Haitao Li*,Yin ZhangHongxi ZhangXiaoqin QinYalin XuRuifang WuZheng Jiang,Yongxiang Zhao*

1 Engineering Research Center of Ministry of Education for Fine Chemicals,Shanxi University,Taiyuan 030006,China

2 Faculty of Engineering and the Environment,University of Southampton,Southampton SO17 1BJ,UK

Keywords:1,4-Butanediol Hydrogenation Ni/SiO2-Al2O3catalyst Deactivation Regeneration

ABSTRACT The deactivation of Ni/SiO2-Al2O3catalyst in hydrogenation of crude 1,4-butanediol was investigated.During the operation time of 2140 h,the catalyst showed slow activity decay.Characterization results,for four spent catalysts used at different time,indicated that the main reason of the catalyst deactivation was the deposition of carbonaceous species that covered the active Ni and blocked mesopores of the catalyst.The TPO and SEM measurements revealed that the carbonaceous species included both oligomeric and polymeric species with high C/H ratio and showed sheet.Such carbonaceous species might be eliminated through either direct H2reduction or the combined oxidation-reduction methodologies.

1.Introduction

1,4-Butanediol(BDO)is highly desired because it has been widely used in the manufacturing of tetrahydrofuran,γ-butyrolactone,polybutylene and polyurethane polymers,poly(butylene succinate),3-buten-1-ol,and other key industrial organic chemicals[1-6].In industry,the Reppe method is the most applied process using acetylene as raw materials to produce BDO via a two-step process involving condensation and hydrogenation(see Fig.1).The first step is the condensation of acetylene(C2H2)and formaldehyde(CH2O)to produce 1,4-butynediol(BYD)on a supported CuO-Bi2O3catalyst,following with the second step,catalytic hydrogenation of BYD to produce BDO on Ni-based catalysts[7,8].In general,the BYD hydrogenation process is carried out in two stages in order to promote heat transfer and improve product quality[9].In the first stage,the BYD aqueous solution is hydrogenated under low temperature and pressure to generate a crude BDO aqueous solution that contains various unsaturated products,such as 1,4-butenediol(BED),4-hydroxybutyraldehyde(HBD),and other carbonyl compounds;in the second stage,the crude BDO solution is further hydrogenated at higher temperature and hydrogen pressure so as to ensure the unsaturated products to be hydrogenated into BDO.This second stage catalytic hydrogenation of crude BDO is paramount in determining the BDO quality that is highly dependent on the performance of catalysts.However,the catalyst is prone to deactivation under such a hydrothermal condition.

In previous studies on the Ni/Al2O3catalyzed liquid-phase hydrogenation of crude BDO in an aqueous solution[4,10],it was found that the Al2O3support would transform into boehmite under the typical hydrothermal reaction conditions.It is the Al2O3transformation that gives rise to the aggregation of active Ni species and decrease of surface area,pore volume,and pore diameter.As a consequence,the hydrogenation catalyst loses its activity.Therefore,improving the hydrothermal stability of Al2O3support is essential to prolong catalyst lifetime.Most recently,we have developed a Ni/SiO2-Al2O3hydrogenation catalyst with remarkable hydrothermal stability through adding SiO2promoter into the Ni/Al2O3[4,11].The catalyst maintained its structural and textural characteristics for a long time under hydrothermal conditions,showing excellent prospects for industrial applications.However,the catalyst significantly lost its activity after thousands of hours of hydrogenation time on stream.Metal catalyst deactivation is extremely complicated in the liquid-phase reactions,where different catalysts show various deactivation behaviors.Also multiple factors are responsible for the drop of the catalyst activity for a single catalyst[12-14].In order to realize the long time operation of the aqueous phase hydrogenation,it is necessary to clarify the key factors determining the catalyst deactivation.

Fig.1.Reaction mechanism for the hydrogenation of 1,4-butynediol.

In the present work,the factors determining the deactivation of the highly hydrothermally stable Ni/SiO2-Al2O3catalyst were verified via investigation of the structure evolution during the 2140 h liquidphase hydrogenation of crude BDO aqueous solution.The spent catalyst was also regenerated by different methods to identify the crucial factor responsible for deactivation.This study may provide paramount information for understanding deactivation of catalysts in some aqueousphase hydrogenations.

2.Experimental

2.1.Materials

The Ni/SiO2-Al2O3catalyst was prepared by the impregnation method.Prior to the impregnation the Al2O3support(SBET280 m2·g-1,VTotal0.58 cm3·g-1,and Dpore11.5 nm)was dried at 393 K for 3 h.100 g Al2O3support was impregnated by 90 ml tetraethyl orthosilicate(TEOS)ethanol solution with a Si concentration of 0.084 g·ml-1.Subsequently,the sample was dried at 393 K and calcined at 723 K for 3 h and then the SiO2-Al2O3support with a Si content of 7 wt%was obtained.The Ni/SiO2-Al2O3catalyst was prepared by wet impregnation method,using nickel nitrate salt(Ni(NO3)2·6H2O)as precursor and the above SiO2-Al2O3as support.After impregnation,the catalyst precursors were dried at 393 K for 12 h,calcined at 673 K for 3 h and then reduced by flowing hydrogen at 450°C for 2 h.The obtained Ni/SiO2-Al2O3catalyst containing 15.7 wt% Ni(marked as Ni-fresh).

2.2.Activity test

The hydrogenation of crude BDO aqueous solution was performed in a fixed bed reactor(i.d.15 mm).About 43 ml(32 g)of catalyst(size:0.8-1.2 mm)was loaded in the reactor.Both the modeled BDO feed and H2were introduced down-flow into the reactor.The reactor was heated to the reaction temperature by an electric furnace controlled by a temperature controller and the temperature was monitored using a thermocouple inserted into the catalyst bed.The reaction was carried out under the following conditions:T=120°C,PH2=4 MPa,LHSV=1.5·h-1,and H2flow rate=～500 ml·min-1.

The product of the first-stage hydrogenation of BYD in Reppe process was modeled for investigating the hydrogenation of the crude BDO aqueous solution,which was composed of 65 wt%H2O,30 wt%BDO,and a minor amount of partially hydrogenated species,BED,HBD,etc.The carbonyl number of the modeled crude BDO solution was approximately 10.6 mg KOH·g-1.In industrial process,the contents of carbonyl compounds in the hydrogenation product are the main factor for measuring the hydrogenation effect.Because the contents of carbonyl compounds in the crude BDO solution are low.Meanwhile,they are very complicated.Consequently,it is difficult to quantify the content of every carbonyl compound by GC or GC-MS methods,while the chemical titration is a suitable method and has been widely used in industrial process.The principle of chemical titration is that the carbonyl CO reacts with hydroxylamine hydrochloride,releasing hydrogen chloride;and then the hydrogen chloride is titrated with potassium hydroxide[4].The carbonyl number of the crude BDO solution in the hydrogenation reaction was used as the main probe to monitor the hydrogenation effect,lower carbonyl number higher activity of catalyst.

Four spent catalyst samples(1.5 g)were extracted from the reactor at the operation times of 54,276,740,and 2140 h(marked as Ni-54,Ni-276,Ni-740,and Ni-2140,respectively).An equivalent weight of fresh catalyst was added to the reactor to compensate the extracted catalyst before restarting the subsequent evaluation.

2.3.Regeneration of the spent catalyst

The 2140 h-aged catalyst(Ni-2140)was regenerated by four distinctive methods:(I)washing with THF,(II)washing with H2O,(III)reduction with H2,and(IV)oxidation with O2followed by reduction with H2.The THF and H2O washing methods were performed in a 100 ml stainless steel autoclave reactor.The washing conditions were 1.0 g of catalyst(particle size 0.28 mm to 0.45 mm),40 ml of H2O(or THF),150°C,and 4 MPa of H2pressure for 3 h.After washing,the sample was dried at 120°C for 3 h in N2flow.The samples after THF and H2O washing were labeled as Ni-2140-THF and Ni-2140-H2O,respectively.Reduction treatment was performed in a quartz tube.The reduction temperature and time were 350°C and 3 h,respectively.The sample was labeled as Ni-2140-R.Oxidation and reduction treatments were also carried out in a quartz tube.The sample was initially oxidized at 350°C for 3 h with air.Then,the sample was reduced at 350°C for 3 h with H2and labeled as Ni-2140-O-R.

The activity of the regenerated and fresh catalysts was tested in a 100 ml stainless steel autoclave reactor.The reaction conditions were 0.2 g of catalyst(0.28 mm to 0.45 mm),40 ml of feed,150 °C,and 4 MPa of H2pressure for 3 h.The carbonyl number of the feed after hydrogenation reaction was determined by a chemical titration method.

2.4.Characterization techniques

BET-specific surface area and pore size distribution were measured by N2physisorption at-196°C using a Micromeritics ASAP 2020 instrument.Prior to the measurement,the samples were degassed under vacuum at 90°C overnight.The BET-specific surface area was calculated from P/Po=0.05 to 0.3 in the adsorption branch,and the BJH pore size distributions were calculated from the desorption branch.

Powder X-ray diffraction(XRD)patterns were recorded using a Bruker D8 Advance X-ray diffractometer with a Cu Kαsource operated at 40 kV and 40 mA,in the 2θ Scanning range from 20°to 80°.

Measurements of temperature-programmed oxidation(TPO)were performed using a Netzsch STA 449 system equipped with a mass detector(HIDEN model QIC-20).50 mg of catalyst sample was heated in the presence of a gas mixture containing 15% O2in N2at a flow rate of 30 ml·min-1.The gaseous effluents were monitored as a function of temperature using a mass detector.A linear heating rate of 10°C·min-1was used throughout the analysis.

The morphology of the catalysts was analyzed by scanning electron microscopy using S4800(HITACHI Corporation,Japan)equipment.

3.Results and Discussion

3.1.Activity test

Fig.2.Change of carbonyl number of product with time over Ni/SiO2-Al2O3catalyst.

Fig.2 shows the results of the hydrogenation of crude BDO aqueous solution over the Ni/SiO2-Al2O3catalyst.As compared to the feed,the carbonyl number of the product after hydrogenation significantly reduced from 10.6 mg KOH·g-1to below 0.25 mg KOH·g-1,indicating that the unsaturated species were effectively converted to BDO through hydrogenation reaction over the Ni/SiO2-Al2O3catalyst.It can be seen,in the initial 54 h hydrogenation period,the carbonyl number of the product decreased from 0.25 mg KOH·g-1to 0.12 mg KOH·g-1.The carbonyl number was then slightly fluctuated between 0.12 and 0.15 mg KOH·g-1,revealing no appreciable deactivation was observed within the 1400 h hydrogenation.Subsequently,as the operation time prolonged from 1400 h to 2140 h,the product carbonyl number increased from 0.12 mg KOH·g-1to 0.25 mg KOH·g-1,indicating the catalyst was significantly deactivated.

3.2.Catalyst deactivation

To investigate the cause of deactivation,four spent catalyst samples were extracted from the reactor at operation times of 54,276,740,and 2140 h(marked as Ni-54,Ni-276,Ni-740,and Ni-2140),respectively.These catalysts were washed with deionized water and dried at 120°C for further characterization by N2physisorption,XRD,SEM,and TPO techniques.

3.2.1.XRD results

XRD patterns for the fresh and spent Ni/SiO2-Al2O3catalysts are comparatively shown in Fig.3.For all catalysts,the broad peaks at around 2θ=37.5°,45.5°,and 66.7°are assigned to γ-Al2O3phase,the other peaks at around 2θ=44.5°,51.8°and 76.4°are associated with metal Ni phase[15,16].For the Ni-fresh sample,the metal Ni diffraction peaks were very low,indicating that the metal Ni was highly dispersed on Al2O3.After the initial 54 h hydrogenation,the intensity of metal Ni diffraction peaks becomes slightly stronger along with the operation time,revealing that the crystallite size of metallic Ni became larger.The mean crystallite sizes of Ni in spent samples are calculated by the Scherrer formula using(200)reflection of fcc Ni.And the dependence of the Ni crystallite sizes on hydrogenation time is shown in Fig.3(b).The gradually increased crystallite sizes of the metallic Ni of the catalyst aged in extended hydrogenation operation indicate that the hydrogenation results in the Ni particle growth.Because the hydrogenation temperature is fairly lower than the reduction temperature,it is reasonable to conclude the growth of Ni should not be due to sintering of metallic Ni.It is no doubt that the aggregation of smaller active Ni species under the hydrothermal hydrogenation conditions favors the irreversible growth of Ni particle,and thus reduces the number of the reactive sites[12].We suppose the aggregation of Ni crystallite would be one cause for the catalyst deactivation,however,it must not be the predominant reason because the catalyst's Ni crystallite size became larger yet its reactivity remained within 1000 h time on stream.

3.2.2.N2physisorption

The textural properties of fresh and spent catalysts were characterized via N2physisorption.The obtained BET specific surface area(SBET)and the porosity results are summarized in Table 1.As compared with the Ni-fresh,the specific surface area of spent catalysts did not vary significantly during the operation time of 2140 h.However,the pore volume and mean pore diameter of the spent catalysts decreased as the operation time prolonged.The N2adsorption-desorption isotherms and the pore size distributions of Ni-fresh,Ni-740 and Ni-2140 samples are shown in Fig.4.The Ni-fresh and Ni-740 samples exhibited a typical IV adsorption-desorption isotherm with H3-type hysteresis loops within the relative pressure(P/Po)region from 0.4 to 0.95,indicating the existence of the slit-shaped pores.The two samples possessed almost the same pore size distribution in a broader region(2-13 nm).The results reveal that the pores of the catalyst were only slightly blocked for the Ni-740 as compared with Ni-fresh catalyst.The Ni-2140 sample still showed a type IV adsorption-desorption isotherm,though the adsorbed volume decreased considerably and the position of hysteresis loop shifted to lower relative pressures(P/Po=0.4-0.85),suggesting that most of the pores were blocked.Indeed,its pore size distribution was reduced to 2-5 nm.

Table 1 Textural properties of fresh and spent Ni/SiO2-Al2O3catalysts

Fig.3.XRD patterns(a)and the dependence of the Ni crystallite sizes on hydrogenation time(b)of the Ni/SiO2-Al2O3catalyst.

Fig.4.N2adsorption-desorption isotherms and pore size distributions of fresh and spent Ni/SiO2-Al2O3catalysts.

Considering the mediate temperature and reactants for the hydrogenation as well as nearly unchanged phase support(XRD section),we suppose the reduction of pore size and diameter for the deactivated catalyst(Ni-2140)should be arisen from carbonaceous species that blocked the catalyst internal pores.The reduced pore sizes would result in the increase of the internal diffusion resistance[17,18],and thus limit the reactants transporting to catalytic active sites.From the microfluidic engineering point of view,the change in textural properties is also an important factor leading to the deactivation of Ni/SiO2-Al2O3catalyst in the reaction time of 2140 h.

3.2.3.TPO results

Fig.5.TG-DTG curves of fresh and spent catalysts.

In order to justify our supposed pore blocking by carbonaceous species,temperature-programmed oxidation(TPO)characterization of the fresh and spent catalysts was performed on TG-MS.The TPO experiment can precisely determine the amount of carbonaceous species and discriminate slight variations in combustive property of the carbonaceous species[19,20].TG-DTG curves for fresh and spent catalysts are shown in Fig.5.For the Ni-fresh sample,the two-stage mass losses can be attributed to desorption of free water(from room temperature to 180°C)and removal of surface hydroxyl groups(from 180°C to 400°C,centered at 280 °C)of the catalyst,respectively.The first stage accounts for a 5.54 wt%mass-loss and the second 3.08 wt%mass-loss.The TG curves of the spent catalysts could be roughly divided into three mass-loss stages according to the DTG curves.The first stage(from room temperature to 180°C)accounts for approximately 4.5 wt%mass-loss due to the desorption of free water,which is slightly less than the fresh catalyst.The second and third stages(from 180°C to 450°C,no obvious boundaries,and centered at 220°C and 280°C,respectively)may be assigned to the removal of hydroxyl groups and burning off carbonaceous species deposited on thespent catalysts.Themass-lossisabout 4.99wt%,5.31wt%,6.23wt%,and 11.31 wt% for the Ni-54,Ni-276,Ni-740,and Ni-2140 spent catalysts,respectively.These results indicate that the carbonaceous species deposited on the Ni/SiO2-Al2O3catalyst increase as the operation time was prolonged while it is not a linear increase.

Fig.6(a)presents the DSC curves of the fresh and spent catalysts.For all the samples,the endothermic peaks with maximum at 85°C can be attributed to desorption of the free water.For Ni-fresh sample,there are no obvious exothermic peaks observed between 180 and 400°C,while the spent samples show obvious exothermic peaks,which are due to burning off carbonaceous species deposited on the catalyst.Only a slight increase is observed from the Ni-54 to Ni-740,indicating that the deposition of carbonaceous species on the Ni/SiO2-Al2O3catalyst is slow during the hydrogenation time spanning from 54 to 740 h.For the Ni-2140,the exothermic peak area is just about double that of Ni-54.

It is worth noting the exothermic peak due to carbonaceous peak may be resolved into two peaks with maxima at 220 and 280°C,suggesting there would be two different types of carbonaceous species deposited on the catalysts.The quantitative analysis of mass-loss of the spent catalysts should deliver the same conclusion.Therefore,the time dependence of mass-loss due to carbonaceous species is plotted in Fig.6(b),where two distinctive slopes can be seen.The sharp rise of the mass-loss within 0 to 54 h time on the stream can be assign to the surface adsorbed BDO products which has just reached an adsorption-desorption equilibrium.The slope of the mass-loss versus time becomes smaller between 54 and 2140 h reaction,suggesting there would be new carbonaceous species formed.Apparently,the TG-DTG and DSC results consistently reveal that the surface-deposited carbonaceous species would be more than one.

To further investigate the nature of the carbonaceous species deposited on the spent catalysts,the major gaseous effluents(CO2and H2O)from the TPO measurements were monitored using a mass spectroscopy.The MS profiles are presented in Fig.7.For each spent catalyst,the CO2profile shows two overlapping peaks with maxima at around 220 and 280 °C,while the H2O profile comprises three distinctive peaks:a sharp peak below 180 °C due to the desorption of the free water and two broad peaks with maxima around 220 and 280°C that may arise from burning off the two types of carbonaceous species.Apparently,compared to the carbonaceous species corresponding to the CO2-MS signal at 220°C,the carbonaceous species burnt off at 280°C would be more inert and responsible for the permanent deactivation.The results confirm our previous speculation in TG analysis that at least two types of carbonaceous species lay down on the catalyst surface.Moreover,the intensity and area of the peaks at 280°C gradually increase as the hydrogenation time is prolonged,especially for the Ni-2140 sample,revealing that the inert carbonaceous species are gradually covering the catalyst along with the hydrogenation reaction,thus result in significant deactivation at 2140 h.

Fig.6.DSC(a)and the time-dependence of the second and third stages mass-loss(b)of fresh and spent catalysts.

Fig.7.MS profiles(CO2and H2O)of spent catalysts.

It is worth comparing the CO2to H2O area ratios()for the peaks at 220°C and 280°C,which may roughly deliver the C/H ratio of the two type carbonaceous species.For convenience,the resonant sensitivity factors of CO2and H2O should be constant,so that we can directly compare their peak area.Thevalues at 220 °C are about 1.3,which are smaller than the values of 1.7 at around 280°C,regardless of the hydrogenation reaction time,revealing that the first type of carbonaceous species burning at 220°C possess low C/H ratio in the compounds and confirming they are due to the feed components strongly adsorbed on the catalyst(see TG-DSC section).The second type of carbonaceous species burning at 280°C has higher C/H mean ratio,which is attributed to oligomeric/polymeric species[21].

As shown in Figs.6(b)and 7,the areas of the exothermic CO2and H2O peaks centered at 280°C gradually increase along with the operation time,which consistently reveal that the oligomeric/polymeric carbonaceous species continuously lay down on the catalyst surface in the prolonged operation time[22,23].Because the carbonaceous species corresponding to the burning off temperature at 220°C was exhibiting in the initial hydrogenation stage(Ni-54)while the catalyst activity was not affected till 2140 h continuous reaction,hence it is reasonable to suppose the catalyst deactivation is less related to this carbon species.The deposited oligomeric/polymeric carbonaceous species would restrict the reactants to access and product desorption from the active site,and even block the pore channels leading to remarkable drop of the catalyst porosity as well as limit matter diffusion(see Section 3.2.2).Therefore,it can be clearly concluded that the surface oligomeric/polymeric species is the major factor leading to the deactivation of the Ni/SiO2-Al2O3catalyst.

3.2.4.SEM results

Fig.8 shows the scanning electron micrographs of the fresh catalyst and the typical spent samples.For the spent catalysts,there are some new sheet materials on their surface,which are related to the oligomeric/polymeric species.The results are corresponding to results of N2physisorption and TPO.Combining with N2physisorption,TPO analysis and reactivity,we may tentatively conclude this oligomeric/polymeric species should be responsible for the significantly deactivation of the catalyst.In order to confirm this speculation,catalyst regeneration experiments were conducted via various routes that helped to distinguish the contribution of the influence of the surface carbon species.

3.3.Characterization and evaluation of the regenerated Ni/SiO2-Al2O3catalyst

The spent Ni-2140 sample was regenerated using different methods as described in the Experimental Section.The TGA results of the regenerated catalysts are comparatively shown in Fig.9.It can be seen that the TG profile of Ni-2140 exhibited a two-stage mass-loss,which was almost overlapped with that of Ni-2140-THF and Ni-2140-H2O.The results reveal that the THF and H2O washing methods cannot effectively remove the carbonaceous species deposited on the surface of the spent catalysts,so that washing treatments would not effectively improve catalytic activity of the dead Ni-2140 catalyst.Indeed,the activity of the washed samples shows the very similar activity to the Ni-2140 in the hydrogenation of crude BDO solution.As shown in Table 2,their carbonyl number in the activity tests is slightly smaller than Ni-2140 which may be due to the partial removal of weakly adsorbed surface carbonaceous.In contrast,distinguishing from the washing treatment,the thermally regenerated samples through reduction(sample Ni-2140-R)or oxidation-reduction(sample Ni-2140-O-R)treatments exhibited interesting TG profiles.

For the sample reduced with H2,Ni-2140-R,two mass-loss stages in the TG curve were also observed but smaller mass-loss in each stage as compared to Ni-2140,suggesting that the reduction treatment can effectively eliminate the carbonaceous species.For the catalyst treated with air followed by H2,Ni-2140-O-R,only one mass-loss stage displays in the TG curve(from room temperature to 220°C),which may be attributed to the desorption of free water.Almost no mass-loss is observed above 220°C on the Ni-2140-R and Ni-2140-O-R samples,indicating that the carbonaceous species are almost removed completely.

Fig.8.SEM images of the fresh and typical spent samples.

In terms of the conclusions from BET tests as regards carbonaceous species removal levels,the porosity of the regenerated samples would differ on the dependence of regeneration methodology.As shown in Table 2,the specific surface area,pore volume and pore size of the Ni-2140,Ni-2140-THF and Ni-2140-H2O samples are almost identical,confirming their pores are still filled carbonaceous species and washing treatments are not effective.However,the porosity properties of the Ni-2140-R and Ni-2140-O-R samples are almost restored to the fresh catalyst levels.It is worth noting that the Ni-2140-O-R sample shows larger pore volume and pore size than the Ni-2140-R sample,suggesting that the oxidation-reduction is more effective to recover the Ni-2140 deactivated catalyst.The fact is that the Ni-2140-O-R sample showed comparable activity and is better than the Ni-2140-R sample in the hydrogenation assessment of the regenerated catalysts.

Fig.10.XRD patterns of spent and regenerated catalysts.

The XRD patterns of the Ni-2140 and samples regenerated via different methods are comparatively shown in Fig.10.Although all the samples exhibit γ-Al2O3and metal Ni diffraction peaks,the intensity of the Ni diffraction peaks of the thermally regenerated samples is much stronger in comparison with that of washed samples,indicating the thermal regeneration led to growth of the Ni particle.The average sizes of Ni crystallites(Table 2)of Ni-2140-R and Ni-2140-O-R are really slightly bigger than those of the Ni-2140-THF and Ni-2140-H2O samples,though the crystallite sizes of Ni-2140-THF and Ni-2140-H2O are also larger than that of Ni-2140.These results indicate that Ni crystallite would grow during all the different regeneration treatments yet the thermal regenerations are leading to the aggregation of metal Ni.

Combining the results of the TGA,N2physisorption and XRD characterizations and the activity of the regenerated samples(summarized in Table 2),it can be found that the activity recovery of the regenerated catalyst is closely related to the removal of carbonaceous species in the catalyst,in particular the inert carbonaceous species.This conclusion can well explain why the thermal regenerated samples,Ni-2140-R,especially the Ni-2140-O-R,are much more active than the washing regenerated samples,despite that the thermal regenerated samples are slightly less active than the fresh catalyst.The slightly lower activity of the thermally regenerated samples can be due to the growth/aggregation of the metallic Ni particle during the long-time hydrogenation reaction and thermal regeneration treatments.The results are not only confirming the thermal regeneration by H2reduction or oxidation-reduction treatments is realistically valid but also more importantly unveiling that the catalyst deactivation is caused by the surface deposition of stable and inert carbonaceous species,rather than the aggregation of metallic Ni particles or degradation of support.

4.Conclusions

The Ni/SiO2-Al2O3catalyst showed significant deactivation in the liquid-phase hydrogenation of crude BDO aqueous solution upon 2140 h operation despite the fact that the catalyst was hydrothermally stable.Through thorough characterizations of the catalyst operating at various hydrogenation periods and the regenerated samples via different treatments,it was found that the deactivation was mainly induced by the surface deposition of sheet inert carbonaceous species,despite the finding that two types of carbonaceous species lay down on the spent catalyst.On the basis of the C/H ratio retrieved from the TPO and regeneration results,the sheet carbonaceous species was attributed to oligomeric/polymeric carbonaceous species due to its high C/H ratio and could be burnt off above 280°C,while the other carbonaceous species would be the surface adsorbed feeding components.Though the aggregation of metallic Ni also slightly contributed to the catalyst deactivation under the present hydrogenation conditions,its influence was ignorable.In addition,it was found that the thermal regenerations of the deactivated catalysts via hydrogen reduction or oxidation-reduction were very effective to restore the catalyst activity.However,the washing treatments,regardless of using organic or water,could not remove the carbonaceous species and thus could not recover the catalyst activity.