Chen Xu,Zhenyi Du,Shiqi Yang,Hongda Ma,Jie Feng

Training Base of State Key Laboratory of Coal Science and Technology Jointly Constructed by Shanxi Province and Ministry of Science and Technology,Taiyuan University of Technology,Taiyuan 030024,China

Keywords:Catalyst Instability Steam reforming of toluene Adsorption Ni/biochar Potassium

ABSTRACT Biochar supported nickel (Ni/BC) has been widely studied as a cheap and easy-to-prepare catalyst with potential applications in tar reforming during the gasification of low-rank fuels,such as brown coal and biomass.However,the role and behaviors of inherent K species,especially their interactions with Ni particles and the biochar support,are not well understood yet.In this work,three Ni/BC catalysts with varying K amount were prepared from raw,water-washed,and acid-washed biomass.They were used in steam reforming of toluene as a tar model compound to elucidate the effects of inherent K on the catalytic activity and stability.Detailed characterization indicated that K enhanced water adsorption due to its hydroscopicity and lowered the condensation and graphitization degrees of biochar,but the alteration to the electronic state of Ni was not observed.These effects together led to a temperature-dependent role of K.That is,at relatively low temperatures of 450 and 500°C,toluene conversion was increased in the presence of K,due to the increased concentration of adsorbed water around Ni particles.By contrast,at relatively higher temperatures of 550 and 600°C,although initial high activity was achieved,Ni/BC with K deactivated rapidly because of the accelerated consumption of the biochar support.

1.Introduction

Gasification is a technology for converting carbonaceous materials to gaseous products,which can be further used for power generation,fuel and chemical synthesis[1–3].Gasification of low-rank fuel,including brown coal and biomass,under relatively low temperatures is beneficial in improving the cold-gas efficiency.However,the high volatile contents in brown coal and biomass result in high tar yield,and it is still one of the major challenges in low-rank fuel gasification [4–7].Tar can cause pipeline corrosion and clogging,and it accounts for 5%–15% of the effective energy in the feedstock [8].Therefore,it is necessary to remove tar for the stable and efficient operation of gasification facilities.

Catalytic steam reforming can effectively remove tar generated during the gasification of low-rank fuels [9],and the development of efficient and low-cost catalysts is the key for this process.Char derived from the pyrolysis/carbonization of brown coal or biomass,is a promising catalyst/support for steam reforming,as it can be produced from the feedstock itself with low cost [10–12].Meanwhile,char-supported catalysts can be gasified or combusted to recover the energy from char when deactivated[13,14].Therefore,char and char-supported metal catalysts have been widely studied in tar reforming [15–17].Shenet al.[18] prepared carbonsupported Ni catalysts and studied the effects of surface properties of carbon support and nickel precursors on Ni nanoparticle sizes and catalytic performances for steam reforming of toluene.In our previous study [19],a series of Ni/biochar (Ni/BC) catalysts were prepared with varying Ni loadings under different pyrolysis temperatures.Since that work was focused on the relationship between the Ni particle size and the catalytic performance,the starting biomass for catalyst preparation was washed with HCl solution to remove the alkali and alkaline earth metallic species(AAEMs),in order to exclude their potential influence on the catalytic performance.AAEMs,particularly potassium(K),exist inherently in biomass materials[20].After pyrolysis,they accumulate in the resultant biochar,which may affect the adsorption and activation of H2O,as reported in the literature [20–22].Huet al.[20]found that the inherent AAEMs promoted water–gas shift reaction,and enhanced the yield of H2and CO2during biomass pyrolysis.Liet al.[21]prepared Na-Ir/TiO2-R catalysts and studied the effects of Na for formaldehyde oxidation,and the results showed that the addition of Na greatly promoted the activation of both chemisorbed oxygen and H2O.Based on density functional theory calculations,Wanget al.[22] reported that the presence of K can promote the catalytic activity of various metals for H2O dissociation to varying degrees.Therefore,the inherent AAEMs in biochar-supported catalysts may improve the catalytic activity by strengthening the adsorption and dissociation of H2O.However,as revealed by Yipet al.[23] and Fenget al.[24] the inherent AAEMs also have significant catalytic effects on the gasification of biochar.In other words,for biochar-supported metal catalysts,the presence of AAEMs may enhance the steam reforming activity of catalysts;on the other hand,it may accelerate the gasification reaction between steam and the biochar support.

Therefore,based on our previous study[19],we conducted this work to elucidate the effects of the inherent AAEMs(particularly K)on the catalytic performance of Ni/BC in the steam reforming of toluene as a tar model compound [25].Three Ni/BC catalysts with different amount of K were prepared from raw,water-washed and acid-washed biomass,and used in toluene steam reforming in the temperature range 450–600°C.

2.Experimental

Corncobs collected from a local farm were crushed into (0.25–0.42)-mm granules as the biomass feedstock.In order to study the catalytic effect of AAEMs,three types of corncobs were prepared,namely raw,water-washed and acid-washed samples.The water-washed samples were prepared by immersing 15 g of corncobs in 150 ml of deionized water for 12 h under stirring.After that,the samples were thoroughly washed with about 5 L deionized water and dried before use.10 g of corncobs were subsequently incipient-wetness impregnated with an aqueous solution of Ni(NO3)2·6H2O,and then the Ni salts-loaded corncobs were stirred for 24 h and dried at 105°C.The dry samples were heated at 10°C·min-1under N2atmosphere to 600°C and held for 0.5 h.After cooling in N2gas flow to room temperature,the catalysts were collected and stored in an airtight container before use.The acid-washed samples were synthesized according to similar procedures to the water-washed ones except that 1 mol·L-1HCl solution was used instead for the thorough removal of AAEMs.The Ni loadings were 5 wt%for all of the samples.The catalysts prepared from raw,water-washed and acid-washed corncob samples are designated as Ni/BC-RAW,Ni/BC-WW,and Ni/BC-AW,respectively.The catalyst evaluation procedures and characterization details are provided in the Supplementary Material.

3.Results and Discussion

3.1.Gasification reactivity of different Ni/BC catalysts

Based on inductively coupled plasma mass spectrometry (ICPMS) analysis,K element (7.1 mg·g-1) was dominant among all of the inherent AAEMs (Ca:0.2 mg·g-1;Na:0.6 mg·g-1;Mg:0.3 mg·g-1) in corncobs,and therefore this study mainly focused on the effects of K.The K contents were 25.0,4.2,and 0.1 mg·g-1for Ni/BC-RAW,Ni/BC-WW,and Ni/BC-AW catalysts,respectively,indicating the effectiveness of water-and acid-washing in removing K species in corncobs.

Thermogravimetric (TG) measurements in steam/Ar gas were performed to study the gasification reactivity of different Ni/BC catalysts.As shown in Fig.1,the major weight loss took place between 450 and 600°C for all three samples.As expected,with decreasing K amount,the temperature of the maximum weight loss rate of the three catalysts increased gradually,which were 511,521 and 541°C,respectively.This is consistent with the literature [26] that the presence of inherent K can improve the gasification reactivity of biochars.

3.2.Structural and surface chemical properties

In the X-ray diffraction (XRD) profiles of catalysts (Fig.2(a)),only the diffraction peaks of Ni were detected,indicating that the impregnated salts were decomposed and carbothermally reduced to metallic Ni nanoparticles during pyrolysis.The crystalline sizes were calculated with the Scherrer equation from the Ni(111)plane at 44.5° and shown in Table 1.Obviously,the crystalline size was related to the K content,i.e.,a higher K content resulted in a larger Ni crystalline size.Transmission electron microscopy(TEM)analysis in Fig.2(b)-(d) showed that the Ni nanoparticles were highly dispersed for all three catalysts,and the average particle size also increased with increasing K amount,which is consistent with the XRD results.Typically a more disordered carbon structure favors the dispersion of metal particles.However,as shown below in Raman analysis,Ni/BC-RAW with a more disordered structure unexpectedly had the largest Ni particle size.Therefore,there must be other underlying factors influencing the Ni particle size.It was reported that the Ni cations interacted strongly with the surface oxygenated groups of biomass during impregnation[27].We infer that the inherent K partially occupied the oxygenated groups of biomass,and reduced the available interaction sites for Ni cations,thereby leading to a larger Ni particle size.

The N2adsorption–desorption isotherms of different Ni/BC catalysts are shown in Fig.S1 in the Supplementary Material.At low relative pressures (P/P0),the considerable uptake of N2indicates the prevalence of micropores for all of the three Ni/BC catalysts.The Ni/BC-WW and Ni/BC-AW catalysts displayed type I isotherms,whereas Ni/BC-RAW catalysts showed a hysteresis loop,implying the existence of mesopores.The development of mesopores also resulted in the reduction of BET surface area of Ni/BC-RAW compared with the other two catalysts as shown in Table 1.

The Raman spectra in the range between 800 and 1800 cm-1were curve-fitted with 10 Gaussian bands(Table S1 in the Supplementary Material)representing the typical carbon structures found in chars from biomass[28],and the peak deconvolution profiles are shown in Fig.S2 in the Supplementary Material.As shown in Table S1 in the Supplementary Material,the G band at 1590 cm-1mainly represents the aromatic ring quadrant breathing and the graphitevibration.The D band at 1300 cm-1mainly represents the defect structures in ordered materials and aromatics with not less than six rings [28].(GR+VL+VR) bands broadly represent amorphous carbon structure with smaller aromatic ring systems of 3–5 fused benzene rings [29].A decrease in the ratio between the D and G band peak areas (ID/IG) is normally expected with increasing extent of graphitization[30],and a decrease in the ratio between the(GR+VL+VR)and D band peak areas(I(GR+VL+VR)/ID) reflects a more condensed carbon structure [31].

Fig.1.TG-DTG curves of Ni/BC catalysts with different K content in steam/Ar atmosphere:(a) Ni/BC-RAW,(b) Ni/BC-WW,(c) Ni/BC-AW.

Fig.2.(a)XRD patterns.TEM images of(b)4Ni/BC-RAW,(c)Ni/BC-WW,(d)Ni/BC-AW.Band are ratio of(e)I(GR+VL+VR)/ID and(f)ID/IG.XPS spectra of(g)C 1 s&K 2p and(h)Ni 2p.

Fig.2(e)and(f)shows theI(GR+VL+VR)/IDandID/IGratios of different Ni/BC catalysts,and they both decrease with the reduction of K content.This implies that the amorphous structure was selectively consumed and preferential formation of graphite structure was more significant for samples with less K amount during pyrolysis.A similar phenomenon was also observed by Xuet al.[32] that chars from sodium-loaded coals showed an increase in these ratios compared with those from acid-washed coals.It has been reported that alkaline metals inhibited the progress of graphitization of carbon,resulting in a more reactive char with more disordered crystalline carbon structure during pyrolysis [33].This should also be partially the reason that the temperature of the maximum weight loss rate of the three catalysts increased gradually with decreasing K amount in the TG measurements in steam/Ar gas.The inherent moisture in low-rank fuels (such as brown coal and biomass) and the pyrolytic water formed during pyrolysis of these fuels can lead to significantin situsteam gasification of the char produced during pyrolysis and also lead to changes in the char structure,which can affect the char reactivity considerably [29,34].Thus the Raman results are reasonable because for the Ni/BC-AW catalysts,in situsteam gasification takes place slowly throughout the biochar (on carbon active sites)to consume the smaller rings selectively,while for the Ni/BC-RAW catalysts,the gasification would be more focused or localized on the K catalytic sites so the activity of carbon active sites becomes less important.

According to previous studies [35–37],K addition can alter the electronic properties of the active metals.In order to find out whether K exerted electronic effects on Ni in the catalysts,X-rayphotoelectron spectroscopy (XPS) analysis was carried out.The XPS profiles of C 1s &K 2p and Ni 2p are shown in Fig.2(g) and(h).The C 1s spectra are deconvoluted into three peaks,and the C-C peak is fixed at 284.8 eV to calibrate the binding energies(BEs) of K and Ni.The other two peaks centered around 286 and 290 eV correspond to C-O and C=O bonds,respectively.The Ni 2p spectra have two spin-orbit components,i.e.,Ni 2p3/2and Ni 2p1/2.The Ni 2p3/2spectra are deconvoluted into three peaks,with BEs between 852.0–853.8 eV (Ni0),853.5–857.2 eV (Ni2+) and 855.2–864.1 eV (Ni2+satellite) [38].As shown in Fig.2(g),two K 2p peaks centered at 293.5 and 296.2 eV are assigned to K+species,and they are not detected in Ni/BC-WW and Ni/BC-AW catalysts due to the low K contents [38].As shown in Fig.2(h),compared with the other two catalysts with less K content,the Ni02p peak of the Ni/BC-RAW catalysts remain the same at 852.9 eV,revealing that the inherent K species had insignificant electronic effects on the active phase Ni.K exists originally in biomass during their growth,but Ni was impregnated during catalyst preparation.Therefore,during the pyrolysis preparation of catalysts,the inherent K species were highly dispersed in the bulk and on the surface of biochar supports,whereas Ni was mainly dispersed on the surface as small particles of 4–7 nm.This is evidenced by the presence of Ni peaks and the absence of K peaks on XRD profiles.It is reasonable to infer that K species may not have direct contact with the Ni particles,and the spacial gap between them limits the electron donation from K to Ni.The surface atomic composition of the catalysts are shown in Table 2.

Table 1Main physicochemical properties of fresh catalysts

As shows in Table 2,the increase of C/O ratio on Ni/BC with the decrease of K content suggested that a more condensed carbon structure was formed,which was consistent with the Raman results.

Table 2Surface elemental composition of Ni/BC catalysts with different K contents as determined by XPS

Fig.3.Light-off performance of different catalysts for toluene steam reforming.

3.3.Catalytic tests of toluene steam reforming

3.3.1.Light-off performance

The light-off curves of toluene steam reforming were measured to evaluate the catalyst performance in a temperature range of 350–800°C at a 1°C·min-1increment.A steam/carbon (S/C) ratio of 3.5 and a space time defined as the ratio between the mass of catalyst and the mass flow rate of toluene(W/F)of 0.5 h were used.Since biochar itself can catalyze tar reforming under certain conditions[39,40],therefore biochar from raw biomass with no Ni loading (BC-RAW) and biochar from acid-washed biomass with no Ni loading (BC-AW) were evaluated as well for comparison.In Fig.3,BC-RAW exhibited very limited catalytic activity,which reached only about 20%toluene conversion even at a high temperature of 800°C.The catalytic activity of BC-AW was further reduced.Nevertheless,on the Ni/BC-AW catalyst,toluene reforming activity was significantly improved as the reaction onset temperature was reduced to 525°C.Complete toluene conversion was achieved at 650°C and the activity was kept with increasing temperature to 800°C.This indicates that Ni nanoparticles were the major catalytic sites for toluene conversion on Ni/BC-AW catalysts.For Ni/BC-RAW and Ni/BC-WW catalysts with inherent K,the onset reaction temperature was further decreased to about 425°C.In addition,the K contents influenced the trends of their light-off curves.The increasing rate of the light-off curve of Ni/BC-RAW catalyst was higher than that of Ni/BC-WW catalyst,suggesting that higher K amount was beneficial for the steam reforming of toluene at low temperatures.However,the toluene conversion over Ni/BCRAW catalyst decreased rapidly above 530°C to zero at 638°C.A similar decreasing trend was also observed over Ni/BC-WW catalyst from 584 to 722°C.This means that,at high temperatures,higher K amount resulted in the faster gasification consumption of the biochar supports,since both Ni/BC-RAW and Ni/BC-WW catalysts disappeared after reactions.However,by further increasing the temperature to 730°C,toluene conversion started to increase on all catalysts except for Ni/BC-AW,which was probably due to the combination effects of thermal cracking and the residual Ni left in the reactor.To confirm this speculation,we carried out the lightoff run with only inert quartz sand(QS)in the catalyst bed and only 6% toluene conversion was reached even at a high temperature of 800 oC.This is consistent with our argument in the manuscript that thermal cracking can convert toluene to a certain extent at high temperatures,and the residual Ni and ashes after biochar gasification further enhanced the cracking process.

Fig.4.Catalytic performance under different temperatures:(a) 450°C,(b) 500°C,(c) 550°C,(d) 600°C.

3.3.2.Activity and stability of the catalysts

To further confirm our observations,we evaluated the steady state catalyst performance at constant temperatures of 450,500,550,and 600°C for 10 h.The evaluation results are shown in Fig.4.

Fig.4(a) and (b) show that the catalytic activity followed the order of Ni/BC-AW

3.4.Role of K in Ni/BC catalysts

Fig.5.MS signals of H2O-TPSR,(a) Ni/BC-AW,(b) Ni/BC-RAW,(c) BC-AW,(d) BC-RAW.

Regarding the role of K,as discussed above for XPS analysis,K did not change the electronic states of Ni.Thus,the effects of K in this study can not be assigned to the electron transfer as described in the literature[21,22].Moreover,XRD and TEM results demonstrate that the presence of K led to a larger Ni particle size,which may also affect the catalytic activity of Ni/BC.However,our previous work[19]and DFT calculations[41]showed that toluene decomposition is a structure-sensitive reaction,wherein smaller Ni particles displayed higher turnover frequency values.If the Ni particle size played a major role,then at 450 and 500°C,the order of catalytic activity should be Ni/BC-AW>Ni/BC-WW>Ni/BC-RAW,given the same Ni loading for all of the catalysts.However,the actual activity followed the opposite order,which excludes the possibility of Ni particle size effects.Therefore,it is speculated that the major role of K playing was enhancing water adsorption and reducing the condensation and graphitization degree of biochar support.

3.4.1.H2O-temperature programmed surface reaction (H2O-TPSR)experiments

To confirm our speculation on the role of K,H2O-TPSR was used to investigate the activity of the catalysts on H2O adsorption and Ni0oxidation,which is expressed according to equation Ni+H2O →NiO+H2.10% H2O/Ar flowed through the catalyst bed while ramping the temperature from 35 to 500°C at a rate of 10°-C·min-1,and the MS signals ofm/z=2,18,28 and 44 were recorded to track the evolution of H2,H2O,CO and CO2,respectively.Fig.5(a) and (b) show the profiles of gas evolution with increasing temperature over the Ni/BC-AW and Ni/BC-RAW catalysts.Since biochar supports can react with steam to generate hydrogen,H2O-TPSR was also performed for BC-AW and BC-RAW with no Ni loading.The results are shown in Fig.5(c) and (d),and due to the extremely low hydrogen intensity,the signals were magnified by 10 times for better visualization.Only above 350°C,the hydrogen intensities started to increase for BC-AW and BCRAW,indicating the occurrence of biochar gasification.To exclude the interference from biochar gasification,only the hydrogen signals below 350°C were integrated to evaluate Ni metal oxidation as shown in Fig.5a and b.The hydrogen area for Ni/BC-RAW is 1.4 times of that for Ni/BC-AW,suggesting a higher degree of Ni oxidation over Ni/BC-RAW.This observation verifies that the major role of inherent K was promoting the adsorption of water on the catalysts,due to the hygroscopic nature of the K species [42–44].

3.4.2.Catalyst characterization after use

The XRD patterns of the Ni/BC catalysts after 10-h runs at constant temperatures are shown in Fig.6(a)and(b).The peaks corresponding to NiO (JCPDS 73–1519) appeared for all of the spent catalysts,suggesting that Ni particles were partially oxidized by water during the reaction.This could be caused by the relatively high S/C ratio of 3.5 applied in this study,and similar phenomena were also observed on metal-oxide supported Ni catalysts [45,46].

At the lowest reaction temperature of 450°C,NiO peaks were more prominent than those for the spent catalysts obtained under higher temperatures.This is because that the reforming activity was too low to generate enough reducing gases and maintain Ni in the reduced form.At 500°C,a diffraction peak of graphite at 26.1owas found on Ni/BC-WW,but this peak did not exist for the other two catalysts.However,reasons for the absence of the graphite peak over Ni/BC-AW and Ni/BC-RAW are different.For Ni/BC-AW,the toluene conversion was below 20% as shown in Fig.4,and thus toluene decomposition did not occur significantly to generate large amount of coke.However,for Ni/BC-RAW,which displayed a higher conversion than Ni/BC-WW,the inherent K species enhanced water adsorption and led to faster coke removal before they condensed into graphitic structures.This is further corroborated by the XRD patterns of spent catalysts at 550°C,where the graphite peak intensity followed the order of Ni/BC-RAW

One might question the origin of the graphite peak,since it was reported that thermal annealing can lead to the condensation of the reactive and amorphous structures into larger and more inert ring systems[47].To clarify this,the Ni/BC-AW catalysts were subjected to different treatments as follows:(1) for Ni/BC-AW (Pyrolysis,10 h),the pyrolysis holding time was prolonged from 0.5 to 10 h;(2)for Ni/BC-AW(toluene,10 h),only toluene was fed during the catalytic test and all the other reaction conditions were kept the same as the one carried out at 600°C for 10 h.As shown in Fig.6(c),prolonging pyrolysis time only resulted in a Ni particle size increase from 4.9 to 5.2 nm,and there was no graphite diffraction peak.In comparison,the graphite peak was found when toluene was fed regardless of the presence of steam,which indicates that graphite structure was mainly derived from the decomposition of toluene.Furthermore,NiO peaks were not detected when solely toluene was fed,again reflecting that NiO was generated due to steam addition.These results prove our statement that steam eliminates coke deposition and the presence of inherent K can enhance this process by strengthening water adsorption.However,at temperatures of 550 and 600°C,the high water adsorption capability also led to the rapid consumption of the biochar supports and further the sharp decline of the catalytic activity.

4.Conclusions

Fig.6.XRD patterns of(a)spent Ni/BC catalysts at 450 and 500°C,(b)spent Ni/BC catalysts at 550 and 600°C,(c)Ni/BC catalysts subjected to various treatment conditions at 600°C.

The inherent K introduced mainly two aspects of effects to the Ni/BC catalysts.First,K enhanced water adsorption due to its hydroscopicity,resulting into a higher concentration of adsorbed water molecules around Ni particles.Second,K inhibited the condensation and graphitization process of biochar support during pyrolysis preparation,and this led to a higher gasification reactivity of Ni/BC catalysts.Under different temperatures,K brought about different results to the catalytic performance of Ni/BC catalysts in toluene steam reforming.At 450 and 500°C,toluene conversion rate was significantly increased in the presence of K,due to the strengthened water adsorption.When increasing the reaction temperature to 550 and 600°C,100%initial conversion of toluene was observed for Ni/BC catalysts with K.However,they experienced fast deactivation,since the more disordered biochar support was consumed more rapidly by the high amount of adsorbed water.In contrast,Ni/BC without K performed more stably at 550 and 600°C.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (2018YFB1501403),the National Natural Science Foundation of China (51776134 and 51406129)and the Technology Foundation for Selected Overseas Chinese Scholars,Ministry of Human Resources and Social Security of China(2016).

Supplementary Material

BET isotherms,Raman spectra deconvolution and band assignment.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.06.010.