given by the difference between the amount of adsorbed acrylic acid and the adsorbed phenol. Before base site measurements, all catalysts were thoroughly degassed under vacuum at room temperature.

Catalytic tests

In order to select the best catalysts, preliminary catalytic tests were performed by contacting synthetic samples of gasoline (100 mL) containing 128.5 ppm t-BuSH with the catalysts (1.3 g) under air flow (5 L/h) at ambient temperature and pressure, under continuous stirring during 4 h. For the series of catalysts presenting higher activity, the effects of the catalyst concentration, the air flow rate and the reaction time were also investigated. Another set of experiments aimed to investigate the resistance of the catalysts under operating conditions and the possibility of using them in six repeated reaction cycles (4 h each). The catalyst used in one reaction cycle was separated from the reaction mixture, washed with diethyl ether and dried under air flow during 1 h at 90 °C, before being introduced in the next cycle. The concentration of t-BuSH in the reaction mixture was determined by gas chromatography using a GC K072320 Thermo-Quest chromatograph equipped with a FID detector and a capillary column of 30 m length with DB5 stationary phase.

Results and discussions

The data presented in Table 1 show that both the ratio Mg/Al and the Mo uptake are different depending on the preparation conditions and the molybdenum source.

All the samples presented traces of carbon suggesting that carbonate is still present no matter what precautions are taken. For both Mo-HT samples prepared using Na2MoO4 as a molybdenum source either by ionic exchange or by co-precipitation at high supersaturation, the Mo-uptake was lower than the expected value considering the atomic ratio in the preparation mixture Mg:Al:Mo in the reaction mixture 3:1:0.25. For the sample prepared by ionic exchange this could be expected since the true exchange capacity of anionic clays (e. g. 1.0–1.5 meq g−1) is usually much lower than the theoretical one (3.3 meq g−1 for HT-NO3 −) [14, 16]. The atomic ratio Mg:Al was also slightly lower than 3, indicating that for these samples, there is probably a leaching of Mg during the washing stage. The differences in the Mg/Al ratios were also confirmed by the results of the XRD and SEM–EDX analyses (see Table 2). The chemical composition of the sample prepared using (NH4)6Mo7O24 as a molybdenum source was very different from that expected considering the composition of the reaction mixture, indicating that a high amount of molybdenum is incorporated in the structure in this case, but there is also a pronounced loss of Mg during the washing step since the ratio Mg/Al is much lower than 3 (e.g. 1.6). Compared to Mo-HTHS-Na2MO4, Mo-HTHS-(NH4)6Mo7O24 also presents a different behavior during thermal decomposition as it may be seen from the results of the thermal analysis presented in Fig. 1. Thus, the total loss of weight during the thermal decomposition of Mo-HTHS-Na2MO4 was 39 wt% (typical for a hydrotalcite-like compound) whereas for Mo-HTHS-(NH4)6Mo7O24, this loss of weight was only 23 wt%, indicating a higher thermal stability. Mo-HTHS-Na2MO4 which has a low concentration of Mo presents two weight losses accompanied by strong endothermic effects at 235 and 419 °C respectively, while the sample Mo-HTHS-(NH4)6Mo7O24, which has a high concentration of Mo, shows four stages of weight loss accompanied by endothermic effects at 177, 445, 526, and 571 °C. The effects noticed at temperatures higher than 500 °C are ascribed to the thermal decomposition of the paramolybdate anions [31].

Table 2 Results of XRD, SEM–EDX and BET analyses for Mo-HT samples
Full size table
Fig. 1
figure 1

TG and HF profiles for Mo-HTHS-Na2MoO4, and Mo-HTHS-(NH4)6Mo7O24

Full size image

The calcined samples Mo-CHT have higher Mg, Al, Mo concentrations since the nitrate anions and most of the hydroxyl groups are lost during calcination. However, since calcination takes place under air flow, traces of carbonate are also found.

The XRD patterns of Mo-HT samples prepared by ionic exchange or co-precipitation at high supersaturation using Na2MoO4 as a molybdenum source (Fig. 2Ia, b) exhibit a typical hydrotalcite structure with sharp and symmetric reflections for (003), (006), (110) and (113) planes and broad asymmetric peaks for (012), (015) and (018) planes (JCPDS 70-2151) [11]. The reflections were indexed in a hexagonal lattice with R3m rhombohedral symmetry. The cell parameters are given by a = 2 × d110 and c =\( \frac{3}{2} \) x (d003 + 2 d006). The interlayer free spacing (IFS) was calculated by subtracting the brucite sheet thickness (e.g. 4.8 Ǻ as reported by Miyata [32]) from \( \frac{c}{3} \) values. These samples have a homogeneous composition as indicated by SEM–EDX analysis. The basal space peaks are very sharp and intense revealing high stacking order along c axis due to a more compact orientation of the anions which are mainly NO3 −. The (0kl) peaks are less asymmetric compared to those of the reference pattern of the natural mineral HT-CO3 2−, showing less turbostratic disorder effects. In their case the decrease of the Mg/Al ratio compared to the standard HT-CO3 2− is accompanied by the decrease of c parameter and consequently of the IFS. This fact indicates stronger electrostatic interactions between layer and interlayer anions due to the higher amount of Al3+ in the brucite sheet.

Fig. 2
figure 2

XRD patterns of the catalysts: (I) Mo-HT precursors: a Mo-HTHS-Na2MoO4, b Mo-HTIE-Na2MoO4, c Mo-HTHS-(NH4)6Mo7O24, (II) Mo-CHT mixed oxides: d Mo-CHTHS-Na2MoO4, e Mo-CHTIE-Na2MoO4, f Mo-CHTHS-(NH4)6Mo7O24, * diffraction lines of Mo compounds impurities

Full size image

The diffraction pattern of Mo-HTHS-(NH4)6Mo7O24 (Fig. 2Ic) shows a considerably disordered structure with broader and less defined peaks. The maximum of the reflection corresponding to (003) plane shifted to lower values of 2θ suggests the location of the paramolybdate species in the interlayer. The significant broadening of the second diffraction line (006), which is also shifted to lower values of 2θ, highlights the disorder in the stacking of the planes (00l) due to the presence of the Mo species similarly to the phenomenon reported in the case of the intercalation of metavanadate species [33, 34]. The wide basal peaks and the broad diffraction band corresponding to the (110) plane suggest the presence of small crystallites. Normally, the smaller size of the crystallites would lead to a larger surface area. However as it may be seen from Table 2, this sample has the lowest surface area. This fact was also confirmed by the SEM microphotographs presented in Fig. 3 where it may be seen that at the same magnification level (e.g. 2700), the particles of Mo-HTHS-(NH4)6Mo7O24 exhibit a quasi-flat surface with small excrescences while the particles of Mo-HTHS-Na2MoO4, which has a well-defined diffraction pattern, show a porous layered structure. Therefore it may be assumed that the poor crystallinity of Mo-HTHS-(NH4)6Mo7O24 may be due to the presence of a compact amorphous phase covered with small crystallites. The less homogeneous composition of Mo-HTHS-(NH4)6Mo7O24 was also revealed by the results of SEM–EDX analysis presented in Table 2.

Fig. 3
figure 3

SEM microphotographs of Mo-HTHS-Na2MoO4, Mo-HTHS-(NH4)6Mo7O24 and Mo-CHTHS-Na2MoO4

Full size image

The XRD patterns of Mo-CHTIE-Na2MoO4 and Mo-CHTHS-Na2MoO4 (Fig. 2IId, e) present the typical features of Mg(Al)O mixed oxide with periclase structure (a = 4.184 and 4.193Å, respectively) and no traces of polymolybdate species have been noticed. Meanwhile, the pattern of Mo-CHTHS-(NH4)6Mo7O24 (Fig. 2IIf) shows a high amount of amorphous phase with several diffraction lines corresponding to molybdenum compounds. The unit cell parameter and the mean crystallite size were evaluated assuming a Voigt profile function. The mean crystallite size was determined using the Debye–Scherrer formula:

$$ {\text{D }}\;=\;\left( {{\text{k}}*\lambda /\beta { \cos }\theta } \right) $$
(4)

where θ is the Bragg angle, β is the fullwidth at half maximum of the peak corrected by the instrumental broadness, λ is the X-ray wavelength (λCuKα = 1.5418 Å) and k is the structural factor (taken as 0.9 for the cubic structure of MgO). The mean crystallite size was calculated as an average value given by the FWHM of (200) and (220) peaks. The results are presented in Table 3.

Table 3 Results of XRD, SEM–EDX and BET analyses for Mo-CHT samples
Full size table

The surface areas of the calcined samples presented in Table 3 are larger than those corresponding to the dried precursors, while respecting the same variation order (e.g. the largest surface area is exhibited by the samples prepared by ionic exchange with Na2MoO4, followed by the sample prepared by co-precipitation at high supersaturation with Na2MoO4 and the lowest surface area is characteristic of the sample prepared at high supersaturation with (NH4)6Mo7O24). The effect of calcination on the surface modification is illustrated by the SEM microphotographs in Fig. 3, where it may be seen that at the same magnification level (e.g. 3500) the dimensions of the particles in Mo-CHTHS-Na2MoO4 are smaller than those in Mo-HTHS-Na2MoO4.

The FTIR spectra of the molybdenum containing catalysts are presented in Fig. 4.

Fig. 4
figure 4

FTIR spectra of Mo-containing samples (a Mo-HTHS-Na2MoO4, b Mo-CHTHS-Na2MoO4, c Mo-HTIE-Na2MoO4, d Mo-CHTIE-Na2MoO4, e Mo-HTHS-(NH4)6Mo7O24, f Mo-CHTHS-(NH4)6Mo7O24, —856 cm−1)

Full size image

The FTIR spectra of the Mo-HT samples (Fig. 4a, c, e) show broad adsorption bands suggesting a high disorder of the molecules in the galleries. Meanwhile, the band corresponding to Mo–O–Mo stretching vibrations at 665 cm−1, the one related to MoO4 2− at 856 cm−1 and the band for Mo=O stretching at 914 cm−1 [25, 35] appear in the same region as those characteristic for a typical HT structure. Therefore, due to the overlapping, it is difficult to comment the nature of Mo species in the modified HTlcs. The most significant difference between the FTIR spectra of the samples Mo-HTHS-Na2MOO4 and Mo-HTHS-(NH4)6Mo7O24 is the higher intensity of the peak at 1400 cm−1 ascribed to the presence of NO3 − anions [36] in the spectrum of Mo-HTHS-Na2MOO4. This band overlaps the one characteristic for the vibrations of CO3 2−, which is present in traces as it was shown by the results of the chemical analysis presented in Table 1. The effect of the ionic exchange for the sample Mo-HTIE-Na2MoO4 is noticed due to the lowering of the band corresponding to NO3 −vibrations compared to the band in the spectrum of Mo-HTHS-Na2MoO4.

In the FTIR spectra of the calcined samples (Fig. 4b, d, f) the decrease of the intensity of the band corresponding to the hydroxyl groups at 3460 cm−1 compared to the spectra of Mo-HT samples was noticed. The FTIR spectrum of Mo-CHTHS-(NH4)6Mo7O24 indicates that this sample has a lower amount of hydroxilic groups and a higher amount of Mo incorporated compared to Mo-CHTHS-Na2MoO4, whose spectrum displays higher intensity bands corresponding to hydroxyl groups while the bands corresponding to Mo–O bonds are noticed as shoulders.

Fig. 5 displays the DR-UV–Vis spectra of the Mo containing catalysts in the region 220–500 nm where the absorption bands characteristics to Mo species appear.

Fig. 5
figure 5

DR-UV–Vis spectra of Mo-containing samples a dried—Mo-HTHS-Na2MoO4; calcined—Mo-CHTHS-Na2MoO4; b dried—Mo-HTIE-Na2MoO4, calcined—Mo-CHTIE-Na2MoO4; c dried—Mo-HTHS-(NH4)6Mo7O24, calcined—Mo-CHTHS-(NH4)6Mo7O24

Full size image

According to literature data, the isolated MoO4 2− species with Mo in tetrahedral coordination present absorption bands in the region 260–280 nm, while the bands in the region 300–320 nm are due to Mo7O24 6− and isolated MoO3 species with Mo6+ in octahedral coordination [37, 38]. As it may be seen from Fig. 5, all the samples contain both Mo in tetrahedral and octahedral coordination. However, in the samples prepared with Na2MoO4, (Fig. 5a, b) the amounts of tetrahedrally coordinated species are prevailing whereas in the samples prepared from (NH4)6Mo7O24 (Fig. 5c) the octahedrally coordinated species prevail.

The results obtained by DR-UV–Vis were confirmed by those of Raman spectroscopy analysis illustrated by the spectra presented in Fig. 6.

Fig. 6
figure 6

Raman spectra of Mo-CHT samples (a Mo-CHTHS-Na2MoO4, b Mo-CHTIE-Na2MoO4, c Mo-CHTHS-(NH4)6Mo7O24, T tetragonal, O octahedral)

Full size image

The Raman spectra revealed the presence of the most intense absorption band corresponding to Mo–O symmetrical stretching vibration in MoO4 2− at 896 cm−1 and the one corresponding to δ(Mo=O) in MoO4 2− at 320 cm−1 [21, 26] in all the calcined samples, even if in the case of Mo-CHTHS-(NH4)6Mo7O24 (Fig. 6c) the intensity of these bands was much lower. The bands characteristic for Mo7O24 6− at 940 cm−1 (highest intensity), 564 and 219 cm−1 (Mo–O-Mo ν and δ vibrations) and 340 cm−1 (δ(Mo=O)) [39] could be discerned clearly in the spectrum of Mo-CHTHS-(NH4)6Mo7O24, while in the spectra of Mo-CHTIE-Na2MoO4 and Mo-CHTHS-Na2MoO4 (Fig. 6a, b), only the band at 940 cm−1 could be noticed as a shoulder.

The results of the preliminary catalytic tests (Fig. 7) performed by contacting the catalysts with synthetic samples of gasoline contaminated with 128.5 ppm t-BuSH under air flow at ambient temperature and pressure showed that for all Mo-HT samples the conversions of t-BuSH were lower by 10–25% than those obtained in the presence of the corresponding derived mixed oxides Mo-CHT. At a first glance, this fact should be explainable by the increase of the surface area in the mixed oxide samples (see Table 3) compared to the hydrotalcite-type precursors (see Table 2).

Fig. 7
figure 7

Comparison of the catalytic activities of precursors and calcined samples, air flow 5 L/h, 1.3 g catalyst, 100 mL gasoline with Co t-BuSH = 128.5 ppm, 4 h reaction time. Precursors: 1 MoHTIE-Na2MoO4; 2 MoHTHS-Na2MoO4; 3 MoHTHS-(NH4)6Mo27O24; Calcined: 1 MoCHTIE-Na2MoO4; 2 MoHTHS-Na2MoO4; 3 MoHTHS-(NH4)6Mo27O24

Full size image

However, it was surprising to notice that the lowest conversion level was obtained in the presence of the samples containing the highest concentration of Mo, e.g. Mo-HTHS-(NH4)6Mo27O24 and Mo-CHTHS-(NH4)6Mo27O24, which yielded only 15 and 24% conversion of t-BuSH. These conversion values are less than half of the conversions reached with the catalysts prepared using Na2MoO4 as a molybdenum source, which have a concentration of Mo sites that is 6-8 times lower than in Mo-HTHS-(NH4)6Mo27O24 and Mo-CHTHS-(NH4)6Mo27O24 samples. Meanwhile, the specific surface areas of these catalysts (35 and 82 m2/g, respectively) are only approximately ¼ lower than those of the samples Mo-HTHS-Na2MoO4 and Mo-CHTHS-Na2MoO4. One fact revealed clearly by the characterization of the samples that might be related to the poor catalytic activity of Mo-HTHS-(NH4)6Mo27O24 and Mo-CHTHS-(NH4)6Mo27O24 could be the low amount of Mo species with tetrahedral coordination included in these samples. Another fact that could be accounted as responsible for their low activity would be a lower basicity of the surface which would lead to an inefficient activation of the alkanethyol reactant.

In order to confirm this supposition, the surface base sites of the catalysts was determined using a method based on the irreversible adsorption of organic acids, (acrylic acid, pKa = 4.2 and phenol pKa = 9.9) that is suitable for both dried and calcined samples [30]. The results presented in Table 4, show that undoubtedly the samples prepared with (NH4)6Mo7O24 have the lowest number of base sites, while the samples prepared either by ionic exchange or by co-precipitation with Na2MoO4 exhibit higher basicity. The values of the base sites numbers obtained for Mo-HTIE-Na2MoO4 and Mo-HTHS-Na2MoO4 are slightly lower than those characteristic to HT-CO3 2− (e.g. 6.47 mmol weak base sites/g and 0.26 mmol strong base sites/g [30]). Meanwhile, those obtained for the corresponding Mo-CHT samples are also just a little bit lower than those obtained for the mixed oxide CHT obtained from HT-CO3 2− precursor (e.g. 7.98 mmol weak base sites/g and 0.38 mmol strong base sites/g [30]).

Table 4 Surface base sites of the samples determined in fully hydrated conditions and by irreversible adsorption of organic acids
Full size table

Based on the above mentioned aspects it may be concluded that both the presence of the base sites as well as the presence of tetrahedral coordinated Mo species are required for the oxidation of t-BuSH. In the first stage, the t-BuSH molecule is activated by dissociation in the presence of the base sites of the solid. In the case of Mo-HT samples, the OH− ions in the interlayer act as Brønsted base sites with a stronger basicity than MoO4 2−, or Mo7O24 6− and NO3 − ions. In the case of Mo-CHT samples, the base surface sites consist of OH− groups (low strength), M–O pairs (medium strength) and O2− (strong) [40]. After the dissociation the t-BuSH− anion is adsorbed preferentially on tetracoordinated Mo6+ sites due to the fact that these ones can undergo easily a change in their coordination state. The Mo6+ is then reduced to Mo5+ and a radical t-BuSH* is generated. Two t-BuSH* radicals will combine to generate the dialkyl disulfide t-Bu-S–S-t-Bu while the reoxidation of the Mo sites will be ensured by the oxygen in the air flow. Further studies using XPS measurements in order to confirm the proposed mechanism are going to be performed in the near future.

Since the calcined catalysts exhibited better catalytic activity, they were selected for further testing in order to establish the optimal reaction conditions. In a first series of tests, the influence of the concentration of catalyst in the reaction mixture has been investigated. The results presented in Fig. 8a show that, for all catalysts, the conversion of t-BuSH increases almost linearly when the concentration of catalyst in the reaction mixture varies from 0.8 up to 1.3 wt%. A further increase of the concentration of the catalyst does not lead to significant increases of the conversion. The highest level of the conversion (80%) is reached with the catalyst Mo-CHTIE-Na2MoO4. For the less active catalyst, the conversion does not exceed 30% even at the highest concentration of solid in the reaction mixture. Taking these results into account, it may be concluded that a concentration of 1.5 wt% catalyst in the reaction mixture would be preferable for obtaining high conversion levels.

Fig. 8
figure 8

Results of catalytic tests for t-BuSH oxidation using 100 mL gasoline with Co t-BuSH = 128.5ppm and Mo-containing catalysts (filled square Mo-CHTHS-Na2MoO4, filled diamond Mo-CHTIE-Na2MoO4, filled triangle Mo-CHTHS-(NH4)6Mo7O24); a effect of catalyst concentration (air flow rate 5 L/h, 4 h reaction time); b effect of air flow rate (concentration of catalyst 1.5 wt%, 4 h reaction time); c effect of reaction time (concentration of catalyst 1.5 wt%, air flow rate 6 L/h)

Full size image

In the second series of tests, the effect of the air flow rate on the conversion of t-BuSH was investigated using the previously selected concentration of catalyst in the reaction mixture. The results plotted in Fig. 8b show that the conversion of t-BuSH increases sharply when the air flow rate varies from 1 up to 5 L/h, while the further increase of the air flow rate does not lead to significant improvements. Therefore it may be considered that an air flow rate of 6 L/h would be optimal for further experiments.

In the third series of tests, the effect of the reaction time on the catalytic activity was investigated using the optimal concentration of catalyst in the reaction mixture and the optimal air flow rate selected. The results plotted in Fig. 8c show that after 4 h reaction time, almost stable values of the conversion of t-BuSH are obtained and there is no decrease of the catalytic activity during 10 h. The fact that the total conversion of this reactant cannot be reached even when the reaction time is extended may be a consequence of two factors: (i) the lower reactivity of t-BuSH compared to the other light alkane thiols due to its molecular structure, (ii) the partial screening of the active sites of the catalyst by the reaction products (disulfide and water) during the process.

The fourth set of experiments aimed to investigate the resistance of the catalysts under operating conditions and the possibility of using them in repeated reaction cycles (4 h each). The results (Fig. 9) indicate that the catalytic activity decreases gradually with about 8% after six reaction cycles. It may be supposed that the deactivation occurs due to the adsorption of the reaction products on the surface of the solids, even though the catalysts had been thoroughly washed with diethyl ether in order to remove the adsorbed disulfide before being used in the next cycle.

Fig. 9
figure 9

Results of catalytic tests for t-BuSH oxidation, air flow rate 6 L/h, concentration of catalyst 1.5 wt%, 100 mL gasoline with Cot-BuSH = 128.5 ppm, repeated reaction cycles

Full size image

The gasoline recovered after six reaction cycles was collected and submitted to ICP-OES analysis (ASTM-E 1479-92) in order to examine the leaching of Mo, Mg, and Al. For both catalysts Mo-CHTIE-Na2MoO4 and Mo-CHTHS-Na2MoO4, the concentrations of Mg, Al and Mo in the recovered liquid phase were lower than the detection limits (e.g. Mg < 0.1 ppm, Al < 0.02 ppm and Mo < 0.05 ppm), suggesting that leaching did not occur under reaction conditions.

Conclusions

The obtained results showed that the higher activity towards the oxidation of t-BuSH is strongly related to two factors: (i) the presence of tetrahedral coordinated Mo species well dispersed in the structure of the base solid, and (ii) the amount of base sites in the solid. The preparation of Mo-HT solids at pH 10, using Na2MoO4 as molybdenum source allows the obtaining of crystalline materials with relatively high basicity and fine dispersion of molybdate species. These solids exhibit the best catalytic activity for the oxidation of t-BuSH (e.g. about 80% conversions). The preparation of Mo-HT solids at pH 10, using (NH4)6Mo7O24 as molybdenum source and NH4OH for pH adjustment, lead to less crystalline solids with smaller surface area and lower basicity. In these solids, which had the lowest catalytic activity, Mo was present mainly as octahedral coordinated species.