Dehydrochlorination of 1,2-dichloroethane over Ba-modified Al₂O₃ catalysts

Abstract

Bimodal mesoporous alumina (Al₂O₃) was prepared using polyethyleneglycol (PEG 20 000) and cetyl trimethyl ammonium bromide as a template. The incorporation of Ba with various loadings was carried out by incipient wetness. Characterization was performed by XRD, N₂ sorption isotherms, and pyridine FTIR. Ba can be highly dispersed on Al₂O₃ covering the strong acid sites of Al₂O₃. In the catalytic dehydrochlorination of 1,2-dichloroethane (1,2-DCE), the Ba/Al₂O₃ catalysts present a high activity, of which Al₂O₃ is most active with 95% conversion at 325 °C, related to the more Lewis acidic Al³⁺ sites in a tetrahedral environment. 1,2-DCE adsorbs dissociatively on Lewis acid–base pair sites, forming chlorinated ethoxy species, which are supposed to be intermediate species for vinyl chloride (VC) production. At a temperature higher than 400 °C, the dehydrochlorination of VC occurs on the strong acid sites of Al₂O₃. Ba can promote greatly the selectivity for VC through a decrease in the strong acid sites. A high stable activity for dehydrochlorination and high selectivity for VC can be obtained over Ba/Al₂O₃ in the presence of oxygen.

---

1. Introduction

Vinyl chloride (VC) has been widely used in the production of popular homopolymeric and copolymeric plastic materials. The demand for VC as a basic commodity increases greatly with the application of vinyl plastic materials. The production of VC through chemical dehydrochlorination of l,2-dichloroethane (l,2-DCE) has been practised on a large scale, of which thermal dehydrochlorination of l,2-DCE at 450–500 °C is industrially a main route with 50% conversion and 98–99% selectivity for VC.^1,2 Catalytic dehydrochlorination is highly effective in VC production. Generally, it was considered that the dehydrochlorination of chloroalkanes was promoted by solid base and acid catalysts. Shalygin reported l,2-DCE dehydrochlorination over a series of silicate catalysts, such as Al₂O₃/SiO₂, Ga₂O₃/SiO₂, ZrO₂/SiO₂, BeO/SiO₂ and Y₂O₃/SiO₂, which are acidic solid materials.¹ There are a few reports on the interaction between chloroalkanes and Al₂O₃. Mochida studied the dehydrochlorination of several chloroalkanes (including 1,1,2-trichloroethane, 1,2-dichloropropane and 1,1,2-di-chloropropane) over Al₂O₃, base and acid catalysts under reductive reaction conditions.³ Dehydrochlorination reactions over dry Al₂O₃ were postulated to proceed through an E2-concerted mechanism, where the chlorine and hydrogen were eliminated almost simultaneously with the basic and acid sites of Al₂O₃. Feijen-Jeurissen⁴ reported the mechanism of catalytic decomposition of 1,2-DCE over γ-Al₂O₃, and proposed that the destruction of 1,2-DCE occurs through dehydrochlorination to VC. Catalytic cracking of DCE is also practised industrially over silicates, metal-promoted alumina, or zeolites, but the advantage balances the drawback of needing catalyst regeneration. The dehydrochlorination of chloride linear paraffins to linear olefins in the production of alkylbenzene sulfonate surfactants was performed catalytically in the presence of silica alumina or on metal packings of reactor columns acting as catalysts. In this case, no information is readily available in the open literature concerning catalyst stability.

Recently, transitional alumina has been among the most used materials in any field of technologies, while details of its physicochemical properties are still a matter of discussion and investigation.⁵ The crystal structure of γ-Al₂O₃ is still a matter of controversy, having a defective non-stoichiometric spinel^6,7 or other cubic or tetragonal structures with occupancy of non-spinel cationic sites.^8–11 Digne et al.^12,13 reported DFT results showing that the two main orientations (under practical operation conditions) of γ-Al₂O₃ facets were 〈100〉 (17%) and 〈110〉 (70%). The 〈100〉 facets were fully dehydrated at 600 K, leading to the formation of coordinatively unsaturated (penta-coordinate) Al³⁺. As reported, the acidity of γ-Al₂O₃ was related to coordination degree of the Al³⁺ species, of which the strongest Lewis acid sites were associated with very low coordination Al cations, such as tri- and tetra-coordinated Al³⁺. In the dehydration of ethanol, the most active sites were believed to be Lewis acidic Al³⁺ sites in a tetrahedral environment located on the edges and corners of the nanocrystals.¹⁴ In this paper, the surface structure of Al₂O₃ was modified with dropping Ba, and the effect of the corresponding acidity of Al₂O₃ on the activity, selectivity and stability in the catalytic dehydrochlorination of l,2-DCE was investigated. TPSR and in situ DRIFTS techniques were utilized to determine reaction intermediates and thus explore a possible reaction mechanism.

2. Experimental

2.1. Catalyst preparation

Bimodal mesoporous alumina (Al₂O₃) was synthesized using polyethyleneglycol (PEG 20 000) and cetyl trimethyl ammonium bromide (CTAB) as a template. The typical procedure is described in ref. 15. CTAB (1.38 g), PEG (2.25 g) and aluminium isopropoxide (Al(O-i-Pr)₃ 5 g) were dissolved in ethanol aqueous solution (63% v/v) with vigorous stirring, and then ammonia (19 mL) as a precipitation agent was added and stirred for 30 min at 50 °C. The produced suspension was aged at 50 °C for 24 h under static conditions. The obtained filter was washed with ethanol and dried at 110 °C overnight. Dried samples were calcined for 3 h at 550 °C in air. Commercial Al₂O₃ (Al₂O₃-C) was referred to as a reference.

A series of Ba/Al₂O₃ catalysts with various Ba loadings were prepared by conventional impregnation methods described elsewhere.⁹ The support Al₂O₃ (2 g) was impregnated with an aqueous solution (4 mL) (the content of barium nitrate was in a range of 0.0048–0.095 g mL⁻¹ (0.05–10 wt%)) and then dried at 80 °C for 12 h and calcined in air at 550 °C for 3 h. The obtained catalysts were denoted as xBa/Al₂O₃ where x presents the content of Ba (wt%). Al₂O₃ used in this work is bimodal mesoporous Al₂O₃ unless stated otherwise.

2.2. Catalyst characterization

The powder X-ray diffraction (XRD) patterns of the samples were recorded on a Rigaku D/Max-rC powder diffractometer using Cu Kα radiation (40 kV and 100 mA). The diffractograms were recorded within the 2θ range of 10–80° with a 2θ step of 0.01° and a time step of 10 s. The nitrogen adsorption and desorption isotherms were measured at 77 K on a Micromeritics ASAP 2400 system operated in static measurement mode. Samples were outgassed at 350 °C for 4 h before the measurement. The specific surface area was calculated using the BET model. The actual Ba content was determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) using a Varian 710 spectrometer. Samples were dissolved using an aqua regia-hydrogen peroxide system to form a homogeneous solution. The X-ray photoelectron spectroscopy (XPS) measurements were made on a VG ESCALAB MK II spectrometer using Mg Kα (1253.6 eV) radiation as the excitation source. Charging of the samples was corrected by setting the binding energy of adventitious carbon (C 1s) at 284.6 eV. The powder samples were pressed into self-supporting disks, and loaded into a sub-chamber, which was evacuated for 4 h prior to the measurements at 298 K. Temperature programmed surface reaction (TPSR) measurement was carried out under the same conditions as those in the catalytic activity tests. First, the feed containing 1000 ppm 1,2-DCE and Ar balance flowed continuously over the samples at 100 °C. After the adsorption–desorption reached an equilibrium, the samples were heated from 100 °C to a specified temperature at a heating rate of 10 °C min⁻¹. The reactant and the products (such as DCE (m/z = 98), CO₂ (44), CO or C₂H₄ (28), ⁺CHO (29), Cl₂ (70), HCl (36), C₂H₂ (26), C₂H₃Cl (62), C4 (41, CH₂=CH–CH₂⁺)) were analyzed on-line over a mass spectrometer apparatus (HIDEN, QIC-20).

2.3. Catalytic activity measurements

Catalytic dehydrochlorination was carried out at atmospheric pressure in a continuous flow microreactor (a quartz tube of 3 mm inner diameter). 75 mg catalyst (grain size, 40–60 mesh) was packed in the reactor bed. Before testing the activity and selectivity, the transport effects were investigated to ensure that the experimental results were not significantly influenced by interphase transportation. Calculation of the theoretical external transfer rate of reactants to the catalytic particles (at a typical temperature of 300 °C) based on estimated mass transfer coefficients gave a value magnitude three orders greater than the measured reaction rates, indicating that the process conditions were far from external diffusional limitations. The effects of external mass transfer resistances were experimentally evaluated by repeating a set of process conditions whilst employing a different linear velocity. The results of these experiments indicated that the conversion was not affected for a linear velocity higher than 7 cm s⁻¹, within the experimental error. Likewise, estimates of the interphase temperature gradients showed fluid-solid differences of less than 1 °C. The possibility of internal pore diffusion was examined by measuring conversions at fixed conditions but varying the catalyst particle size. Results showed that the pore diffusional resistance was absent for particles less than 1 mm in diameter. Intraparticle mass transfer resistances were theoretically evaluated by computing effectiveness factors, which were calculated to be greater than 0.98, indicating that intraparticle mass transfer resistances were not significant. Finally, internal thermal gradients also proved to be negligible over the range of conditions evaluated in this study. The feed flow through the reactor was set at 40 mL min⁻¹ (linear velocity of 9.4 cm s⁻¹) and the gas hourly space velocity (GHSV) was maintained at 30 000 h⁻¹. The feed stream to the reactor was prepared by delivering liquid 1,2-DCE with a syringe pump into dry Ar and the injection position was electrically heated to ensure complete evaporation of the liquid reaction feeds. The temperature of the reactor was measured with a thermocouple located just at the bottom of the microreactor. The effluent gases were analyzed on-line at a given temperature using gas chromatography (GC9790, FULI) equipped with a Kromat-KB-5-30 m × 0.32 mm × 0.50 μm capillary column and a flame ionization detector (FID) for quantitative analysis of the organic chlorinated reactant. The catalytic activity was measured over the range 150–450 °C and the conversions were calculated by subtracting the outlet concentration from the inlet concentration of the reactant and dividing by the inlet concentration. These conversions were obtained at different temperatures under a steady state at each temperature. All the reactions were repeated three times to assure reproducibility. Furthermore, carbon balances could be as accurate as within 5%.

2.4. In situ FTIR

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments were conducted on a Nicolet 6700 FTIR fitted with a liquid nitrogen-cooled mercury–cadmium–telluride detector (MCT). The DRIFTS cell (Harrick, HVC-DRP) fitted with ZnSe windows was used as the reaction chamber that allowed samples to be heated to 550 °C. All the spectra were obtained averaging 64 scans. Considering the instrumental optics and the strong framework adsorption of alumina, the useful spectral range was 4000–1100 cm⁻¹. Prior to 1,2-DCE adsorption experiments, the samples were pre-treated in Ar at 550 °C for 2 h. Then the samples were cooled down to 50 °C in order to remove the contaminants. The spectra of the samples (20–30 mg) were recorded from 100 to 400 °C after 1,2-DCE adsorption following sweeping with Ar.

3. Results and discussions

3.1. Catalyst characterization

3.1.1 Physical properties. Fig. 1 shows N₂ sorption isotherms of the samples with the pore size distribution determined based on the adsorption branch using the BJH method. The synthesized Al₂O₃ presents two continuous type-IV curves with an H1-type hysteresis loop, which locate at a relative pressure of 0.4–0.55 and 0.55–0.95, respectively. Two kinds of pore appear with a radius of 3.9 and 7.4 nm in the pore distribution curve, respectively. The BET area is estimated to be 330 m² g⁻¹ (Table 1). With the incorporation of Ba into Al₂O₃ by 1–2 wt%, similar N₂ sorption isotherms are observed. A further increase in the Ba loading makes the two kinds of pore decrease to 3.7 and 5.0 nm, respectively. BET areas of the Ba/Al₂O₃ samples are in the range of 310–334 m² g⁻¹ (Table 1). Commercial Al₂O₃ (Al₂O₃-C) has a narrow pore size distribution centered at 6.1 nm with a surface area of 117.9 m² g⁻¹. XRD patterns of the samples are presented in Fig. 2. All of the samples exhibit diffraction peaks at 19.4, 37.6, 45.8 and 67.0°, ascribed to the 〈111〉, 〈311〉, 〈400〉 and 〈440〉 planes of γ-Al₂O₃, respectively (JCPDS 10-0425). Note that no diffraction peaks due to Ba species are detected for the Ba/Al₂O₃ samples with ≤3 wt% Ba loading, implying that Ba species are highly dispersed on the surface of Al₂O₃. With a further increase in the Ba loading up to 4 wt% or higher, new diffraction peaks appear at 23.9, 24.3, 33.7, 34.1, 34.6 and 42.0°, which result from the BaCO₃ crystalline phase (JCPDS card no. 05-0378). The formation of BaCO₃ should be related to the interaction of BaO with CO₂ during the calcination due to the strong basicity of BaO.

Fig. 1 N₂ sorption isotherms (a) and pore size distribution (b) of Al₂O₃, Ba/Al₂O₃ and Al₂O₃-C samples.

Table 1 Physicochemical properties of Al₂O₃ and Ba/Al₂O₃ with various Ba loadings

Sample	SBET (m^2 g^−1)	Vpore (cm^3 g^−1)	Dpore (nm)	Total acidity^a (mmol per g cat)		Activity (°C)
				Weak acid	Strong acid	T50	T95
a Determination by NH3-TPD tests.
Al2O3	330	0.434	3/7	0.149	0.082	308	355
1Ba/Al2O3	301	0.366	3/6	0.162	0.100	315	360
2Ba/Al2O3	321	0.383	3/6	0.173	0.070	329	381
4Ba/Al2O3	334	0.390	2.7/5	0.108	0.060	338	404
10Ba/Al2O3	304	0.302	2.5/5	0.101	0.052	392	472
Al2O3-C	118	0.253	6	0.088	0.043	386	441

---

Fig. 2 XRD patterns of Al₂O₃ and Ba/Al₂O₃ samples with various Ba loadings.

Fig. 3 shows HRTEM images of the synthesized Al₂O₃ and 4Ba/Al₂O₃ samples. With the presence of Ba, no evident change in morphology is observed. The size of the observed Al₂O₃ particles is in the range of 3–7 nm (white circles in Fig. 3a). The small Al₂O₃ particles were stacked together to form a large plane. The lattice spacing was measured to be 0.14 and 0.20 nm, in good agreement with those of the 〈440〉 and 〈400〉 crystal planes of the standard γ-Al₂O₃ sample (JCPDS 10-0425). The observation of electron diffraction rings in the selected area electron diffraction (SAED) patterns (inset of Fig. 3a) suggests the formation of a polycrystalline structure. The images also show that the catalysts are homogeneous with the absence of crystalline Ba oxide or BaCO₃ phases. To further characterize the distribution of Ba in 4Ba/Al₂O₃, STEM mapping of Ba, Al and O was conducted. As shown in Fig. S1,† Ba species were uniformly and highly dispersed on the Al₂O₃ surface. This result shows the strong interaction between Ba and Al₂O₃, as reported elsewhere.^9,16,17 A low BaO coverage of 2 wt% on γ-Al₂O₃ monomeric BaO units was present almost exclusively and these molecularly dispersed BaO units were concentrated on the (100) faces of the alumina crystallites.⁹ Density functional theory calculations predicted that the energetically most favorable BaO monomer and dimer units anchor to the penta-coordinate Al³⁺ sites on the (100) faces of γ-Al₂O₃ in such geometries that maximize their interactions with the support surface.¹⁶ In our case, Al₂O₃ exposed mainly 〈400〉, which is really different from the above results.^9,16 The BaO monomer and dimer units should be formed because the aggregation of BaO can not be observed on 4Ba/Al₂O₃ where the Ba species is highly dispersed on Al₂O₃ (confirmed by HRTEM). Additionally, the crystalline particles of BaCO₃ are not detected by HRTEM, indicating that XRD diffraction from BaCO₃ may be due to a separate phase from Al₂O₃ produced during calcination.

Fig. 3 HRTEM and SAED (inset) images of Al₂O₃ (a) and 4Ba/Al₂O₃ (b) samples.

3.1.2 Pyridine adsorption. Pyridine, as a basic probe molecule, can be used for characterization of catalyst surfaces, allowing also a definite determination of the existence of Lewis acidity. The spectra recorded after adsorption of pyridine on the Ba/Al₂O₃ samples are presented in Fig. 4. In the Py-FTIR spectra, the typical features of Lewis bonded pyridine can be observed. Over Al₂O₃, three 8a components observed at 1574–92, 1612 and 1624 cm⁻¹ reveal the existence of at least three different families of Lewis acid sites, as discussed previously for transitional alumina.^5,18–22 The typical assignments for the first and the last of these three components are to pyridine bonded to octahedral and tetrahedral Al³⁺ ions, respectively, both with single coordinative unsaturation, thus being penta- and tri-coordinated, respectively, before pyridine adsorption. The band in the middle can result from pyridine species interacting with ​tetra-coordinate Al³⁺ ions that have a Lewis acid strength slightly lower than that of tri-coordinated Al³⁺ ions. These sites may be tri-coordinated too, but with a nearest cation vacancy. The spectra observed for the samples containing Ba are very similar, as also are most spectra reported in the literature for pyridine adsorbed on alumina.²³ As previously reported, the penta-coordinate aluminum ions on γ-Al₂O₃ were identified as the preferential anchoring points for Ba¹⁶ and Pt.¹⁷ It should be noted that with an increase in the Ba loading, the intensity of the band at 1624 cm⁻¹ decreases gradually and the 8a band at 1574–92 cm⁻¹ becomes weaker and narrower. This is possibly due to the coverage of Ba species on tri- and penta-coordinated Al³⁺ ions, thus leading to a decrease in the pyridine molecules adsorbed on these sites. In the spectrum of commercial Al₂O₃-C, the features of pyridine adsorbed on Lewis sites are also evident. However, the band at 1624 cm⁻¹ due to pyridine species interacting with tri-coordinated Al³⁺ species becomes weak, indicating that the number of strong Lewis acid sites on Al₂O₃-C is smaller than that on Al₂O₃ and Ba/Al₂O₃ with a low Ba loading.

Fig. 4 Py-FTIR spectra of Al₂O₃ and Ba/Al₂O₃ samples at 200 °C.

3.1.3 NH₃-TPD analyses. Fig. 5 shows the NH₃-TPD profiles of the Al₂O₃ and Ba/Al₂O₃ samples. The NH₃-TPD profiles can be divided into two desorption temperature ranges of 150–300 °C and 300–450 °C, corresponding to ammonia desorption from weak and strong acidic sites, respectively. The weak and strong acid amounts of the sample with 1% Ba increase by 8% and 23% compared with those of Al₂O₃, respectively (Table 1). As mentioned previously, the addition of Ba will create an interface between Al–O and Ba–O species which distorts the surface structure of γ-Al₂O₃, probably with electronic unbalance due to the change in length of the Al–O bonds, which can contribute to the increase in acidity. On a further increase in the Ba loading, however, both strong and weak acids decrease significantly. Additionally, BaO is a solid base, and it is expected that the acid amount decreases with the formation of BaO particles over Al₂O₃. Thus, for the samples with a Ba loading of higher than 4%, the total acidity is inversely proportional to the Ba content (Fig. 5, inset). In the Py-FTIR spectra of all samples, characteristic peaks appear of pyridine adsorbed on Lewis acid sites at ca. 1442, 1574–92, 1612 and 1624 cm⁻¹ (Fig. 4).¹⁹ Obviously, Al³⁺ ions coordinated with different numbers and Ba²⁺ ions for the samples with a low Ba content contribute to the Lewis acid sites.

Fig. 5 NH₃-TPD profiles and the amount of NH₃ desorbed (inset) at 125–300 °C and 300–450 °C from Al₂O₃ and Ba/Al₂O₃ samples with various Ba loadings.

3.2. Dehydrochlorination

3.2.1 Activity test. The total conversion of 1,2-DCE over Al₂O₃ and Ba/Al₂O₃ catalysts in a dry feed of 1000 ppm 1,2-DCE and Ar balance as a function of temperature is shown in Fig. 6. Al₂O₃ presents a considerably high activity and the conversion of 1,2-DCE increases quickly on raising the temperature. T_95% (the temperature needed for 95% conversion of 1,2-DCE) is 325 °C. The main products are composed of VC, aldehyde, ethyne and C₄ (butene or chloro-butene) (Fig. 6), which are confirmed by TPSR (Fig. S2†). C₄ appears only at high temperature with a low selectivity (below 1%). The selectivity for VC is promoted by raising the temperature with a significant decrease in the aldehyde content and reaches 95% at 350 °C. After that, the selectivity for VC decreases quickly, and is only 60% at 425 °C. Generally, it is considered that the formation of VC is through the dehydrochlorination of 1,2-DCE.⁴ The Cl atom of 1,2-DCE adsorbs on the Al³⁺ acidic site and surface O²⁻ species or hydroxyl groups from the surface of γ-Al₂O₃ do a nucleophilic attack on the carbon atom of 1,2-DCE and dehydrochlorination occurs. The formation of aldehyde was considered to be the result of the reaction of VC with surface hydroxyl groups over acidic catalysts.²⁴ As the reaction proceeds with the increase in temperature, the surface oxygen species become less and less, and the formation of aldehyde becomes difficult. At a temperature of 375 °C or higher, VC can be further dehydrochlorinated into ethyne.

Fig. 6 The conversion curves of 1,2-DCE and TOF (inset, based on the mole number of converted 1,2-DCE at 300 °C per second per acidic site) over different catalysts. Reaction gas: 1000 ppm 1,2-DCE and Ar balance; SV = 30 000 mL g⁻¹ h⁻¹.

With incorporation of Ba, the conversion curve shifts to a high temperature. T_95% is proportional to the Ba loading (Table 1) and increases from 350 °C for Al₂O₃ to 471 °C for 10Ba/Al₂O₃, indicating that the activity is inhibited by Ba to some extent. A similar product distribution over catalysts containing Ba is available. However, the aldehyde distribution becomes broader and higher with an increase in the Ba loading and the selectivity for aldehyde over 4Ba/Al₂O₃ reaches 60% at 225 °C. In all cases, the higher the 1,2-DCE conversion is, the higher the VC selectivity is; and the lower the DCE conversion is, the higher the aldehyde selectivity is. At complete conversion, aldehyde is not formed at all, suggesting that the 1,2-DCE partial pressure available is a key factor favoring aldehyde (at low conversion, with more aldehyde available) or VC (at high conversion). In fact, the dependence of the aldehyde formation is expected to have a higher reaction order with respect to 1,2-DCE than VC synthesis. On the other hand, the presence of Ba can produce additional oxygen species as strong basic sites, and so may be favorable for the formation of aldehyde through nucleophilic attack (see later). Moreover, the formation of ethyne during increasing the temperature decreases gradually with Ba addition. At 425 °C, the selectivity for ethyne decreases from 60% over Al₂O₃ to 5% over 4Ba/Al₂O₃. The elimination of HCl from chloroalkanes is promoted by solid base and acid catalysts. Generally, the reactivity of the chlorinated ethylenes decreases with the increasing chlorine content in the molecule. The rate determining step probably does not involve breaking the C–Cl or C–H bonds. Bond et al.²⁵ suggested that over a Pt/Al₂O₃ catalyst, the rate determining step is the removal of a chlorine atom. Chintawar et al. extensively studied the oxidation of chlorinated ethylenes.²⁶ They found a strong correlation between the adsorption capacity of the molecule on the catalyst and the reactivity of the molecule. The elimination reactions of the chlorocompounds over Al₂O₃ in a dry feed were postulated to proceed through an E2-concerted mechanism where the chlorine and proton were eliminated almost simultaneously by the acidic and basic sites of Al₂O₃. If the acidity of the proton was weak enough, such as that in the case of VC, the removal of the Cl atom on the strong acid sites was critical. At that time, the elimination may proceed via an E1-concerted mechanism. For Ba/Al₂O₃ catalysts, the decrease in strong acid sites is really not favorable for the activation of VC, even though the basicity of the Ba/Al₂O₃ catalysts is strong. As expected, Al₂O₃-C with less strong acidity shows a poor selectivity for ethyne at high temperature. Al₂O₃-C with less strong basic sites (basic oxygen species such as BaO possesses) presents a lower selectivity for aldehyde (Fig. 7).

Fig. 7 The distribution of products obtained in the reactions under the same conditions as those described in Fig. 6.

In order to investigate the effect of acidity on catalysts on the kinetics of 1,2-DCE dehydrochlorination, TOF based on the mole number of converted 1,2-DCE molecules per second per mole of acidic site on the surface of catalysts can be used for comparing the activity difference among the acidic sites of Al₂O₃ and Ba/Al₂O₃. Fig. 6 (inset) shows the TOF at 300 °C as a function of Ba loading. It can be seen that the TOF decreases almost linearly with the Ba loading. A few studies have been published on the interaction between chloroethanes and Al₂O₃. A significant difference in TOF indicates that the rate determining step probably does not involve breaking C–Cl bonds. The removal of Cl species from the surface may be a slow step (Fig. 8).²⁷

Fig. 8 Correlation of ethyne selectivity with I₁₆₂₄/I₁₄₄₈ for catalysts at various temperatures. Reaction gas: 1000 ppm 1,2-DCE and Ar balance; SV = 30 000 mL g⁻¹ h⁻¹.

3.2.2 The stability. Considering the possible effects of Cl and carbon depositions on the surface of catalysts, stability tests in the feed of 1000 ppm 1,2-DCE and Ar balance were carried out at 400 °C and the results are shown in Fig. 9. It can be seen that the conversion on Al₂O₃ and 4Ba/Al₂O₃ decrease by 50% on stream within the first 45 and 18 h, respectively. The used catalysts in the stability test became black. EDS analyses showed that there was 1.5–2.7% carbon deposition. This phenomenon implies that polymerization, fuse and carbonization can occur during the reaction. Moreover, about 2% Cl was detected by XPS and EDS analyses. As is known, the removal of Cl species deposited on Al₂O₃ can be considered to be a rate-controlling step because of the strong adsorption of Cl species.²⁸ The reaction of chloro-organics with Al₂O₃ catalysts is also described in the literature concerning the preparation of highly acidic chlorinated alumina catalysts. Chlorination was successfully performed by the reaction of chlorinated alkanes and alkenes containing at least two chlorine atoms. Several chlorinated alkanes and alkenes can be considered to be less reactive in chlorinating Al₂O₃.^28,29 Chlorine adsorbed over alumina, as reported by Muddada et al.,^30,31 reduced the amount and the acidic strength of the Lewis sites, saturating the coordinate vacancies of Al³⁺ by the formation of Al–Cl species. The higher deactivation rate of 4Ba/Al₂O₃ can be ascribed to stronger adsorption of HCl on basic sites. As reported in our previous work, the removal of Cl species from the catalyst surface can be promoted in the presence of water through providing hydrogen atoms.²⁷ In this work, the deactivation of Al₂O₃ can not be inhibited by adding water at 350 °C (Fig. S3†), although water promotes the activity of the fresh catalysts at low temperature to some extent (Fig. S4†). XPS and EDS analyses showed that Cl deposition on the surface of the used Al₂O₃ and 4Ba/Al₂O₃ on stream of the wet feed at 350 °C decreases by 50% or more. In the case of the reaction of dichloromethane with water, the balance of a small amount of Cl adsorbed on the Al₂O₃ catalyst can be maintained and the activity becomes constant at 300 °C.²⁷ Obviously, the deactivation of Al₂O₃ on the wet stream at 350 °C should be caused only by carbon deposition (Fig. S3†). When Al₂O₃ and 4Ba/Al₂O₃ became deactivated during the stability test, with the addition of 5% O₂ (Fig. 9), the activity of the catalysts can restore almost to the level obtained on the fresh catalysts at 400 °C. TG-MS confirmed that the removal of black carbon deposited on the surface of the catalysts in the presence of oxygen occurred at 430 °C. In fact, with the addition of oxygen, the activity of the dehydrochlorination over Al₂O₃ can not be restored completely at 350 °C (Fig. S5†). Here, the removal of Cl species as HCl from the surface of the catalysts was promoted by water produced from the oxidation of a small amount of 1,2-DCE.

Fig. 9 The stability on the feed streams at 400 °C of the Al₂O₃ (a) and 4Ba/Al₂O₃ (b) catalysts. Reaction gas: 1000 ppm 1,2-DCE and Ar balance; SV = 30 000 SV = 30 000 mL g⁻¹ h⁻¹.

In the dry and oxygen-free feed, the selectivity of VC increases gradually with the deactivation of Al₂O₃ and reaches 96% at a conversion of 50%, due to the decrease in strong acidic sites through carbon deposition. However, with the removal of carbon from the strong acidic sites in the feed containing oxygen, the dehydrochlorination of VC becomes significant, and the selectivity for VC finally decreases to about 65%. This phenomenon indicates that the strong acidic sites on which black carbon deposits are responsible for the second dehydrochlorination. It is interesting to find that for 4Ba/Al₂O₃, the selectivity for VC maintains 90% during the stability test (at least 80 h) and almost a zero amount of ethyne was detected. 4Ba/Al₂O₃ in particular was tested in the feed containing 5% oxygen for another 80 h at 400 °C with a conversion as high as 90–92% and the selectivity for VC was as high as 88–90% (ESI†). This result is really different from those obtained previously, in which catalytic cracking at 300–400 °C on pumice (SiO₂, Al₂O₃, alkalis) or on charcoal, impregnated with BaCl₂ or ZnCl₂ has not found more widespread application due to the limited life of the catalysts of about 10 h and VC can be further oxychlorinated to ethyl trichloride and further transformed in the presence of oxygen.³¹ To the best of our knowledge, the catalysts used for dehydrochlorination of chlorinated linear paraffins were acidic catalysts, such as silicates and zeolites.^1,3,32,33 Similar results were found in the dehydrochlorination of 1,2-dichloropropane over silica-alumina catalysts where 20% selectivity for allyl chloride and 10 h stability were available.³⁴ The theoretical and experimental studies showed that the addition of Ba species can cover strong acidic sites of Al₂O₃.¹⁶ It provided an opportunity to develop a new pathway of 1,2-DCE dehydrochlorination over catalysts with strong basicity and weak acidity. The synergy of a strong base and weak acid is critical to ensure a good performance of 4Ba/Al₂O₃. The carbon balance can reach 95% or higher. CO can be detected in effluent due to the oxidation of 1,2-DCE. Inorganic Cl species leave from the reactor as HCl but not as Cl₂. In fact, a significant Deacon reaction (HCl + O₂ → Cl₂ + H₂O) occurs at 700 °C or higher over Al₂O₃.³⁵

In order to understand the high selectivity for VC in the presence of oxygen, the oxidation of VC over Al₂O₃ and 4Ba/Al₂O₃ was conducted. The activity of Al₂O₃ and 4Ba/Al₂O₃ for VC oxidation is poor, and the conversion reaches 56% and 20% up to 350 °C (Fig. S6†). Moreover, the addition of 1,2-DCE can retard the conversion of VC, indicating that 1,2-DCE is more favorable for adsorption on active sites. High selectivity for VC and a stable activity over 4Ba/Al₂O₃ in the presence of oxygen results from its high activity for dehydrochlorination of 1,2-DCE and low activity for the second dehydrochlorination and VC oxidation.

3.2.3 The effect of space velocity. The effect of space velocities on DCE dehydrochlorination over Al₂O₃ and 4Ba/Al₂O₃ was investigated at 275, 290 and 305 °C within the space velocities of 30 000–90 000 mL g⁻¹ h⁻¹. The results (Fig. 10 and S7†) show that the increase in space velocities does not decrease the rate (based on the mole number of 1,2-DCE converted per second per square meter), which is not as expected. The highest reaction rate was obtained at 60 000 mL g⁻¹ h⁻¹ and the rates at 30 000 and 90 000 mL g⁻¹ h⁻¹ are almost equal. These anomalous data arevassociated with the promotion of the removal of Cl species produced during the reaction by a higher linear rate of feed in the reaction bed, implying that at a lower feed rate, HCl desorbed from the surface can go back to the surface. Busca observed this anomalous phenomenon in the dehydration of ethanol over Al₂O₃ containing more chlorine impurities.¹⁴

Fig. 10 Relationship between the space velocities and rates of 1,2-DCE dehydrochlorination over Al₂O₃ catalyst.

3.3. In situ FTIR spectra

3.3.1 Al₂O₃. FTIR spectra collected at different temperatures during the treatment of the synthesized Al₂O₃ with the reaction feed of 1000 ppm 1,2-DCE in Ar for 1 h after the treatment in Ar at 550 °C are shown in Fig. 11. The adsorption of 1,2-DCE on Al₂O₃ has been studied between RT and 400 °C. The spectra were recorded in the absence of gas phase 1,2-DCE unless stated otherwise. It can been seen that at RT, intense bands at 2845 and 2945 cm⁻¹ and a weak band at 3040 cm⁻¹ (difficult to observe, due to a low signal-to-noise ratio) are observed, indicating that several types of C–H bonds are present. On raising the temperature, the peaks shift to high wavenumbers, 2899 and 2966 cm⁻¹ at 200 °C, indicating the modification of the chemical environment. In the section of the hydroxyl group, the spectra exhibited negative bands in the 3788–3637 cm⁻¹ range, which usually are assigned to strong surface hydroxyl groups, along with a weak positive band between 3630 and 3581 cm⁻¹ (centered at approximately 3608 cm⁻¹). These spectra suggest that the Al₂O₃ surface hydroxyl groups interact with 1,2-DCE molecules during adsorption, leading to the formation of weaker hydrogen-bonded OH groups.^36,37 After progressive heating in the range of 100–400 °C, the negative bands in the 3788–3637 cm⁻¹ range become strong with the temperature in parallel to the increase in the bands due to chlorinated ethoxy groups. Indeed, at RT, 1,2-DCE can react with surface hydroxyl groups. Vigué et al.³⁸ assumed that the reactivity of alumina surfaces toward halogenated molecules highly suggested that the substitution of surface OH groups by halogenide ions should be easier for the most basic OH groups, corresponding to monocoordinated hydroxyl groups. Consequently, the disappearance of the bands at 3788–3637 cm⁻¹ corresponds to the formation of Al–Cl bonds. In the frequency range between 1800 and 1000 cm⁻¹, weak bands at 1160 and 1194 cm⁻¹ corresponding to chlorinated ethoxy are observed within the experimental temperature, which shift to higher values, compared with those for CH₃–CH₂O– (1075 and 1116 cm⁻¹).³⁹ Previous investigations proposed that the first step in the catalytic oxidation of chlorinated methane over Al₂O₃ catalysts was a nucleophilic substitution.⁴⁰ During the nucleophilic substitution, the chlorine atom was abstracted and replaced by oxygen species, forming surface methoxy species and HCl gas.⁴⁰ It can be expected that the formation of chlorinated ethoxy species on Al₂O₃, as the first step of dehydrochlorination is a synergistic effect of nucleophilic attack by the basic surface hydroxyl groups and the abstraction of a Cl atom by an acidic Al³⁺ site. Meanwhile, the maxima at 1462 cm⁻¹ (δ_asC–H of CClH₂) and 1400 cm⁻¹ (δ_symCClH₂) are due to deformation modes of the CClH₂ group to which the CH₂ scissoring mode is superimposed. Compared with the case of CH₃, these C–H bands shift to high values slightly, probably due to substitution of Cl for hydrogen. It is interesting to note that, at low temperature, in the region from RT to 200 °C, the absorption band centered at 1647–1665 cm⁻¹ is predominant, accompanied by the appearance of two broad strong bands centered at 1409 and 1310 cm⁻¹. Three bands were attributed to a surface enolic species, as similar bands to those were observed on Al₂O₃ during the adsorption of CH₃CHO (not shown). One evidence is that the IR spectra of syn-vinyl alcohol (CH₂=CHOH) in the gas phase show a strong absorption band between 1644 and 1648 cm⁻¹, which is accompanied by two bands at 1409–1412 and 1300–1326 cm⁻¹.^41,42 The IR spectrum of adsorbed catechol on a TiO₂ colloid also gives a similar band at 1620 cm⁻¹.⁴³ Their common characteristic is an enolic structure. On the other hand, a band at 1605 cm⁻¹ grows substantially with temperature on stream. Dreoni et al.⁴⁴ assigned a strong band observed at 1598 cm⁻¹ during adsorption of cyclohexanone on silica at 200 °C to a C=C stretching vibration. Furthermore, these authors also assigned bands at 3075 and 3045 cm⁻¹ to unsaturated =CH– bond stretching. Indeed, a similar band at 3040 cm⁻¹ is also observed in our case during the adsorption of 1,2-DCE and can be assigned to such a vibration mode. Additional bands in the same region at 2939 and 2867 cm⁻¹ correspond to C=C and –CH– bonds and are also present in the spectra of adsorbed 1,2-DCE. The presence of C=C and –CH– bonds further verifies the formation of the enolic form.⁴⁴ The bands at 1386, 1268 and 1605 cm⁻¹ (assigned to δCH₂, ρCH and γC=C of VC) become significant up to 250 °C, where the conversion to VC in a parallel kinetic reaction is significant. At the same time, new bands appear at 1680 and 1730 cm⁻¹, ascribed to the –C=O group. Generally, adsorbed acetaldehyde corresponded to coordination to a Lewis acid, R–CH=O–Al³⁺.⁴ In fact, –CHO was detected in TPSR as m/z = 29, and in a parallel kinetic reaction, a significant amount of aldehyde in the effluent can be observed. Chintawar et al. observed for the adsorption of vinyl chloride on chromium-exchanged zeolite Y a band at 1678 cm⁻¹ and assigned this band to an adsorbed aldehyde or ketone.⁴⁰ Zhou observed the formation of aldehyde during the decomposition of 1,2-DCE at 260–360 °C.⁴⁵ VC can readily be protonated in the presence of acid catalysts, like AlCl₃, in the presence of OH surface species into a stable reactive carbonium ion which then was attacked by a nucleophilic oxygen species (basic site of alumina or adsorbed water), leading to the CH₃–CHCl–O species, which would readily decompose to form acetaldehyde and leave a chloride ion on the surface. Moreover, new bands appear at 1338–1356 and 1565 cm⁻¹ (assigned to carbonate bidentate), 1268, 1425 and 1538 cm⁻¹ (asymmetric stretching vibration of carboxylates of the acetate type),^40,46–48 and at 1470 and 1356 cm⁻¹ (monodentate bonded carbonates)⁴⁹ at 250 °C or higher, probably due to the presence of surface oxygen species, such as basic hydroxyl groups.

Fig. 11 In situ FTIR spectra in the 1100–4000 cm⁻¹ region for Al₂O₃ in a 1000 ppm 1,2-DCB/Ar stream from 50 to 400 °C after treatment in Ar at 550 °C.

3.3.2 4Ba/Al₂O₃. For 4Ba/Al₂O₃, the bands ascribed to 1,2-DCE adsorption appear at 2880 and 2965 cm⁻¹ (several types of C–H bond, Fig. 12), consistent with that observed on Al₂O₃. However, the band intensity seems very weak. Additionally, the band splits and various range sections can be observed, suggesting 1,2-DCE adsorption on two different types of sites associated with BaO and Al₂O₃. At the same time, the bands resulting from oxidized surface species such as carbonate bidentate (appearing at 1338–1356 and 1565 cm⁻¹), asymmetric stretching vibration of carboxylates of the acetate type (appearing at 1268, 1425 and 1540 cm⁻¹) and monodentate bonded carbonates (appearing at 1473 and 1356 cm⁻¹) become much stronger at 200 °C. Oxidation products were seen in the absence of gas phase oxygen, indicating involvement of surface oxygen in this process. As reported, formaldehyde and formic acid were detected in the spectra during the adsorption of dichloromethane on Al₂O₃.⁵⁰ In fact, their formation resulted from the nucleophilic attacks by surface hydroxyl groups or oxygen species to dichloromethane to produce methoxy species and the following disproportionation. Probably, these oxidized products in this work are related to incorporation of hydroxyl groups and basic oxygen species into the adsorbed 1,2-DCE molecules. A small amount of alkali metal was known to make alumina basic.⁵¹ The addition of Ba increases inevitably the strong basic oxygen species. As expected, there are more oxidized products on 4Ba/Al₂O₃. It should be noted that the band corresponding to the carbonyl group of the aldehyde is not observed on 4Ba/Al₂O₃, while aldehyde was detected in the effluent in a significant amount. Based on the fact that 4Ba/Al₂O₃ possesses less strong acid sites, it can be concluded that the aldehyde produced mainly adsorbs on strong acid sites as R–CH=O–Al³⁺. The same phenomenon is observed on Al₂O₃-C with less strong acid sites (Fig. S8†), where the carbonate bidentate, carboxylates of the acetate type and monodentate bonded carbonates can not be observed within the experimental temperature. These results imply that the oxidation involving surface oxygen species on Al₂O₃-C is weak, probably related to the lack of strong basic sites or strong acid sites.

Fig. 12 In situ FTIR spectra in the 1100–4000 cm⁻¹ region for 4Ba/Al₂O₃ in a 1000 ppm 1,2-DCB/Ar stream from 50 to 400 °C after treatment in Ar at 550 °C.

Based on these experiments it was postulated that the first step in the interaction of 1,2-DCE with Lewis acid sites occurs and the adsorbed 1,2-DCE can undergo nucleophilic attack by oxygen to form a chlorinated ethoxy intermediate through the removal of a chlorine atom. The formation of VC occurs through transfer of a proton and rearrangement of chlorinated ethoxy. The rearrangement of the ethoxy intermediate was considered to be an indispensable step in the dehydration of ethanol on pure Al₂O₃.¹⁴ At the same time, VC adsorbed on two sites, a pair of acidic-basic sites can be converted into carbonyl species which may or may not contain chlorine, and be stabilized by resonance (an enolic structure). And with the destruction of VC, the formation of aldehyde was proposed. A following oxygen attack on the carbonyl compound results in the formation of carboxylate and carbonate species.⁴⁰ On the other hand, on strong acidic sites, the chlorine atom of VC can be abstracted with the transfer of a proton, and ethyne is formed. In order to obtain deep insights into this reaction at a molecular level, a mechanism over Al₂O₃ consisting of three elementary steps is drawn (Fig. 13): (1) formation of the 1,2-DCE adsorption complex during the interaction of 1,2-DCE with Lewis acid sites; (2) chlorine abstraction by nucleophilic oxygen (surface hydroxyl groups) to form the chlorinated ethoxy intermediate; and (3) dehydroxylation through rearrangement to form VC. Other side reactions include: (1) attack of basic oxygen species to VC; (2) formation of oxygenate species, such as aldehydes, carbonate bidentates, monodentates and partially oxidized surface species such as enolic species, acetates, and carboxylates of the acetate type; and (3) the formation of ethyne through the interaction of VC with strong Lewis acid sites, such as tetrahedral Al³⁺ ions at high temperature.

Fig. 13 Reaction pathway for 1,2-DCE dehydrochlorination over γ-alumina.

4. Conclusions

Bimodal mesoporous Al₂O₃ was prepared using polyethyleneglycol (PEG 20 000) and cetyl trimethyl ammonium bromide as a template. Ba/Al₂O₃ catalysts with a Ba loading of 1–10 wt% obtained by incipient wetness were characterized by XRD, BET and porosity measurements, and Py-FTIR, and used in the catalytic dehydrochlorination of 1,2-DCE. The results showed that the Ba species was highly dispersed on Al₂O₃ probably as monomeric or dimeric BaO. The number of surface tetrahedral Al³⁺ ions decreases with the Ba loading, which corresponds to the decrease in strong acid sites. In the catalytic dehydrochlorination of 1,2-DCE, the Ba/Al₂O₃ catalysts present a high activity, of which Al₂O₃ is most active with 95% conversion at 325 °C. The products are composed of VC, aldehyde, ethyne and butene. The formation of aldehyde occurs at low 1,2-DCE conversion (at low temperature) and butane appears at a higher temperature with a low selectivity (below 1%). For Al₂O₃, the selectivity for VC reaches 95% at 350 °C, however, at a higher temperature, the selectivity decreases quickly and is only 40% at 425 °C, which is related to a significant formation of ethyne. The addition of Ba really promotes the selectivity for VC at high temperature through the decrease in strong acidic sites which are favorable for the second dehydrochlorination. Al₂O₃ and 4Ba/Al₂O₃ deactivate heavily on the stream within 20 h at 400 °C, due to the deposition of carbon and chlorine species. The presence of oxygen in a small amount can effectively promote the stability, in which 90% conversion and 90% selectivity for VC on 4Ba/Al₂O₃ is available. In situ FTIR showed that the first step in the interaction of 1,2-DCE with Lewis acid sites occurs and the adsorbed 1,2-DCE can undergo nucleophilic attack by oxygen to form a chlorinated ethoxy intermediate through the removal of a chlorine atom. The formation of VC occurs through transfer of a proton and rearrangement of chlorinated ethoxy. Different types of partially oxidized products, including the enolic form of the aldehyde-type intermediate, were observed because of the attack to VC by hydroxyl or basic oxygen species existing on the surface of the Al₂O₃ or 4Ba/Al₂O₃ catalysts.

Acknowledgements

This research was supported by Development Program for National Natural Science Foundation of China (No. 21277047, 21477036 and 21307033).

References

1. A. S. Shalygin, L. M. Koval, L. V. Malysheva, N. S. Kotsarenko and E. A. Paukshtis, Chem. Sustainable Dev., 2009, 17, 417–422.
2. A. Lakshmanan, W. C. Rooney and L. T. Biegler, Comput. Chem. Eng., 1999, 23, 479–495.
3. I. Mochida, A. Uchino, H. Fujitsu and K. Takeshita, J. Catal., 1978, 51, 72–79.
4. M. M. R. Feijen-Jeurissen, J. J. Jorna, B. E. Nieuwenhuys, G. Sinquin, C. Petit and J. P. Hindermann, Catal. Today, 1999, 54, 65–79.
5. G. Busca, Catal. Today, 2014, 226, 2–13.
6. B. C. Lippens and J. H. de Boer, Acta Crystallogr., 1964, 17, 1312–1321.
7. A. R. Ferreira, M. J. F. Martins, E. Konstantinova, R. B. Capaz, W. F. Souza, S. S. X. Chiaro and A. A. Laitão, J. Solid State Chem., 2011, 184, 1105–1111.
8. R. S. Zhou and R. L. Snyder, Acta Crystallogr., Sect. B: Struct. Sci, 1991, 47, 617–630.
9. J. H. Kwak, J. Z. Hu, D. H. Kim, J. Szanyi and C. H. F. Peden, J. Catal., 2007, 251, 189–194.
10. G. Paglia, A. L. Rohl, C. E. Buckley and J. D. Gale, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 224115–224116.
11. L. Smrcok, V. Langer and J. Krestan, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2006, 62, 83–84.
12. M. Digne, P. Sautet, P. Raybaud, P. Euzen and H. Toulhoat, J. Catal., 2004, 226, 54–68.
13. M. Digne, P. Sautet, P. Raybaud, P. Euzen and H. Toulhoat, J. Catal., 2002, 211, 1–5.
14. T. K. Phung, A. Lagazzo, M. Á. R. Crespo, V. S. Escribano and G. Busca, J. Catal., 2014, 311, 102–113.
15. P. Bai, P. P. Wu and Z. F. Yan, Microporous Mesoporous Mater., 2009, 118, 288–295.
16. H. Kwak, D. Mei, C. W. Yi, D. H. Kim, C. H. F. Peden, L. F. Allard and J. Szanyi, J. Catal., 2009, 261, 17–22.
17. J. H. Kwak, J. Z. Hu, D. Mei, C. W. Yi, D. H. Kim, C. H. F. Peden, L. F. Allard and J. Szanyi, Science, 2009, 325, 1670–1673.
18. H. Knözinger and P. Ratnasamy, Catal. Rev.: Sci. Eng., 1978, 17, 31–70.
19. C. Morterra and G. Magnacca, Catal. Today, 1996, 27, 497–532.
20. X. Liu and R. E. Truitt, J. Am. Chem. Soc., 1997, 119, 9856–9860.
21. D. T. Lundie, A. R. McInroy, R. Marshall, J. M. Winfield, P. Jones, C. C. Dudman, S. F. Parker, C. Mitchell and D. Lennon, J. Phys. Chem. B, 2005, 109, 11592–11601.
22. S. Bordiga, C. Lamberti, F. Bonino, A. Travertd and F. T. Starzyk, Chem. Soc. Rev., 2015, 44, 7262–7341.
23. G. Busca, Chem. Rev., 2010, 110, 2217–2249.
24. A. E. Kulikova and E. N. Zilberman, Russ. Chem. Rev., 1971, 40, 256–271.
25. G. C. Bond and N. Sadeghi, J. Appl. Chem. Biotechnol., 1975, 25, 241–248.
26. P. S. Chintawar and H. L. Greene, Appl. Catal., B, 1997, 13, 81–92.
27. L. Ran, Z. Qin, Z. Y. Wang, X. Y. Wang and Q. G. Dai, Catal. Commun., 2013, 37, 5–8.
28. G. Clet, J. M. Goupil and D. Cornet, Bull. Soc. Chim. Fr., 1997, 134, 223–233.
29. A. G. Goble and P. A. Lawrance, in Proc. 3rd Int. Cong. Catal., North Holland, Amsterdam, 1964, vol. I, p. 320.
30. N. B. Muddada, U. Olsbye, T. Fuglerud, S. Vidotto, A. Marsella, S. Bordiga, D. Gianolio, G. Leofanti and C. Lamberti, J. Catal., 2011, 284, 236–246.
31. G. Leofanti, A. Marsella, B. Cremaschi, M. Garilli, A. Zecchina, G. Spoto, S. Bordiga, P. Fisicaro, G. Berlier, C. Prestipino, G. Casali and C. Lamberti, J. Catal., 2001, 202, 279–295.
32. K. Weissermel and H. J. Arpe, Industrial Organic Chemistry, VCH, Weinheim, Germany, 3rd edn, 1997, p. 219.
33. E. L. Dreher, Ullmann’s Encyclopedia of Industrial Chemistry, VCH, Weinheim, Germany, 1986, vol. A6, p. 289.
34. C. Pistarino, E. Finocchio, M. A. Larrubia, B. Serra, S. Braggio and G. Busca, Ind. Eng. Chem. Res., 2001, 40, 3262–3269.
35. M. W. M. Hisham and S. W. Benson, J. Phys. Chem., 1995, 99, 6194–6198.
36. D. Carmello, E. Finocchio, A. Marsella, B. Cremaschi, G. Leofanti, M. Padovan and G. Busca, J. Catal., 2000, 191, 354–363.
37. B. H. Aristizábal, C. M. de Correa, A. I. Serykh, C. E. Hetrick and M. D. Amiridis, Microporous Mesoporous Mater., 2008, 112, 432–440.
38. H. Vigué, P. Quintard, T. Merle-Méjean and V. Lorenzelli, J. Eur. Ceram. Soc., 1998, 18, 305–309.
39. A. Starczewska, R. Wrzalik, M. Nowak, P. Szperlich, Ł. Bober, J. Szala, D. Stroz and D. Czechowicz, Infrared Phys. Technol., 2008, 51, 307–315.
40. P. S. Chintawar and H. L. Greene, J. Catal., 1997, 165, 12–21.
41. M. Mikami, I. Nakagawa and T. Shimanouchi, Spectrochim. Acta, 1967, 23A, 1037–1053.
42. M. Rodler, C. E. Blom and A. Bauder, J. Am. Chem. Soc., 1984, 106, 4029–4035.
43. T. Rajh, L. X. Chen, K. Lukas, T. Liu, M. C. Thurnauer and D. M. Tiede, J. Phys. Chem. B, 2002, 106, 10543–10552.
44. D. P. Dreoni, D. Pinelli, F. Trifiro, G. Busca and V. Lorenzelli, J. Mol. Catal., 1992, 71, 111–127.
45. Q. Q. Huang, Z. H. Meng and R. X. Zhou, Appl. Catal., B, 2012, 115–116, 179–189.
46. S. Krishnamoorthy, J. A. Rivas and M. D. Amiridis, J. Catal., 2000, 193, 264–272.
47. V. E. Suprunov and A. A. Ivanov, React. Kinet. Catal. Lett., 1987, 33, 75–80.
48. V. S. Escribano, G. Busca and V. Lorenzelli, J. Phys. Chem., 1990, 94, 8939–8945.
49. M. L. Hair, Infrared Spectroscopy in Surface Chemistry, Marcel Dekker, New York, 1967, pp. 205–208.
50. I. Maupin, L. Pinard, J. Mijoin and P. Magnoux, J. Catal., 2012, 291, 104–109.
51. H. Pines and W. O. Haag, J. Am. Chem. Soc., 1960, 82, 2471–2483.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra08855d

---