A chemical titration method for quantification of carbenes in Mo- or W-containing catalysts for metathesis of ethylene with 2-butenes: verification and application potential

Tatiana Otroshchenko *, Ole Reinsdorf , David Linke and Evgenii V. Kondratenko *
Leibniz-Institut für Katalyse e.V. an der Universität Rostock, Albert-Einstein-Str. 29 A, D-18059 Rostock, Germany. E-mail: tatiana.otroshchenko@catalysis.de; evgenii.kondratenko@catalysis.de

Received 23rd August 2019 , Accepted 20th September 2019

First published on 20th September 2019

We introduce and experimentally validate a simple titration method for quantifying the number of carbenes acting as active sites in the metathesis of ethylene with 2-butenes over Mo- or W-containing catalysts. This method is based on the experimentally proven mechanistic sequence of carbene formation through the reaction of ethylene with oxidized metal oxide species: (i) reduction of M(VI) to M(IV) by ethylene, (ii) addition of ethylene to M(IV) to yield M[double bond, length as m-dash]C2H4 carbene and (iii) transformation of the latter to M[double bond, length as m-dash]CH2 upon reaction with C2H4 yielding C3H6. As M[double bond, length as m-dash]CH2 cannot produce C3H6 when reacting with C2H4, the amount of C3H6 formed in step (iii) corresponds to the number of catalytically active carbenes. The applicability of this method was supported by establishing a correlation between this number and the steady-state rate of propene formation in the metathesis of ethylene with 2-butenes.

1. Introduction

Metathesis of ethylene with 2-butenes is one of the large-scale processes for on-purpose production of propene.1,2 Several catalytic systems such as WOx/SiO2, MoOx/Al2O3 and ReOx/Al2O3 are known to be active in such a reaction.3–5 From a mechanistic viewpoint, the reaction occurs on metal-carbenes, which must be formed in situ upon the reaction of feed olefins with supported metal oxides species.6,7 Therefore, the ability of the catalyst to generate the desired metal-carbenes is a key parameter affecting its activity. Another important activity-governing factor is the intrinsic activity of metal-carbenes formed from differently structured (isolated or polymerized) supported metal oxide species. To elucidate these fundamentals, it is necessary to quantify the number of catalytically active metal carbenes. Several methods were described in the literature and used for such purposes.8–11 A brief discussion about their applicability and drawbacks is given below.

Chauvin et al.8 made an attempt to determine the number of active sites at room temperature for ReOx-based catalysts by means of a chemical titration reaction. The catalysts were exposed to a flow of one olefin and then evacuated. Hereafter, so-generated surface intermediates were titrated with another olefin. It is, however, worth mentioning that the amount of active sites determined by such a method is more likely underestimated because some of the intermediates can be desorbed upon catalyst evacuation.

Amakawa et al.9 used ethylene-d4 (CD2[double bond, length as m-dash]CD2) as a probe olefin for the determination of the number of metal-carbene species on the surface of MoOx/SiO2. The procedure included a propene metathesis stage performed at 50 °C followed by a purging stage with Ar. Hereafter, the latter was replaced by a flow of C2D4 and the formed propene-1,1-d2 (CD2[double bond, length as m-dash]CH–CH3) was quantitatively analyzed with a mass spectrometer. It should be mentioned that only one catalyst was tested in such an experiment and therefore no conclusion about the influence of different factors (e.g., the amount and/or the kind of supported MOx species) on the formation of metal-carbenes could be made.

Mougel et al.10 tried to quantify the desired sites on the surface of a WOx-based catalyst by exposing the catalyst subsequently to non-labeled ethylene and 13C-ethylene (13CH213CH2). These two stages of the procedure were separated by a phase of catalyst evacuation in a high vacuum. Gas chromatography coupled with mass spectrometry was applied for monitoring the formation of mono-labeled ethylene during the second stage. Again, only one catalyst was investigated in such an experiment.

Recently, Lwin and Wachs11 demonstrated that titration procedures involving evacuation or flushing steps significantly underestimate the number of metal-carbenes. The authors tried to quantify metal-carbenes in ReOx/Al2O3 through adsorption of 2-butene followed by titration of the species with ethylene. It was shown that the number of surface intermediates is strongly influenced by experimental conditions. Evacuation or flushing was found to decrease the number of detected species by about 90% because of partial desorption of the adsorbed olefins. It was also noted that titration of metal-carbenes at low temperature unlikely provides the correct information about their number, since strongly bound intermediates are incompletely titrated under such conditions.

Against the above background, the purpose of the current work was to elucidate the potential of an alternative chemical titration method for the quantification of metal-carbenes. In contrast to most of the used titration procedures, this method does not include flushing or evacuation stages leading to an underestimation of the number of intermediates, which could desorb before the titration. To verify the method, we applied it for a series of MoOx- or WOx-based catalysts strongly differing in (i) their activity for propene formation in the metathesis of ethylene with 2-butenes and (ii) speciation of the supported metal oxide. From a fundamental viewpoint, the aim was to understand how the kind of support (mixed SiAlOx with different SiO2 contents), the kind of active metal (Mo versus W) and the structure of supported metal oxide species affect the formation and the intrinsic activity of the corresponding metal carbenes. To this end, the titration tests were combined with catalyst characterization and steady-state experiments.

2. Experimental

2.1. Catalyst preparation

(NH4)6Mo7O24·4H2O (Riedel de Haen) and (NH4)6H2W12O40·xH2O (Aldrich) were used as Mo and W precursors for the preparation of Mo- and W-containing catalysts, respectively. Commercial mixed SiAlOx materials with different Si/Al ratios (Siral 10 (10 wt% SiO2 in Al2O3, Sasol), Siral 40 (40 wt% SiO2 in Al2O3, Sasol) and Siral 70 (70 wt% SiO2 in Al2O3, Sasol)) were applied as supports. The catalysts were synthesized by impregnation of 5 g of each pre-calcined (500 °C, 8 hours in static air) support with a required amount of the aqueous solution of the metal precursor to obtain the catalysts with a nominal metal surface density of 0.15, 1.5, 3 and 5 nm−2. Hereafter, the catalysts were dried at 110 °C overnight and finally calcined in static air at 500 °C for 8 hours. The catalysts were abbreviated as XM/Y with “X” standing for the nominal surface density of Mo or W and “Y” representing the support. The amounts of the metal precursor and water required for catalyst synthesis were calculated taking into account the specific surface areas of the bare supports and their absorptive capacity (see Table S1 in the ESI).

2.2. Catalyst characterization

Nitrogen physisorption experiments were carried out at 77 K to determine the specific surface areas (SBET) of the samples using a Belsorp mini II setup (Bel Japan). Desorption isotherms were evaluated according to the BET method.

UV-vis measurements were carried out using an Avantes spectrometer (AvaSpec-2048-USB2-RM) equipped with a high-temperature reflection UV-vis probe, an Ava-Light-DH-S-BAL deuterium–halogen light source and a CCD array detector. The probe consisting of six radiating optical fibers and one reading fiber was threaded through the furnace to face the wall of the quartz tube reactor at the position where the catalyst (200–300 mg) was located. Before recording the UV-vis spectra, the sample was in situ calcined in flowing air at 500 °C for 1 h and cooled down to room temperature. The UV-vis spectra were recorded at room temperature in the range from 200 to 800 nm. Barium sulfate (99.998%, Aldrich) was used as a white standard.

X-ray diffraction (XRD) measurements were performed on a Theta/Theta diffractometer X'Pert Pro (Panalytical) with CuKα radiation (λ = 1.5418 Å, 40 kV, 40 mA) and an X'Celerator RTMS detector. The phase composition of the samples was determined using the program suite WinXPOW (Stoe & Cie) with inclusion of the powder diffraction file PDF2 of the International Centre for Diffraction Data.

Temperature-programmed reduction (TPR) experiments were carried out in an in-house developed set-up containing eight individually heated continuous-flow fixed-bed quartz reactors. Prior to the experiments, all the samples (80–200 mg) were calcined in flowing air at 500 °C for 1 h, cooled down to room temperature in the same flow and then flushed with Ar for 15 min. Hereafter, the samples were heated in a flow of 5 vol% H2 in Ar (10 ml min−1) up to 900 °C with a heating rate of 10 K min−1. The consumption of H2 was determined from the obtained H2 profiles recorded with the help of an on-line mass spectrometer (Pfeiffer Vacuum OmniStar GSD 320). Signals at atomic mass units (AMUs) of 2 (H2) and 40 (Ar) were collected. Argon was the internal standard.

Temperature-programmed desorption of ammonia (NH3-TPD) was carried out in the same set-up used for TPR experiments. The samples (200 mg) were exposed to a flow of 1 vol% NH3 in Ar (10 ml min−1) at 120 °C for 1 h. Then the catalysts were flushed with Ar for 5 h to remove weakly bound NH3 and cooled down to 80 °C. Hereafter, the catalysts were heated in Ar flow up to 900 °C at a heating rate of 10 K min−1. Desorbed ammonia was registered by an on-line mass spectrometer (Pfeiffer Vacuum OmniStar GSD 320). Signals at AMUs of 15 (NH) and 40 (Ar) were used. Prior to the experiment, all the samples were in situ calcined in flowing air at 500 °C for 1 h.

2.3. Catalyst testing

Catalytic tests were performed at 1.25 bar (abs.) in an in-house developed set-up equipped with 14 continuous-flow fixed-bed reactors. The catalysts (20 mg for Mo-containing and 150 mg for W-containing catalysts) were put into the reactors, heated in a flow of N2 up to 500 °C and then calcined at the same temperature in air flow for 3 hours. Hereafter, they were cooled down in N2 flow to 50 °C and exposed to a flow of C2H4/trans-2-C4H8/N2 in a molar ratio of 5/5/1 (total flow was set to 22 mL min−1). The initial rate of propene formation (r(C3H6)) was measured after 420 s on stream. Eqn (1) was used to calculate this rate.
image file: c9cy01697j-t1.tif(1)

Here, Ffeed is the volumetric flow rate of the feed gas (mL min−1) under reference conditions (0 °C, 1 atm), N2 with superscripts “in” or “out” stand for the molar flow (mol mL−1) of N2 at the reactor inlet or outlet, xout (C3H6) is the molar fraction of propene at the reactor outlet, Vm is the molar volume (22[thin space (1/6-em)]414 mL mol−1), and mcat is the weight of the catalyst (g). The feed components and the reaction products were analysed using an on-line gas chromatograph (Agilent 6890) equipped with an AL/S capillary column (for hydrocarbons), connected to a flame ionization detector and a PLOT/Q (for CO2)/Molsieve 5 (for H2, O2, N2, and CO) capillary column combination connected to a thermal conductivity detector.

2.4. Titration of carbene species

Titration experiments were performed in the same set-up as the catalytic tests. The catalysts (100 or 150 mg) were loaded into quartz reactors and heated in N2 flow up to 500 °C. Hereafter, they were calcined in air at the same temperature for 3 hours. Then the catalysts were cooled down in N2 flow to 50 °C and exposed to a flow of C2H4/N2 in a molar ratio of 1/1 (total flow was set to 25 mL min−1). Formed C3H6 was quantified using the above-mentioned gas chromatograph. The experiment was carried out until C3H6 was no longer detected. In cases when the molar fraction of C3H6 registered at the end of the experiment was not equal to zero, the temporal profile of the C3H6 molar fraction was linearly extrapolated to zero. The amount of C3H6 calculated through integration of the temporal profile of the C3H6 molar fraction was considered to be equal to the amount of metal carbenes formed.

3. Results and discussion

3.1. Catalysts and their characterization

All the catalysts were characterized by BET, XRD, TPR and NH3-TPD methods. Selected properties of the samples are listed in Table 1. Analysing these data, one can see that the specific surface area gradually decreases with increasing metal loading, indicating the blockage of pores of support materials by MOx (M = Mo, W). Thus, the surface area of Siral 10 decreased by 34% after depositing Mo to obtain 5Mo/Siral 10. In general, W-containing catalysts possess a lower surface area than the corresponding Mo-containing catalysts.
Table 1 Content of Mo or W in wt% assuming that metal exists in a form of MO3 in a fresh catalyst, specific surface area (SBET), temperature of the maximum H2 consumption (Tmax-H2) obtained from the TPR experiment, temperature of the maximum NH3 desorption (Tmax-NH3) and density of acidic sites (N(ac.s.)) calculated from the NH3-TPD experiment
Catalyst ω (M), wt% S BET, m2 g−1 T max-H2, °C T max-NH3, °C N(ac.s.), nm−2
0.15Mo/Siral 10 0.8 340 >900 290, 652 0.34
0.15Mo/Siral 40 1.1 390 893 297, 693 0.32
0.15Mo/Siral 70 0.8 320 807 350, 685 0.31
1.5Mo/Siral 10 7.4 315 511, >900 282 0.44
1.5Mo/Siral 40 9.7 365 571, 858 272 0.41
1.5Mo/Siral 70 7.2 275 629 280 0.42
3Mo/Siral 10 13.4 255 504, 622, 854 277 0.58
3Mo/Siral 40 16.9 221 544 253 0.46
3Mo/Siral 70 13.0 213 511, 767, 824 277 0.49
5Mo/Siral 10 19.7 232 507, 589, >900 277 0.68
5Mo/Siral 40 24.2 163 544, 595 253 1.0
5Mo/Siral 70 19.2 157 612, >900 259 1.3
0.15W/Siral 10 1.6 295 286 0.42
0.15W/Siral 40 2.1 335 305 0.4
0.15W/Siral 70 1.5 290 359 0.39
1.5W/Siral 10 13.3 270 >900 270 0.48
1.5W/Siral 40 17.1 315 804, >900 293 0.45
1.5W/Siral 70 12.9 285 585, >900 316 0.50
3W/Siral 10 22.8 221 574, >900 277 0.69
3W/Siral 40 28.1 237 568, >900 277 0.53
3W/Siral 70 22.2 169 565, >900 298 0.60
5W/Siral 10 31.9 178 566, 884 273 0.71
5W/Siral 40 37.9 170 565, >900 277 0.61
5W/Siral 70 31.2 169 569, >900 273 0.62

In order to investigate the speciation of MOx in the catalysts with low metal atomic surface density (0.15 and 1.5 nm−2), we applied UV-vis spectroscopy. The UV-vis spectra of Mo- and W-containing catalysts are shown in Fig. S1(a) and (b) in the ESI, respectively. Fig. S2 in the ESI shows the edge energy (Eg) values obtained from the spectra according to ref. 12. The spectra of the catalysts with 0.15 nm−2 Mo (Fig. S1(a) ) are characterized by strong absorption bands at around 250 and 290 nm, which can be ascribed to the ligand-to-metal charge transfer (LMCT) transitions in highly dispersed and small linear-chained tetrahedral species, respectively.13 Increasing the Mo atomic density up to 1.5 nm−2 results in a shift of the absorption peak towards higher wavelengths due to the appearance of a new band at around 330 nm related to the LTMCT in polymerized octahedral MoOx species.14

The presence of a weak absorption band at around 400 nm can be assigned to the appearance of small MoO3 crystallites.15 The UV-vis spectra of 0.15W/Y catalysts (Fig. S1(b)) are characterized by a band at around 250 nm which corresponds to LMCT of isolated WO4 species.12,15 Increasing the W atomic surface density to 1.5 nm−2 gave rise to a new band located at around 280 nm, which can be assigned to octahedral polytungstate species.16 The presence of a weak band at around 400 nm in the UV-vis spectra of 1.5W/Siral 40 and 1.5W/Siral 70 can be related to bulk WO3.17 It should be noted that for both Mo- and W-containing catalysts the Eg value decreases with increasing surface MOx coverage (Fig. S2), indicating an increase in MOx domain size.18–20 Moreover, for the catalysts with the same metal atomic surface density, the Eg value decreases with increasing fraction of SiO2 in the support. Such a result can be explained by the weaker interaction of MOx species with silica-rich than with alumina-rich supports resulting in agglomeration of MOx species.

The XRD method was used to investigate the phase composition of the catalysts with atomic surface density of 1.5 and higher. The XRD patterns of the catalysts containing 1.5, 3 or 5 Mo per 1 nm2 on the surface of Siral 10, Siral 40 or Siral 70 are shown in Fig. S3(a) in the ESI. For 1.5Mo/Siral 10, no reflections related to crystalline Mo-containing phases could be detected due to the strong interaction of MoOx with an alumina-rich support, enabling high dispersion of MoOx.13 For all other samples with lower Al2O3 loading, the reflections related to orthorhombic MoO3 and/or monoclinic Al2(MoO4)3 phases21 were identified. The weaker interaction of MoOx with a silica-rich support induces the formation of crystalline MoO3. The higher the Mo loading in these materials, the higher the fraction of the Mo-containing phases.

The corresponding XRD patterns of W-based catalysts with a W surface density of 1.5, 3 or 5 nm−2 are shown in Fig. S3(b) in the ESI. Similar to their Mo-based counterparts, no reflections related to crystalline W-containing phase(s) could be seen in 1.5W/Siral 10. The XRD patterns of the other samples possess reflections of monoclinic WO3.22,23 Their intensity increases with W loading, thus indicating that this phase becomes more evident. Moreover, with respect to the kind of support, the following order of the fraction of WO3 was established: Siral 10 < Siral 40 < Siral 70. This order can be explained by the weaker interaction of WOx species with silica-rich than with alumina-rich supports.24 Siral 10 with only 10 wt% SiO2 favors strong interaction with WOx species, thus enabling their high dispersion and preventing the formation of a crystalline phase. With increasing SiO2 content in the support, the interaction between the WOx species and support decreases resulting in agglomeration of WOx and formation of WO3.

As the redox properties of supported MoOx and WOx species are decisive for carbene formation, we applied TPR for analysing these catalyst properties. The obtained TPR profiles in a form of the ratio of H2 to Ar as a function of temperature are shown in Fig. S4(a) in the ESI for Mo-containing catalysts. The corresponding values of the temperature (Tmax-H2) of the maximum hydrogen consumption are shown in Table 1. Hydrogen uptake increases with rising metal loading. In addition, the shape of the H2-TPR profiles and the Tmax-H2 values are influenced by Mo loading and the kind of support. For most catalysts, the profiles are characterized by two or three maxima which are due to the presence of differently reducible MoOx species (octahedral and tetrahedral species, crystalline MoO3).21,25,26 With increasing Mo surface density from 0.15 to 3 nm−2, the Tmax-H2 shifts to lower temperatures. This is due to the decrease in strength of the interaction between Mo and the support as described elsewhere.21 It is also well-known that at low Mo loading, MoOx species mostly form tetrahedral structures which are difficult to reduce due to the strong interaction with the support. At high Mo loadings, octahedral MoOx species dominate. Such species are easily reducible. For Mo/Siral 70 samples, the first reduction peak shifts to higher temperature when the Mo loading increases further up to 5 nm−2. Such behavior can be explained by the presence of MoO3 crystallites.26

Fig. S4(b) in the ESI shows the H2-TPR profiles of W-containing samples. The corresponding values of Tmax-H2 are listed in Table 1. Regardless of the kind of support, no hydrogen consumption was observed for the 0.15W/Y samples. 1.5W/Siral 10 consumed only a small amount of hydrogen at temperatures higher than 800 °C. For the other samples, the H2-TPR profiles are characterized by hydrogen consumption in the temperature range from 400 to 650 °C and above 670 °C. As in the case of Mo-containing catalysts, the hydrogen uptake increases with increasing W loading. The presence of several reduction peaks is in good agreement with previous work on WOx-based materials.27 Low-temperature reduction is related to WOx species in octahedral coordination, while a broad peak at higher temperatures is associated with the reduction of tetrahedrally coordinated species. It should be noted that with respect to the effect of the kind of support on WOx reducibility, for most of the catalysts the following order of the low-temperature Tmax-H2 was established: Siral 10 > Siral 40 > Siral 70. Taking these data, the XRD data and previous work of Debecker et al.24 into account, it can be concluded that weakly bound WOx species on a silica-rich support (Siral 70) can be reduced much easier than highly dispersed strongly bound WOx species on an alumina-rich support (Siral 10).

To check if acidic properties are relevant for the transformation of metal oxide species into carbenes (see section 3.3), we determined the overall catalyst acidity by means of NH3-TPD tests. Fig. S5 in the ESI shows the obtained NH3-TPD profiles defined as the ratio of MS signals at AMUs of 15 (NH) and 40 (Ar). All the profiles are characterized by a broad asymmetric NH3 desorption peak in the temperature range from 180 to 600 °C. For the catalysts with low metal loading (0.15Mo/Siral 10, 0.15Mo/Siral 40, 0.15Mo/Siral 70, 0.15W/Siral 10, 0.15W/Siral 40, and 0.15W/Siral 70), an additional small broad high-temperature peak between 580 and 700 °C is observed. This means that these samples additionally possess strong acidic sites. The position of the low-temperature desorption peak slightly shifts to lower temperatures, while the high-temperature desorption peak disappears when the metal loading increases.

This observation is in agreement with a previous study26 and suggests a decrease in the strength of acidic sites with rising metal loading. Table 1 shows the temperature of the maximum NH3 desorption and the number of acidic sites per 1 nm2 of catalyst calculated under the assumption that one NH3 molecule is adsorbed by one acidic site. Regardless of the kind of support, the amount of NH3 desorbed increases with increasing metal atomic surface density for both Mo- and W-containing catalysts, indicating an increase in the number of acidic sites. This result is in good agreement with a previous study.13 In most cases, the catalysts containing Siral 10 as a support demonstrated higher density of acidic sites than the catalysts containing Siral 40 or Siral 70.

3.2. Catalytic test

All the catalysts were tested for their activity in the metathesis of ethylene with trans-2-butene to propene at 50 °C. Before the reaction, all the catalysts were in situ calcined at 500 °C for 3 hours and then cooled down to the reaction temperature in a flow of nitrogen. The rate of propene formation over Mo-containing catalysts as a function of Mo surface density is shown in Fig. 1(a). It is clearly seen that for XMo/Siral 10 catalysts the activity passes through a maximum at a Mo surface density of about 3 nm−2, while for XMo/Siral 40 and XMo/Siral 70 this maximum is shifted to 1.5 nm−2. The dependence of the rate of propene formation on W surface density for W-containing catalysts is shown in Fig. 1(b). It should be noted that the rate of propene formation over W-based catalysts is much lower than that over the corresponding Mo-based catalysts (up to 100 times). The activity of XW/Siral 10 samples gradually increases with increasing W loading, while the activity of XW/Siral 40 and XW/Siral 70 passes through a maximum at a W surface density of 3 and 1.5 nm−2, respectively. For benchmarking our catalysts, we compare the metathesis activity of some of our catalysts with that of state-of-the art Mo- or W-containing catalysts tested under similar conditions. As seen in Table S2 in the ESI, our catalysts demonstrated much higher activity (up to 4 times higher).
image file: c9cy01697j-f1.tif
Fig. 1 The rate of propene formation at 50 °C as a function of (a) Mo atomic surface density and (b) W atomic surface density.

It should be mentioned that no direct correlation between the catalytic activity and the redox or acidic properties of the samples could be established (see Fig. S6(a)–(d) in the ESI). Thus, the main activity-determining factors should be (i) the ability of MOx (M = Mo, W) species to be in situ transformed into metal carbenes and/or (ii) the intrinsic activity of the latter formed from differently structured supported MOx. To check if and how these two catalyst properties depend on metal loading and the kind of support, we tried to quantify the number of metal-carbenes and relate the obtained data with catalyst activity. The results are presented and discussed in the next section.

3.3. Titration of carbene species

In order to obtain a quantitative estimation of the number of carbenes formed from MOx (M = Mo, W) species, we performed titration experiments as follows. The catalysts were initially treated in the same way as before steady-state metathesis tests (see section 2.4 for details). Hereafter, they were exposed at 50 °C to a C2H4/N2 = 1 feed. The idea behind such experiments is that active carbene species are formed through the reaction of ethylene with MOx as schematically shown in Fig. 2 and experimentally validated in our previous study for two Mo-containing samples.28 According to this scheme, an M(VI) species is initially reduced by ethylene to M(IV). In a second step, the reduced metal oxide reacts with another ethylene molecule to form carbene species 3 in Fig. 2. This carbene reacts further with ethylene, producing a metallacyclobutane intermediate (4) which decomposes into gas-phase propene and a new carbene species (5 in Fig. 2). It should be noted that propene can be formed only from carbene species 3, while carbene species 5 when reacting with ethylene will again give ethylene but not propene. As a consequence, propene formation will stop after all carbene species 3 are transformed into carbene species 5. On the basis of this mechanistic concept, the amount of propene formed upon catalyst treatment with ethylene corresponds to the number of MOx species being able to be transformed into carbene species (3 and 5 in Fig. 2).
image file: c9cy01697j-f2.tif
Fig. 2 A scheme of the formation of M-carbenes during the reaction of MOx species with ethylene.

Temporal changes of the molar fraction of registered propene related to the catalyst amount are shown in Fig. S7(a) and (b) in the ESI for Mo- and W-containing catalysts, respectively. As expected, the highest amount of propene was detected at the beginning of the experiment where carbene species 3 are formed and react with ethylene to propene and carbene species 5. After passing a maximum, the concentration of propene decreased as carbene species 5 cannot form propene from ethylene. Besides propene, trace amounts of C4–C5 olefins were also observed. Such products are suggested to be produced through ethylene dimerization and consecutive metathesis. Since the amount of such products was negligibly small, we did not consider them when counting carbene sites.

Although 2-C4 olefins can form propene through the metathesis reaction with ethylene, this pathway should play a minor role in comparison with that shown in Fig. 2 due to the very low concentration of the higher olefins. The amount of propene formed during these titration experiments was determined by integrating the area under the curve from Fig. S7 (in the ESI) and using eqn (2).

image file: c9cy01697j-t2.tif(2)

The obtained results are shown in Fig. 3 as a function of Mo or W surface density. One can see that the dependence of the amount of propene formed in the titration tests on the surface density is very similar to that derived for the rate of propene formation in the ethylene–2-butene metathesis (Fig. 1). For illustrating the relationship between the steady-state activity and the amount of propene determined from titration tests, we constructed Fig. 4 for all the catalysts tested in the present study. In general, the rate of propene formation increases, independently from the kind of support, with increasing amount of carbene species (which is considered to be equal to the amount of propene formed during titration tests). Such a correlation suggests that our titration method can be used for the quantification of the metathesis active carbenes formed from supported MoOx or WOx species. It should, however, be mentioned that the method provides an overall number of carbene sites but cannot distinguish between carbenes differing in their intrinsic activity for propene formation under steady-state conditions.

image file: c9cy01697j-f3.tif
Fig. 3 Dependence of the amount of C3H6 formed during the titration experiment at 50 °C on metal atomic surface density for (a) Mo-containing or (b) W-containing catalysts. Note: The titration experiment was repeated for selected samples; the difference in C3H6 amount for two independent tests did not exceed 3%.

image file: c9cy01697j-f4.tif
Fig. 4 Dependence of the rate of propene formation at 50 °C on the amount of propene formed during the titration experiment at 50 °C for (a) Mo-containing or (b) W-containing catalysts.

3.4. Factors affecting formation and intrinsic activity of carbenes

Taking into account the correlation in Fig. 4, we discuss now the fundamentals of different activities of Mo- and W-containing catalysts. In order to simplify the comparison, we selected only the catalysts mostly containing isolated tetrahedral species (catalysts 0.15 M/Y). Hereby, we can exclude the difference in activity due to the presence of different species (tetragonal, octahedral, and non-active crystalline MO3). To estimate the fraction of MOx species converted into carbenes in the catalysts 0.15 M/Y, we related the number of propene molecules formed in our titration tests to the total amount of Mo or W in the catalysts. The obtained data are shown in Fig. 5(a) (the data for 0.15 M/Siral 10 are not presented due to a big error associated with the detection of a tiny amount of carbenes). Regardless of the kind of support, the fraction of MoOx species being able to form carbenes was found to be much higher (up to 100 times) than that of the WOx species. As seen in Fig. 1, W-containing catalysts in comparison with Mo-containing catalysts revealed a lower rate of propene formation in the metathesis of ethylene with 2-butenes. It is also obvious that a much lower amount of carbenes (expressed as the amount of propene in Fig. 3) was generated on the surface of the former catalysts. Thus, one can suggest that the lower propene formation rate over W-containing catalysts is related to the lower ability of WOx to transform to carbenes.
image file: c9cy01697j-f5.tif
Fig. 5 (a) Fraction of active metal determined for 0.15 M/Siral 40 and 0.15 M/Siral 70; (b) turnover frequency of propene formation for 0.15 M/Siral 40 and 0.15 M/Siral 70.

To check how the kind of MOx influences the intrinsic activity of carbenes, we calculated the apparent turnover frequency (TOF) of propene formation as follows. The steady-state rate of propene formation was related to the amount of carbene species determined from our titration tests (Fig. 3). The TOF values determined for Mo- and W-containing catalysts with an apparent surface density of metal of 0.15 nm−2 are shown in Fig. 5(b). It is worth mentioning that the TOF values do not significantly differ for different samples. Based on such observation, we can conclude that carbenes formed from MoOx or WOx species possess similar intrinsic activity. Thus, the different activities of W- and Mo-containing samples are mostly related to the different abilities of WOx and MoOx species to form carbenes.

4. Conclusions

It was demonstrated that the amount of carbenes formed from supported MoOx or WOx species is strongly influenced by the kind of support, metal loading and its chemical nature. In general, Mo-containing catalysts are able to generate a higher amount of carbenes than the corresponding W-containing ones. Depending on the kind of support, between 20 and 30% highly dispersed MoOx species are transformed into Mo-carbenes, while this number is about 100 times lower for the W-based counterparts. Using these quantitative data and the results of steady-state tests on ethylene–2-butene metathesis, the intrinsic activity of Mo-carbenes for propene formation was demonstrated to be similar to that of carbenes formed from corresponding WOx species. Thus, the higher activity of Mo-containing catalysts for propene formation in comparison with their W-containing counterparts is related to the higher ability of the former catalysts to in situ form active carbenes.

Conflicts of interest

There are no conflicts to declare.


The authors thank Dr. Henrik Lund for the XRD analysis. Financial support by Deutsche Forschungsgemeinschaft (DFG, project OT 586/1-1) is gratefully acknowledged.


  1. S. Maksasithorn, D. P. Debecker, P. Praserthdam, J. Panpranot, K. Suriye and S. K. N. Ayudhya, Chin. J. Catal., 2014, 35, 232–241 CrossRef CAS.
  2. J. C. Mol, J. Mol. Catal. A: Chem., 2004, 213, 39–45 CrossRef CAS.
  3. H. A. Wittcoff, B. G. Reuben and J. S. Plotkin, in Industrial Organic Chemicals, John Wiley & Sons, Inc., 2005, pp. 57–99 Search PubMed.
  4. N. Popoff, E. Mazoyer, J. Pelletier, R. M. Gauvin and M. Taoufik, Chem. Soc. Rev., 2013, 42, 9035–9054 RSC.
  5. W. Jiang, R. Huang, P. Li, S. Feng, G. Zhou, C. Yu, H. Zhou, C. Xu and Q. Xu, Appl. Catal., A, 2016, 517, 227–235 CrossRef CAS.
  6. Y. Chauvin, Angew. Chem., Int. Ed., 2006, 45, 3740–3747 CrossRef PubMed.
  7. X. Li, J. Guan, A. Zheng, D. Zhou, X. Han, W. Zhang and X. Bao, J. Mol. Catal. A: Chem., 2010, 330, 99–106 CrossRef CAS.
  8. Y. Chauvin and D. Commereuc, J. Chem. Soc., Chem. Commun., 1992, 462–464 RSC.
  9. K. Amakawa, S. Wrabetz, J. Kröhnert, G. Tzolova-Müller, R. Schlögl and A. Trunschke, J. Am. Chem. Soc., 2012, 134, 11462–11473 CrossRef CAS PubMed.
  10. V. Mougel, K.-W. Chan, G. Siddiqi, K. Kawakita, H. Nagae, H. Tsurugi, K. Mashima, O. Safonova and C. Copéret, ACS Cent. Sci., 2016, 2, 569–576 CrossRef CAS PubMed.
  11. S. Lwin and I. E. Wachs, ACS Catal., 2015, 5, 6823–6827 CrossRef CAS.
  12. E. I. Ross-Medgaarden and I. E. Wachs, J. Phys. Chem. C, 2007, 111, 15089–15099 CrossRef CAS.
  13. T. Hahn, U. Bentrup, M. Armbrüster, E. V. Kondratenko and D. Linke, ChemCatChem, 2014, 6, 1664–1672 CrossRef CAS.
  14. C. C. Williams, J. G. Ekerdt, J. M. Jehng, F. D. Hardcastle, A. M. Turek and I. E. Wachs, J. Phys. Chem., 1991, 95, 8781–8791 CrossRef CAS.
  15. E. L. Lee and I. E. Wachs, J. Phys. Chem. C, 2007, 111, 14410–14425 CrossRef CAS.
  16. S. Huang, F. Chen, S. Liu, Q. Zhu, X. Zhu, W. Xin, Z. Feng, C. Li, Q. Wang and L. Xu, J. Mol. Catal. A: Chem., 2007, 267, 224–233 CrossRef CAS.
  17. H. Liu, S. Huang, L. Zhang, S. Liu, W. Xin and L. Xu, Catal. Commun., 2009, 10, 544–548 CrossRef CAS.
  18. J. E. Herrera, J. H. Kwak, J. Z. Hu, Y. Wang, C. H. F. Peden, J. Macht and E. Iglesia, J. Catal., 2006, 239, 200–211 CrossRef CAS.
  19. A. Chakrabarti and I. E. Wachs, ACS Catal., 2018, 8, 949–959 CrossRef CAS.
  20. I. E. Wachs, T. Kim and E. I. Ross, Catal. Today, 2006, 116, 162–168 CrossRef CAS.
  21. S. Rajagopal, H. J. Marini, J. A. Marzari and R. Miranda, J. Catal., 1994, 147, 417–428 CrossRef CAS.
  22. Y. Wang, Q. Chen, W. Yang, Z. Xie, W. Xu and D. Huang, Appl. Catal., A, 2003, 250, 25–37 CrossRef CAS.
  23. X. Xia, R. Jin, Y. He, J.-F. Deng and H. Li, Appl. Surf. Sci., 2000, 165, 255–259 CrossRef CAS.
  24. D. P. Debecker, M. Stoyanova, U. Rodemerck, F. Colbeau-Justin, C. Boissère, A. Chaumonnot, A. Bonduelle and C. Sanchez, Appl. Catal., A, 2014, 470, 458–466 CrossRef CAS.
  25. J. Brito and J. Laine, Polyhedron, 1986, 5, 179–182 CrossRef CAS.
  26. K. V. R. Chary, K. R. Reddy, G. Kishan, J. W. Niemantsverdriet and G. Mestl, J. Catal., 2004, 226, 283–291 CrossRef CAS.
  27. F. Di Gregorio and V. Keller, J. Catal., 2004, 225, 45–55 CrossRef CAS.
  28. V. A. Kondratenko, T. Hahn, U. Bentrup, D. Linke and E. V. Kondratenko, J. Catal., 2018, 360, 135–144 CrossRef CAS.


Electronic supplementary information (ESI) available: Databases containing characterization and catalytic data. See DOI: 10.1039/c9cy01697j

This journal is © The Royal Society of Chemistry 2019