Yanyong Li,Meng Ge,Jiameng Wang,Mengquan Guo,Fanji Liu,Mingxun Han,Yanhong Xu,Lihong Zhang,

1 Department of Catalysis Science and Technology,Tianjin Key Laboratory of Applied Catalysis Science &Technology,School of Chemical Engineering and Technology,Tianjin University,Tianjin 300350,China

2 Department of Materials Engineering,Xuzhou College of Industrial Technology,Xuzhou 221140,China

Keywords:Isobutane dehydrogenation Catalyst Nanoparticles Silica LaAlO3 perovskite Pt-Cu interaction

ABSTRACT In this study,isobutane dehydrogenation to isobutene reaction was carried out in a series of Pt-Cu bimetallic catalysts prepared by co-impregnation method.The catalysts were characterized by means of several techniques,including XRD,N2 adsorption–desorption,TEM,XPS,H2-TPR and TG.The results show that the existence of LaAlO3 perovskite can enhance the dispersion and sintering resistance of metal nanoparticles and facilitate the transfer of carbon deposits from active sites to the support.Interestingly,the perovskite nanoparticles can also inhibit the reduction of CuOx and the formation of PtCu alloys,resulting in the suitable interaction between Pt and Cu.The Pt-Cu/LaAlO3/SiO2 catalyst exhibits the optimal dehydrogenation performance with an isobutane conversion of 47%and isobutene selectivity of 92%after 310 min reaction,which was ascribed to the unique role of LaAlO3 perovskite as well as the appropriate Pt-Cu interaction.

1.Introduction

Isobutene,as an important chemical building block,is widely utilized for producing butyl rubber,polyisobutylene and other chemicals[1].At present,isobutene is mainly obtained as a byproduct in steam cracking of naphtha and fluid catalytic cracking of heave oil [2].However,with the ever-increasing market demand for isobutene,these conventional methods impede further application attributed to their high-energy consumption and relatively low selectivity[3].Therefore,the direct dehydrogenation of isobutane to isobutene has drawn extensive attention as an alternative to convention methods [4,5].

At present,researchers have developed different catalysts for the non-oxidative dehydrogenation of alkanes,such as Pt-based[6],CrOx-based [7],Pd-based [8] and Co-based catalysts [9],etc.Al2O3-supported Pt and CrOx-based catalysts are mainly used in industry[10].Considering the carcinogenicity to humans and hazards to the environment of Cr,hence,the Pt-based catalyst has been widely used for isobutane dehydrogenation in industry[2,11],due to the strong ability to break C-H bond of isobutane[12,13].Nevertheless,isobutane dehydrogenation is a highly endothermic reaction accompanied by volume expansion,which requires a higher temperature and lower pressure to obtain high yields of isobutene [14].Under these harsh conditions,catalyst deactivation caused by coke formation or sintering of metal nanoparticles and some undesired side reactions resulting in lower isobutene selectivity are inevitable [15–17].Consequently,it is essential to design a catalyst of superior dehydrogenation performance with resistance to carbon deposition and Pt sintering.

Proper catalyst support is very important to inhibit the sintering of Pt nanoparticles and improve the dispersion [18–23].Over the last decades,perovskite-type oxides (PTO) have been extensively investigated due to their structural stability and high thermal stability [24].However,PTO possesses a large size and relatively small surface area attributed to be synthesized by calcination at high temperature [25].Therefore,a feasible method to enhance surface area and decrease the particles size is to spread PTO nanoparticles in the pores of support material with high surface area [26].Generally,the carbon formation is closely related to strong acid sites on the support.Hence,mesoporous SiO2(S)with weak acid sites is employed as a high surface area support to limit the size of LaAlO3(LA)PTO along with metal nanoparticles.The stable structure of PTO can be easily maintained under a reducing atmosphere at high temperature due to its particularly weak reducibility,which is beneficial for stabilizing the metal nanoparticles[27].Furthermore,the presence of the oxygen vacancies in PTO helps to suppress the carbon formation.The reason is that the oxygen vacancies can lower the energy barrier for the oxygen activation and improve the mobility of the lattice oxygen.The lattice oxygen in PTO plays a major role in suppressing the carbon formation,so the existence of the oxygen vacancies can further improve lattice oxygen’s role and suppress the formation of carbon deposition [28,29].

To date,numerous researches have been also conducted to modify and develop Pt-based catalysts to improve the dehydrogenation performance mainly by adding additives(such as alkaline metals [30,31],Mg [32],La [33],Fe [34],Zn [35,36],In [17],Sn[15,37].However,Pt-based catalysts using Cu as a promoter to study the dehydrogenation performance of isobutane are rarely reported.The Cu addition can suppress carbon deposition and undesirable side reactions and increase energy barrier for C—C bond cleavage [11,38].Therefore,it’s of great significance to develop a Pt-Cu bimetallic catalyst to optimize the dehydrogenation performance of isobutane.

In this study,a series of SiO2supported Pt-based catalysts with or without LA were prepared by co-impregnation method.The effects of LA and Cu on the physico-chemical properties of catalysts were evaluated.Meanwhile,the relationship between the isobutane dehydrogenation performance,and the structure of Pt-based catalysts was investigated.The current work mainly studies the role of LA in Pt-based catalysts as well as the interaction among different metal particles and supports.

2.Experimental

2.1.Materials

All experimental materials are of analytical grade.Citric acid was obtained from Tianjin Yuanli Chemical Co.Ltd.,glycol was supplied by Tianjin Fengchuan Chemical Regent Co.Ltd.,H2PtCl6-﹒6H2O was purchased from Tianjin Kermel Regent Co.Ltd.,and SiO2was bought from Qingdao Bangkai High-tech Materials Co.Ltd.La(NO3)3﹒6H2O,Al(NO3)3﹒9H2O and Cu(NO3)2﹒3H2O were provided by Tianjin Fuchen Chemical Regents Factory.

2.2.Catalyst preparation

The SiO2-supported LaAlO3PTO sample(named simply as LA/S)was prepared by citric complexing method combined with an incipient wetness impregnation method.The support SiO2was pretreated at 700°C for 5 h in a muffle furnace.Firstly,La(NO3)3-﹒6H2O and Al(NO3)3﹒9H2O with a La/Al molar ratio of 1:1 were dissolved in deionized water.Then,the citric acid and ethylene glycol in a molar ratio of total metal ions:citric acid:ethylene glycol=1.00:1.20:0.24 were added to the mixed solution,stirred until completely dissolved.After that,SiO2was immersed into the above aqueous solution at room temperature overnight.Subsequently,the obtained solid was dried at 80°C for 6 h and 120°C for 12 h.Finally,the calcination was carried out to obtain the target sample at 350°C and 700°C for 2 and 5 h,respectively,at a ramping rate of 2°C﹒min-1.In this paper,the mass loading of PTO was controlled at 30.0 wt%.

The above as-prepared sample LA/S supported Pt-Cu catalyst was synthesized by co-incipient wetness impregnation (co-IWI)method.In brief,LA/S was impregnated in a mixed aqueous solution of H2PtCl6﹒6H2O and Cu(NO3)2﹒3H2O.After aging overnight at room temperature,the resulting solid was dried at 120°C for 12 h and calcined at 550°C for 4 h with a heating rate of 2°-C﹒min-1.Additionally,the Pt-Cu/S and monometallic catalysts,such as Pt/S,Cu/S,Pt/LA/S and Cu/LA/S,were also prepared by the same process.In all cases,the loading of Pt and Cu was 0.5 wt% and 3 wt%,respectively.

2.3.Catalyst characterization

X-ray diffraction (XRD) measurements were carried out on a D8-Foucas(Bruker)diffractometer equipped with a Cu Kα radiation(λ=0.15418 nm),operated at 40 kV and 40 mA.The patterns were recorded from 10° to 90° with a scanning speed of 8(°)﹒min-1.

N2adsorption–desorption isotherms were obtained with an automatic analyzer (QUA211007,Quantachrome,USA) at-196°C.Prior to the measurements,the samples were degassed under vacuum at 300°C for 4 h.The specific surface areas were determined based on the Brunauer–Emmett–Teller (BET) method,and the pore size distributions were calculated from the desorption branch of isotherms employing the Barrett–Joyner–Halenda (BJH)method.

Inductively coupled plasma-optical emission spectroscopy(ICPOES) was performed using a Perkin Elmer Optical Emission Spectrometer Optima 5300 DV to determine the elemental composition of the catalysts.

The temperature-programmed reduction (TPR) experiments were implemented with a conventional TPR apparatus.Prior to the TPR tests,0.05 g of fresh catalysts were pretreated in flowing N2at 400°C for 1 h.After the baseline was stabilized for 30 min,the reactor was heated from room temperature to 900°C at a heating rate of 10°C﹒min-1.5 vol%H2/N2was used as the reducing gas at a flow rate of 30 ml﹒min-1.The amount of H2uptake during the reduction was measured by a thermal conductivity detector(TCD).

X-ray photoelectron spectroscopy(XPS)was conducted to study the surface electronic states of the reduced catalyst components on a Thermo ESCALAB 250Xi spectrometer using Al Kα radiation.Before the test,all the samples were reduced under 5 vol% H2/N2(30 ml﹒min-1)at 600°C for 2 h,and the binding energies were calibrated by C1s level at 284.8 eV.

Transmission electron microscopy(TEM)images were obtained using a JEM-2100F field-emission transmission electron microscope.Prior to the test,the pre-reduced catalysts were prepared by grinding into powder,suspending and sonicating them in absolute ethanol,and then the suspension was dropped on a copper grid.

The NH3temperature-programmed desorption(NH3-TPD)measurements were carried out to determine the acid properties of the catalysts using a Micromeritics AutoChem II 2920 apparatus.Prior to NH3adsorption,120 mg sample was heated to 300°C at a heating rate of 10°C﹒min-1,and then 300°C for 2 h under He stream with 50 ml﹒min-1.After cooling to 100°C,NH3was adsorbed using a flow of 10 vol%NH3/He(50 ml﹒min-1)for 1 h.The NH3desorption was firstly performed in He (50 ml﹒min-1) for 1 h to remove the physically adsorbed NH3on the surface of samples,and then the samples were heated to 700°C in He (50 ml﹒min-1) under a ramp rate of 10°C﹒min-1,and NH3desorption profile was registered with a thermal conductivity detector (TCD).

Thermogravimetric and differential thermo analysis (TG-DTA)was carried out to determine the amount of carbon deposited on the spent catalysts on a DTG-50/50H thermal analyzer.The sample was heated from room temperature to 900°C with a ramp of 10°-C﹒min-1in air.

2.4.Catalytic activity measurements

The catalytic activity measurements were evaluated at atmospheric pressure and 600°C in a stainless-steel continuous fixedbed reactor with an inner diameter of 8 mm.Prior to reaction,0.5 g of catalyst samples (40–60 mesh) were reduced in 5 vol%H2/N2atmosphere(30 ml﹒min-1)at 600°C for 2 h.After reduction,isobutane and hydrogen (C4H10:H2molar ratio=1:1) were introduced into the reactor using mass flow controllers with a total flow of 20 ml﹒min-1.The weight hourly space velocity(WHSV)of isobutane was 3 h-1.The composition of the gaseous product was analyzed by an online gas chromatograph equipped with a fame ionization detector (Al2O3packed column).

Conversion (Conv) and selectivity (Sel) were calculated using the following equations:

where nC4H10,inand nC4H10,outstand for the number of moles of isobutane in feed and exit gases respectively.ni,outrepresents the number of moles of hydrocarbon product (CH4,C2H4,C2H6,C3H6,C3H8,C4H8) in the effluent gas.

3.Results and Discussion

3.1.Structure and phase composition of fresh catalysts

3.1.1.XRD

The wide-angle XRD patterns of the calcined and reduced catalysts are exhibited in Fig.1.Based on previous studies [27,39,40],the PTO nanoparticles should be easily constructed on SiO2or Al2O3support.However,the characteristic diffraction peaks assigned to PTO cannot be detected owing to the small size and high dispersion [27].It can be seen that all of the samples show a broad diffraction peak of SiO2(JCPDS file No.39–1425).Especially for LA/S supported catalysts,the diffraction peaks of SiO2become significantly weaker,which are attributed to the formation of PTO nanoparticles.

From Fig.1(a),some strong diffraction peaks of CuO (JCPDS file No.80-1916) and Pt (JCPDS file No.87-0646) appear in the XRD patterns of calcined Cu/S and Pt/S catalysts,respectively.This indicates the interaction between monometallic element and SiO2support is weak.As a result,the aggregation of CuO and Pt is inevitable.

As for calcined Pt-Cu/S,the CuO peaks become weak along with the disappearance of Pt peaks and appearance of PtO2diffraction peak at 2θ of 34.7° (JCPDS file No.71-0568).It implies that there is the interaction between Pt and Cu species,which is beneficial to disperse from each other.Compared with the SiO2supported catalysts,the diffraction peaks of the corresponding metal species on the LA/S supported catalysts are so weak that some characteristic peaks cannot be detected.It means that the addition of LA plays an important role in promoting the dispersion and inhibiting the aggregation of metal species,which should be conducive to the dehydrogenation of isobutene.Even,the high dispersion of metal species can be maintained for the reduced LA/S-containing catalysts as shown in Fig.1(b).Additionally,the Pt diffraction peaks become sharper in the reduced Pt/S sample,but disappear in the reduced Pt-Cu/S.This further proves that the Cu addition can improve the Pt dispersion [41].Meanwhile,one PtCu alloy peak can be seen for the reduced Pt-Cu/S due to the excessive Cu addition[42,43],whereas no PtCu alloy peak appears in the reduced Pt-Cu/LA/S.It can be speculated that the presence of LA nanoparticles can improve the PtCu alloy dispersion beyond the XRD detection limit or may be detrimental to form the PtCu alloy.Next,this phenomenon will be discussed in detail combined with a series of characterization methods.

3.1.2.N2adsorption–desorption

The low-temperature N2adsorption–desorption isotherm curves and pore size distributions (PSDs) of SiO2and different Ptbased catalysts after calcination are represented in Fig.2.The corresponding textural parameters are summarized in Table 1.All samples depicted in Fig.2(a)show a type IV isotherm with H2hysteresis loops,revealing the existence of mesopores based on the IUPAC classification [44].Furthermore,it also indicates the mesoporous structure is still retained even for the supported samples.

Compared to the pure SiO2,all the supported catalysts show an obvious decrease in the SBET,Vpand Dpvalue,which can be attributed to the blockage of the pores by nanoparticles[45].Meanwhile,it can be found that the incorporation of LA results in a most significant decrease in the SBETand Vpvalue of Pt/LA/S and Pt-Cu/LA/S(Table 1).This can be explained by the formation of LA with high loading in the porous channels of SiO2,where small LA particles with high dispersion can be produced,which is consistent with the XRD results.

As listed in Table 1,the Vpvalue of Pt/S is smaller than that of Pt-Cu/S due to agglomeration and stacking of metal nanoparticles.However,the Dpand Vpvalues of Pt-Cu/LA/S are slightly lower than that of Pt/LA/S.It suggests that the presence of LA can prevent the metal nanoparticles from being stacked and promote high dispersion,which is in line with the XRD analysis.

3.1.3.ICP-OES

The actual metal contents of Pt and Cu in all catalysts determined by ICP-OES technique are listed in Table 2.It can be seen that the actual Pt and Cu content are similar to their theoretical values,indicating complete impregnation of the metal ions on the support in the synthesis process.

Table 1 Textural properties of SiO2 support and calcined catalysts

Table 2 Elemental compositions of different catalysts determined by ICP-OES

Fig.1.XRD patterns of the calcined catalysts (a) and reduced catalysts (b).

Fig.2.(a) Nitrogen adsorption and desorption isotherm curves and (b) pore size distributions of calcined catalysts.

3.1.4.TEM

The morphology and particle size distribution of the metallic phases in the reduced catalysts are presented in Fig.3 and the average particle sizes of the reduced and spent catalysts are listed in Table 3.As can be seen from Fig.3(b),(d),(h),the lattice spacing of the metal particle is 0.226 nm corresponding to the Pt (111)plane.Additionally,the Cu (200) plane can be easily found over the Pt-Cu/LA/S catalyst (Fig.3(h)),which conformed from the lattice spacing of 0.181 nm.It is worth noting that the lattice fringes of LA are exhibited in Fig.3 (h),confirming that LA has been successfully synthesized on SiO2support.As displayed in Fig.3 (f),the lattice spacing of the metal particles is 0.215 nm,indicating the formation of PtCu alloys in the reduced Pt-Cu/S catalyst [11],which is in good agreement with the XRD results.On the contrary,no PtCu alloy phase is found in the reduced Pt-Cu/LA/S catalyst,which is in accord with the analysis of XRD above.

The monometallic Pt/S catalyst represents a wide distribution of Pt nanoparticle size and a big average particle size of (2.50±0.62)nm,which can be ascribed to the weak interaction between the Pt nanoparticles and the SiO2support.For comparison,the Pt/LA/S catalyst depicted in Fig.3(c)displays a relatively narrow distribution range of Pt nanoparticle size with small average particle size of(1.36±0.23)nm and the Pt nanoparticles still retain a relatively small size of (1.40±0.38) nm after 310 min reaction (Table 3).It can be determined that adding LA is in favor of improving the dispersion of Pt nanoparticles and increasing the Pt-support interaction.

For the bimetallic catalysts,as shown in Fig.3 (e),(g),the Cu doping can further shrink the metal particle size distribution range and decrease the metal average particle size.This should be attributed to the Pt-Cu interaction and indicates that the dispersion effect of Cu on Pt is remarkable.Certainly the introduction of LA has a further significant effect on the dispersion of the metal particles based on the decreasing particle size.However,as listed in Table 3,it seems clear that the reaction significantly increases the metal average particle size of Pt-Cu/S,while the growth ratio of the metal particle size of the reduced Pt-Cu/LA/S sample is far below that of Pt-Cu/S sample.This further illustrates that the LA play an important role on stabilizing and dispersing metal sites.

Taking what has been discussed above about TEM results into consideration,it can be deduced that the good stabilization and high dispersion of active Pt species on the reduced Pt-Cu/LA/S sample are depending on the relative strength of Pt-Cu and Pt-support interaction.Just LA plays a coordinating role between two different interactions.

3.1.5.XPS

The Chemical states of Cu element in representative samples,such as the reduced Pt-Cu/LA/S and Pt-Cu/S catalysts,are investigated by employing the XPS analysis technique.The Pt 4f spectrum is not studied here due to the difficulty to distinguish from the Al 2p peak [6].Fig.4 shows the XPS spectra of Cu 2p3/2region of the reduced bimetallic catalysts.The binding energies of various valence states of Cu 2p are particularly similar,which is difficult to distinguish [43].However,for the reduced Pt-Cu/LA/S catalyst,the existence of the shake up satellite at around 944.2 eV generally validates the presence of CuO [46].Additionally,it should be known that in addition to CuO,the presence of metallic copper cannot be ignored,as evidenced by the results of XRD and TEM analysis.It can be concluded that the addition of LA can strengthen the interaction between the metal nanoparticles and the support and inhibit the reduction of CuO,which makes it difficult to form PtCu alloys.As a contrast,no satellite peak appears in the Cu 2p3/2region of the reduced Pt-Cu/S catalyst.Combined with the XRD and TEM analysis,it can be deduced the CuO phase should be fully reduced into the form of metallic state[47],which be able to interact with metal Pt nanoparticles to easily form PtCu alloy in the reduced Pt-Cu/S catalyst.

Fig.3.HR-TEM images and the statistics of particle size distribution in the insets of TEM pictures of the reduced catalysts.(a,b)Pt/S;(c,d)Pt/LA/S;(e,f)Pt-Cu/S and(g,h) Pt-Cu/LA/S.

3.1.6.H2-TPR

H2-TPR tests are carried out to investigate the reducibility of the calcined catalysts exhibited in Fig.5.No any peak corresponding to the reduction to PtOxcan be found for the monometallic Pt-based catalysts (Pt/S and Pt/LA/S) due to the low loading of platinum.Another important reason is that the platinum species is completely reduced to Pt0in the calcination process[48,49],in consis-tent with the XRD results.The H2consumption values corresponding to different reduction peaks are provided in Table 4.Pt/S and Pt/LA/S are not listed in table due to the lack of reduction peaks.

Table 3 Variation in average particle size determined from TEM of reduced and spent catalysts

Table 4 Quantified H2 consumption for the different catalysts

Fig.4.XPS spectra of the reduced Pt-Cu/LA/S and Pt-Cu/S.

Fig.5.H2-TPR profiles of the calcined catalysts.

It can be observed that the TPR profiles of all the Cu-containing catalysts except for Pt-Cu/S can be divided in two major reduction peaks.For Cu/S and Cu/LA/S,the peak at low temperature can be assigned to the reduction of the well dispersed copper oxide species while the peak at high temperature is mainly due to the reduction of bulk CuO [50].Apparently,compared with Cu/S,two reduction peaks of Cu/LA/S shift to a higher temperature after adding LA meanwhile the intensity decreases obviously,indicating that the presence of LA can enhance the interaction between metal element and support and inhibit the reduction of copper oxide species.For Pt-Cu/LA/S,the peak centered at around 230°C can be attributed to the co-reduction of PtOxand highly dispersed CuO species,the peak located at approximately 310°C is mainly ascribed to the reduction of bulk CuO phase [51].

For Pt-Cu/S,the successive peak can be deconvoluted into three reduction peaks.The first peak at around 101°C should be assigned to the reduction of PtOxspecies.However,it can be noted that the hydrogen consumption value of Pt-Cu/S (see Table 4) at this temperature has exceeded the theoretical hydrogen consumption value of complete reduction of PtOxspecies.This indicates that there is also a reduction of some CuOxspecies near the Pt species.Combined with the XRD results,the reduced Pt and Cu species are tended to form Pt-Cu alloy through the interaction with each other.The attribution of other two peaks centered at 190°C and 254°C is same with that of two peaks of Cu/S,respectively.The main difference is that the peak temperatures of Pt-Cu/S are lower than that of Cu/S,suggesting that the addition of Pt species can promote the reduction of copper oxide species.

Fig.6.NH3-TPD profiles of LA and LA/S catalysts.

As shown in Table 4,compared with the catalyst without LA,the total hydrogen consumptions of the catalysts with LA are significantly lower than that of catalysts without LA.It can be speculated that the copper oxides in the Pt-Cu/LA/S catalyst have not been completely reduced,which can be assigned to the poor reducibility of the highly dispersed copper oxide species strongly interacting with the support [50,52].

3.1.7.NH3-TPD

The acidity of catalysts was measured by NH3-TPD,the collected TPD profiles were presented in Fig.6.For LA perovskite,two obvious desorption peaks are found ascribed to the weak (α peak) and the medium (β peak) acid sites,respectively.Moreover,no more desorption peak was found at higher temperature,suggesting the absence of strong acid sites [53,54].When LA perovskite is loaded on SiO2support，only a broad desorption peak attributed to the weak acid site is observed for LA/S,indicating that the presence of SiO2weakens the acidity.In addition to improving the dehydrogenation conversion,the presence of weaker acid sites can also suppress some side reactions,such as polymerization,isomerization and cracking [27].

3.2.Catalytic performance

The catalytic performance of isobutane dehydrogenation over a series of reduced catalysts as functions of time on stream is represented in Fig.7.It is observed that the isobutane conversion,isobutene selectivity and yield of these catalysts decrease in the order of Pt-Cu/LA/S>Pt/LA/S>Pt/S>Pt-Cu/S>Cu/S ≈Cu/LA/S after reaction of 310 min.It can be found that there is no clear difference in dehydrogenation performance between Cu/S and Cu/LA/S with the very poor performance.This suggests that the Cu element is only a promoter and not the active component.

On the other hand,in comparison with the Pt/S catalyst,the Pt/LA/S catalyst shows slightly higher activity and selectivity.Especially,the crucial role of LA in the dehydrogenation reaction is extraordinarily obvious by comparing the activity of Pt-Cu/S catalyst and the Pt-Cu/LA/S catalyst.It is revealed that the incorporation of LA is advantageous to the reaction,mainly due to its synergistic effect with SiO2to stabilize metal particles and improve the dispersion of metal Pt.

Fig.7.Catalytic performance of isobutane dehydrogenation over different catalysts:(a) isobutane conversion and isobutene selectivity;(b) isobutene yield (reaction conditions:600°C,atmospheric pressure,H2:i-C4H10=1:1 (molar ratio),WHSV=3 h-1,mcat=0.5 g).

Fig.8.(a) TG profiles of the spent Pt-based catalysts;(b) TG-DTA profiles of the spent Pt-Cu/S and Pt-Cu/LA/S catalysts.

Furthermore,it should be noted that the co-impregnation of Pt and Cu on the LA/S support results in a superior catalytic performance and much better than the corresponding monometallic Pt catalyst.The initial and final yields of isobutene of Pt-Cu/LA/S are 60.2% and 43.7%,respectively,and the selectivity of isobutene is always stable at approximately 92%over the 310 min run.All these should be assigned to the close interaction between Pt and Cu.In fact,the Cu and LA addition appear to regulate the interaction of metal Pt with support,which can improve the Pt dispersion and prevent the aggregation of Pt species on the surface of LA/S,but also build an appropriate chemical environment for the rapid desorption of isobutene[7].Nevertheless,as displayed in Fig.7(a),the addition of Cu into Pt/S catalyst gives rise to an evident decline in the activity and selectivity,probably ascribed to the strong Pt-Cu interaction resulting in the formation of PtCu alloy,which adversely affecting the activation ability of Pt.For the active Pt sites,the formation of PtCu alloys can reduce the Pt surface,which may weaken the ability of Pt in dehydrogenation [42].

All in all,LA and SiO2synergism act as a carrier together and Cu as an auxiliary agent are beneficial to the dehydrogenation of isobutane.The Pt-Cu/LA/S catalyst has exhibited superior activity and excellent stability among the several catalysts studied,which is a quite promising catalyst applying in the process of alkane dehydrogenation.

3.3.Analysis of spent catalysts

3.3.1.TG

Generally,carbon formation has been considered to be one of the main reasons of catalyst deactivation arising from the coverage of active metal sites by coke deposits[11,55].After reaction,the TG tests are carried out to determine the deposited carbon content of spent catalysts and the results are shown in Fig.8.The mass loss is attributed to the burning off of carbon deposits on the spent catalysts.

As displayed in Fig.8(a),a decreased amount of carbon deposits is observed in accordance with the order of Pt-Cu/LA/S>Pt/LA/S>Pt/S(≈Pt-Cu/S).This tendency is nearly consistent with the corresponding catalytic performance as described in Fig.7.The low deposited carbon amount on the Pt/S and Pt-Cu/S catalysts is obvious related to their poor activities determined by the sintered metal nanoparticles,as described in the XRD and TEM results.

The DTA profiles of Pt-Cu/S and Pt-Cu/LA/S in Fig.8 (b) show two peaks.The first peak at lower temperature is related to the combustion of carbon deposition on the metal site.Another peak at higher temperature can be assigned to the burning of carbon formation on the support [6,56].Compared with Pt-Cu/S,it can be found that the carbon deposition of Pt-Cu/LA/S is mainly on the support.Furthermore,it can be observed that the introduction of LA perovskite reduces the combustion temperature of carbon deposition,which is beneficial to the elimination of carbon deposition.The Pt-Cu/LA/S catalyst,which possesses the most amount of carbon content,shows the excellent dehydrogenation performance.This indicates that the addition of LA facilitates the transfer of carbon deposits from the active site to the support [27].Simultaneously,LA plays an important role in preventing irreversible deactivation caused by the sintering of metal nanoparticles.

3.3.2.TEM

The TEM images of the spent Pt-Cu bimetallic catalysts are represented in Fig.9.It can be seen that the metal particles in the Pt-Cu/S catalyst undergo severe agglomeration during the dehydrogenation reaction and shows a relatively wide distribution range of metal particles size.From TG analysis,the Pt-Cu/S catalyst has a small amount of carbon deposition.Hence,the agglomeration and sintering of metal particles is the main reason for the deactivation of the Pt-Cu/S catalyst.

As shown in Fig.9(b),a great deal of graphite carbon and amorphous carbon can be clearly observed in the spent Pt-Cu/LA/S catalyst,which is consistent with TG result [57].Most interestingly,although there exists a lot of carbon deposition,many highly dispersed metal particles can still be found in the TEM image,further indicating that carbon deposition is mainly on the support.In spite of the large amount of carbon formation,the deposited carbon did not deactivate the Pt-Cu/LA/S catalyst (see Fig.7).These indicate that the carbon deposits must be transferred from the metal active sites to the support,in order to avoid the active component being covered [55].Therefore,it can be inferred that the addition of LA facilitates the transfer of carbon deposits from the active site to the support.

3.3.3.XRD

After a period of dehydrogenation reaction,the wide-angle XRD patterns of the spent catalysts are shown in Fig.10.Compared with the reduced Pt/S catalyst (Fig.1(b)),the diffraction peaks of Pt in the spent Pt/S catalyst become a little sharper and higher.This indicates that the Pt metal nanoparticles undergo the agglomeration and sintering in the reaction process,resulting in the increase of metal particles size,which is in accordance with the results of TEM analysis.The same phenomenon also occurs in the spent Pt-Cu/S sample.In contrast,no diffraction peak corresponding to Pt is appeared for Pt/LA/S even undergone a period of reaction,which once again shows the vital importance of LA in stabilizing the metal nanoparticles and improving the dispersion.

Fig.9.HR-TEM images of the spent Pt-Cu bimetallic catalysts.(a) Pt-Cu/S;(b) Pt-Cu/LA/S.

Fig.10.XRD patterns of the spent catalysts.

For Pt-Cu/LA/S with the optimal dehydrogenation performance,the diffraction peak of carbon with graphitic nature can be observed at 2θ of 26.2°(JCPDS file No.75-1621).However,no peak assigned to the graphitic carbon is detected for other Pt-based catalysts.This is closely related to the amount of carbon deposits as depicted in TG results.Most interestingly,for Pt-Cu/LA/S,in addition to the diffraction peaks of Cu,the peaks of CuO and Cu2O(JCPDS file No.34-1354) are also found,whereas not be appeared for Pt-Cu/S.Generally speaking,CuO is easily reduced to metallic Cu at low temperature [58].Combining with the XPS and H2-TPR results,this can be explained by the fact that the presence of LA restricts the ability of the oxidized Cu to be reduced to metallic Cu,which leads to the partial reduction of CuOx.On the other hand,LA can enhance the dispersion of metal nanoparticles.Hence,considering these above two factors,it can be concluded that it’s not readily to form the PtCu alloy for Pt-Cu/LA/S,as clarified by the results of XPS and H2-TPR analysis.

4.Conclusions

In summary,the method of utilizing porous SiO2to confine the size of PTO nanoparticles and then co-impregnating Pt and Cu is particularly feasible to improve the dehydrogenation performance of isobutene.Pt-Cu/LA/S shows an initial and final isobutane conversion of 67% and 47%,respectively,and the conversion shows a quite stable trend during the reaction.The optimal dehydrogenation performance of Pt-Cu/LA/S can be ascribed to the unique role of LA and appropriate interaction between Pt and Cu.The presence of LA can improve the dispersion and sintering resistance of metal nanoparticles and promote the transfer of carbon deposits from active sites to the support.Furthermore,another important role of LA can inhibit the partial reduction of CuOx,which is not conducive to the formation of PtCu alloys.From the analysis results of a series of characterizations,appropriate interaction between Pt and Cu without forming PtCu alloys is beneficial for isobutane dehydrogenation,which opens a new path for future research.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21776214) and State Key Laboratory of Chemical Resource Engineering.