Revisiting Oxidative Dehydrogenation of Ethane by W Doping into MoVMn Mixed Oxides at Low Temperature

The catalytic performance ofMoVMnWmixed oxides was investigated in the oxidative dehydrogenation of ethane at three different reaction temperatures (235, 255, and 275C) using oxygen as an oxidant.The catalysts were characterized by using X-ray diffraction, temperature-programmed reduction, and scanning electron microscopy.TheMoVMnWmixed oxide catalyst showed the 70–90% of ethylene selectivity at the reaction temperatures. However, a significant decrease in the selectivity of ethylene was observed by increasing the reaction temperature from 235C to 275C.


Introduction
Ethylene (C 2 H 4 ) is an important building block used in the petrochemical industry [1].Its primary production method is the energy-intensive steam cracking process [2] which is usually carried out at  > 800 ∘ C [3].Over 100 million metric tons of ethylene is produced annually by means of the steam cracking of long carbon-chain hydrocarbons.In addition to the high energy cost, unavoidable side reactions and the deactivation of catalysts by carbon deposition are other technical shortcomings of the process.Increasing attention has been directed towards the oxidative dehydrogenation (ODH) of alkanes because of its potential benefit of using abundant and inexpensive raw materials [4][5][6].The ODH of light alkanes offers a potentially attractive route to alkenes since the reaction is exothermic [2] and avoids the thermodynamic constraints of nonoxidative routes by forming by-product water [7].However, a low yield and selectivity of olefins still retard the industrial application of the ODH.
To date, about a 20% yield of ethylene was reported in the literature, and for industrial applications, it requires the ethylene productivity of above 1 kg C 2 H 4 /kg cat −1 h −1 [8].The ODH of ethane to ethylene has received less attention than the nonoxidative cracking [9].However, ethane is the secondlargest component of natural gas and the primary product of the methane conversion by oxidative coupling.Thus, a large number of catalysts have been investigated for the ODH of ethane and propane, for example, using Mo-V based catalysts [10][11][12][13].It was reported that a V based catalyst is one of the most active and selective single metal catalysts in the ODH [14,15], while V [10,[16][17][18], Mo [12,13,[19][20][21], Pt [22][23][24][25] based catalysts, and MoV mixed oxides [26][27][28] are the most widely studied transition-metal catalysts for the ODH of ethane.Furthermore, a more complicated material (i.e., Mo 1 V 0.25 Nb 0.12 Pd 0.0005 O x [29][30][31]) with two different catalytic centers, such as the ODH of ethane and the heterogeneous Wacker oxidation of ethylene to acetic acid, was reported.The catalytic activity of MoV based catalysts supported on TiO 2 and Al 2 O 3 was investigated in the ODH of both ethane and propane [32].Recently, the catalytic properties of the acidic and basic forms of Ni-, Cu-, and Feloaded Y zeolites were investigated in the ODH of ethane to ethylene [33].As discussed above, although numerous studies on the ODH of ethane using MoV based mixed oxides were reported [29,31,32,34,35], to the best of our knowledge, little studies using W doping into MoV mixed oxide catalysts have been reported in the literature [35,36], demonstrating the ethane conversion is the best using a 6.  40-60 mesh particles.Reactants and products were analyzed using an online gas chromatograph equipped with a double detector (i.e., thermal conductivity detector (TCD) and flame ionization detector (FID)) and two columns of HayeSep D 80/100 mesh and LAC446.We confirmed the reproducibility of the experiment results within the uncertainty of ±2%.
Then, we obtained the conversion ( ethane ), selectivity to product  (  ), and yield to product  (  ). ) were noticed as the major phases (Figure 1).TPR in H 2 was performed to examine reduction characteristics of the mixed oxide catalysts as displayed in Figure 2.There are two reduction peaks observed for W 0 at 587 ∘ C and 677 ∘ C which correspond to the reduction of isolated V 5+ species in the bulk structure of the catalyst [38,39] and Mo 6+ species [40], respectively.In addition, the observed shoulder at a temperature around 470 ∘ C is assumed to be H 2 uptake associated with the reduction of MnO 2 to Mn 2 O 3 [41].The position of the peak at 677 ∘ C is shifted to higher temperatures of 769 ∘ C (W 0.06 ) and 772 ∘ C (W 0.14 ) and becomes broadened without a significant change in peak intensities.Solsona and coworkers [42] reported a similar shift when W was added to NiO and proposed that the shift was related to an interaction between NiO and tungsten oxide nanoparticles.As shown in the peaks of W 0.06 , the H 2 consumption at 470 ∘ C and 587 ∘ C was enhanced by the addition of W loading (W 0 versus W 0.06 ).We observed that the reduction occurred slightly at higher temperatures by adding more W into the MoVMn based catalysts (W 0.06 versus W 0.14 ).

Results and Discussion
Figure 3 shows typical SEM images of W 0 , W 0.06 , and W 0.14 doped catalysts, displaying that all samples have irregular shaped particles.The SEM image of W 0 without W ascertains a rough surface with variable shapes and sizes, and a few regions show a stack of fine crystallites (Figure 3(a)).Similarly, W 0.06 has a rough surface with variable shapes and sizes, but without a stack of fine crystallites (Figure 3(b)).The SEM images of W 0.14 determine a stack of fine crystallites, leading to a flower-like morphology (Figure 3(c)).

Catalytic
Performance from the ODH of Ethane.Figure 4 shows the effect of W loading on the ethane conversion and product selectivity from the ODH of ethane at 275 ∘ C and at a reaction time of 0.65 sec.Ethylene is observed as the major product from the ODH of ethane, while acetic acid, CO, and CO 2 are also detected.The highest ethane conversion (∼21.0%) is observed from the MoVMn sample with W = 0.063.The increase in the amount of W over the MoVMn sample leads to a reduction in the conversion of ethane.Similarly, the addition of W results in an increase in the selectivity of ethylene with a maximum selectivity (∼ 74%) achieved at W = 0.063.The lowest selectivity of acetic acid (∼12%) is detected at W = 0.14.The ODH experiment of ethane was carried out at two more temperatures of 235 and 255 ∘ C using the W 0.06 catalyst with the contact times of 0.08, 0.12, 0.16, 0.24, 0.32, 0.49, and 0.65 sec (Figure 5).As shown in Figure 5, ethane conversions of approximately 7.0, 11.3, and 21.0% are achieved at 235, 255, and 275 ∘ C, respectively, at a contact time of 0.65 sec.The conversion of ethane is improved as the contact time increases.In this study, the maximum ethane conversion was obtained at 275 ∘ C. The product distribution from the ODH of ethane on the W 0.06 catalyst at 235, 255, and 275 ∘ C is presented in Figure 6.At all temperatures studied, ethylene is the major product, while acetic acid and carbon oxides are also observed (i.e., ∼3.5% and ∼1.7%, resp., at 275 ∘ C). Figure 7 presents the influence of reaction temperatures and contact times for the selectivities of ethylene, acetic acid, and carbon oxides (CO and CO 2 ).The selectivities of ethylene decrease by the increase of contact times at all temperatures studied.The highest selectivity of ethylene is observed at a reaction temperature of 235 ∘ C. Increasing the reaction temperature from 235 ∘ C to 275 ∘ C leads to a reduction in the selectivity of ethylene, while an enhancement of selectivities of acetic acid, CO, and CO 2 are observed (Figure 7(b) and Table 1).Table 1 compiles theoretical selectivities of ethylene and CO x at each reaction temperature.The trend of the reaction products over the W 0.06 catalyst indicates that ethylene is a key source of acetic acid and carbon oxides.Furthermore, the reduction of the selectivity of ethylene along with the corresponding increase of the by-products (acetic acid and carbon oxides) clearly manifests that acetic acid and carbon oxides are the secondary products from the consecutive oxidation of ethylene.The effect of ethane conversion on the selectivities of ethylene, acetic acid, CO, and CO 2 at 235, 255, and 275 ∘ C over W 0.06 is plotted in Figure 8, demonstrating a strong dependence of ethane conversion and selectivity.It is observed that the selectivity decreases as ethane conversion increases.Also, the selectivity of by-products (i.e., acetic acid, CO 2 , and CO) shows a linear relation with the conversion of ethane.As the reaction temperature increases, the selectivity of ethylene decreases.Our experimental data suggest that ethylene undergoes substantial further oxidation, and the ODH of ethane can be described by the parallel consecutive scheme [43]. is shown in Scheme 1 [3,33,44].The kinetic parameters  0 ,   , and  for the ODH of ethane over the W 0.06 catalyst were obtained using a nonlinear regression.develop a kinetic model for the ODH of ethane, three assumptions were made: (1) the ODH is isothermal, (2) a catalyst deactivation is a function of time-on-stream (TOS), and (3) a single deactivation function is defined for all reactions, and a thermal conversion is neglected.Based on the assumptions, rates of the disappearance of ethane ( C 2 H 6 ) and the formation of ethylene ( C 2 H 4 ), acetic acid ( CH 3 COOH ), and carbon oxides ( CO  ), respectively, are described as follows:

Kinetic
Scheme 1: Schematic of catalytic reactions of the ODH of ethane.
where   is a concentration of species  in the units of mol/m 3 ,   is a reaction in the unit of time (i.e., at inlet of the catalyst bed,   = 0, and at the outlet of the catalyst bed,   = ),   is an apparent rate constant for the th reaction in the unit of s −1 , and  is the apparent deactivation function.The measurable variables from our chromatographic analysis are the weight fraction of the species,   , in the system.By definition the molar concentration,   , of every species in the system can be related to its mass fraction,   by the following relation: where  TM is the total mass flow rate (kg/min), MW  is the molecular weight of species  in the system, and   is the total volumetric flow rate (m 3 /min).Regarding catalyst deactivation, as deactivation functions can be expressed in terms of the catalyst time-on-stream ( = exp(−)), deactivation can also be related to the progress of the reaction [45], where  is a catalyst deactivation constant.The deactivation function based on time-on-stream was initially suggested by Voorhies [46].Substituting ( 5) into ( 1)-( 4), we have the following first order differential equations which are in terms of weight fractions of the species: where  1 ,  2 ,  3 ,  4 , and  5 , respectively, are given by The temperature dependence of the rate constants was represented with the centered temperature form of the Arrhenius equation; that is, where   is an average temperature introduced to reduce parameter interaction in the unit of K [47],   is the rate constant for reaction  at   (s −1 ), and   is the activation energy for reaction  (kJ/mol).Since the experimental runs were done at 235, 255, and 275 ∘ C, T o was calculated to be 255 ∘ C. The values of the model parameters along with their  ) and alkene combustion ( 3 ), respectively, over an alumina-supported vanadia catalyst.The apparent activation energies obtained for ethane and ethylene combustion to form carbon oxides in this present study show a similar order of magnitude with those reported in the literature.Using the estimated kinetic parameters, we carried out kinetic modeling to examine the validity of the simulated results, and compared them with our experiments finding (i.e., Figures 5 and 9).The fitted parameters were substituted into the comprehensive model developed for this scheme, and the equations were solved numerically by using the 4th-order-Runge-Kutta routine.It was found that our prediction is in good agreement with experiment (Figures 5  and 9), demonstrating the validity of the proposed kinetic model.It is assumed that the slight deviation observed in Figure 9 at 275 ∘ C might be due to the significant side products generated at this reaction temperature.It is proposed to carry out microkinetic modeling based on a mechanistic study using density functional theory (DFT) simulations [48,49], providing more detailed information of the conversion, selectivity, and yield of C 2 H 4 , CH 3 COOH, CO, and CO 2 on W-doped surfaces.The DFT based study would facilitate rationally design of novel MoVMnW mixed oxide catalysts.

Conclusions
MoVMnW mixed oxide catalysts were revisited to examine the ODH of ethane.It was found that the addition of W to MoVMn mixed oxide catalysts improves the catalytic activity toward C 2 H 4 , while lowering the formation of by-products (CH 3 COOH, CO, and CO 2 ).Using the MoV 0.4 Mn 0.18 W 0.063 O x (W 0.06 ) catalyst, we observed that the primary product of ethane oxidation is ethylene, which may go through consecutive reactions to CH 3 COOH, CO, and CO 2 .The best selectivity of ethylene on the catalyst was ∼90% at 235 ∘ C.However, a significant decrease in the selectivity of

Figure 4 :
Figure 4: Influence of W loading into MoVMn mixed oxide samples at 275 ∘ C and at a reaction time of 0.65 sec.

Figure 5 :Figure 6 :
Figure 5: Variation of ethane conversion against contact time achieved over the W 0.06 catalyst at 235 (◼), 255 (e), and 275 ∘ C ().The solid curves are predicted results based on the kinetic modeling.

Figure 8 :
Figure 8: Dependence of ethane conversion versus selectivity for (a) CH 3 COOH, CO, and CO 2 at 275 ∘ C and (b) ethylene at 235, 255, and 275 ∘ C in the ODH of ethane over the W 0.06 catalyst.The data sets are from different contact times.

Figure 9 :
Figure 9: Ethylene yield versus ethane conversion at the reaction temperatures of 235, 255, and 275 ∘ C. The solid, dashed, and dotted curves are predicted results based on the kinetic model.
[36]entioned, 6.3 mol% of W doping to MoVMn mixed oxide catalysts showed the best conversion of ethane by the ODH[35].Based on the previous study, we applied three materials of MoV 0.4 Mn 0.18 W 0 O x (W 0 ), MoV 0.4 Mn 0.18 W 0.063 O x (W 0.06 ), and MoV 0.4 Mn 0.18 W 0.14 O x (W 0.14 ).We considered W 0 and W 0.14 as the extreme cases without W and the highest doping, respectively, and W 0.06 as the optimal one in terms of the ethane conversion from the ODH.The synthesis method is described in detail elsewhere[36].Briefly, ammonium metavanadate (Fluka Chemical) was added to deionized water and heated to 85 or WO 4 , MoO 3 , Mo suboxide such as Mo 8 O 23 or Mo 9 O 26 , (M x Mo 1−x ) 5 O 14 (M = V or Mn or W), W 3 O, V 0.95 Mo 0.97 O 5 , and Mn 2 O 3 , respectively.
3 mol% W doping to MoVMn mixed oxide catalysts.In this study, to obtain a more systematic, optimal condition of MoVMnW based catalysts2.Experimental2.1.Catalyst Preparation.∘ C, with a constant stirring.A yellow colored solution was obtained.Oxalic acid (Riedel-de Haën Chemicals) was added with water to the solution with constant stirring, while the required amount of ammonium tungstate (BDH Chemicals) and manganese acetate (Riedel-de Haën Chemicals, resp.) was slowly added to the mixture.Then, ammonium heptamolybdate (Riedel-de Haën Chemicals) was added to the solution.Precipitates were dried in an oven at 120 ∘ C and calcined at 350 ∘ C for 4 hours.

Table 1 :
Contact-time dependence of selectivities for (a) CH 3 COOH, CO, and CO 2 at 275 ∘ C and (b) ethylene at 235, 255, and 275 ∘ C in the ODH of ethane over the W 0.06 catalyst.Theoretical selectivities a of ethylene and CO  .
Studies.A kinetic model was developed to understand the ODH of ethane on MoVMnW mixed oxide catalysts.A generally accepted scheme for the ODH of ethane a They were obtained by linear regression analyses at their zero conversion.

Table 2 :
Estimated kinetic parameters for the ODH of ethane on the W 0.06 catalyst.

Table 2 ,
while the resulting cross-correlation matrices are also given in Table3.The cross-correlation matrices give good results as indicated by a low correlation between most of the parameters, with only a few exceptions.From the results of the kinetic parameters presented in Table2, it was observed that the catalyst deactivation was found to be small,  = 0.03, indicating a low coke formation.As mentioned, the ODH of ethane can occur via the reaction network as shown in Scheme 1, in which ethane reacts with oxygen to form ethylene with a rate constant  1 , or carbon oxides (CO x ) with a rate constant  2 .Ethylene, then, undergoes a subsequent oxidation to CO x and CH 3 COOH with rate constants  3 and  4 , respectively.Rate constants  1 and  2 can be used to determine the ODH of ethane activities, while the ratios of rate constants (i.e.,  1 / 2 and  1 / 3 ) are used to determine the ethylene selectivity.High ethylene yields at a reasonable contact time apparently require high values of  1 and low values of  2 / 1 and  3 / 1 .The kinetic parameters ( 1 ,  2 ,  3 , and  4 ) summarized in Table2confirms the high selectivity of ethylene noticed in the ODH of ethane over the W 0.06 catalyst.

Table 3 :
[7]]elation matrix for the ODH of ethane on the W 0.06 catalyst.COOH from C 2 H 4 ( 4 ), respectively, following the order of  4 <  1 <  3 <  2 .Recently, Lin et al.[33]reported apparent activation energy of ∼51.5 kJ/mol for the formation of ethylene over a K-Y zeolite.Similarly, an apparent activation energy ( 1 ) of ∼63.2 kJ/mol was reported for the formation of ethylene over VO x /SiO 2 catalyst[6].These values are in good agreement with that calculated over the W 0.06 catalyst.Argyle and coworkers[7]reported apparent activation energies ∼115 kJ/mol and ∼60-90 kJ/mol for ethane combustion ( 2