A comparative study of the thermal and hydrothermal aging effect on Cu-SSZ-13 for the selective catalytic reduction of NOx with NH3

Huawang Zhao, Xiaomin Wu, Zhiwei Huang, Ziyi Chen, Guohua Jing

Department of Environmental Science & Engineering, College of Chemical Engineering, Huaqiao University, Xiamen 361021, China

Keywords:Emission control Selective catalytic reduction Cu-SSZ-13 Hydrothermal aging Thermal aging

ABSTRACT In this work, the characterizations of Cu-SSZ-13 after hydrothermal aging (HTA) and thermal aging (TA)at different temperatures (750, 800, and 850 °C) are compared, and the differences between those two serious aged samples are analyzed.With this data, the effect of steam on catalysts deactivation during hydrothermal aging is analyzed.The TA at 750 and 800 °C causes the dealumination and the agglomeration of Cu2+ions to CuO,resulting in the activity loss of Cu-SSZ-13.The presence of steam upon HTA at 750 and 800 °C aggravates the catalyst deactivation by increasing the Al detachment and the Cu2+agglomeration.The structure and cupric state are almost the same in the Cu-SSZ-13 after TA and HTA at 850 °C, respectively, indicating that the steam has little influence on the deactivation.The formation of CuAl2O4 spinel is the primary reason for the deactivation after both HTA and TA at 850 °C, probably attributed to the strong interaction between Cu2+ ions and framework Al sites at high temperatures.

1.Introduction

To meet ultimate zero emission standards, numerous efforts have been made to eliminate the NOxfrom diesel-powered engine[1-3].Selective catalytic reduction of NOxwith NH3(NH3-SCR) is one of the most efficient techniques [1,2].The Cu-CHA catalyst,i.e.,Cu-SSZ-13,can achieve 100%NOxconversion in a wide temperature range (200-500 °C), which has been commercialized in the past decade [4-6].In addition, as soot is often removed from the DPF at high temperatures (＞650 °C), which is located upstream of SCR catalysts.On the other hand, the moisture always exists in the vehicle exhaust [7,8].Therefore, high hydrothermal durability is required for the SCR catalyst to remain effective for NOxemission control.However, Cu-SSZ-13 suffers inevitable deactivation when hydrothermal aged above 750 °C, imposing a challenge to meet the life-time requirements for the commercial after-treatment catalysts [9-12].The deactivation mechanism of Cu-SSZ-13 after hydrothermal aging has been extensively investigated in previous studies, and the loss of Cu2+active sites as well as the destruction of the zeolite framework by dealumination are considered the two main reasons to the degradation [13-16].

The relationship between the dealumination and the agglomeration of Cu2+ions to CuO during hydrothermal aging (HTA) are widely discussed [4,10,17-19].In the initial investigation period,the dealumination is proposed to contribute to the aggregation of Cu2+ions to CuO, since the framework Al provides the ionexchange sites for Cu2+ions, [11,12,16,20].Then, Nam and coworkers [4]reported that high exchange levels in Cu-SSZ-13 catalysts led to lower hydrothermal stability at 850 °C, which is independent with the Si/Al ratio; they pointed that the excess of Cu2+ions would facilitated the formation and growth of CuO particles,which destroyed the structure of SSZ-13 and resulted in the deactivation of Cu-SSZ-13.Songet al.[10]stated a similar scheme that the agglomeration of Cu2+ions results in dealumination, and they further proposed that the hydrolysis of Cu2+ions by steam to Cu(OH)2intermediates accelerated the formation of CuO.Recently,Shanet al.[14]tested the hydrothermal stability of a serious Cu-SSZ-13 samples with same Si/Al ratio, but containing different Cu loading.They found that an appropriate amount of Cu loading(1%-3.8%, mass) reduces the Br?nsted acid sties content which are the most vulnerable sites to be attacked by steam, reducing the dealumination, while the formation of big CuO particles destroys the CHA structure when the Cu loading exceeds 4.8%(-mass) [14].

Based on the investigations above,it is apparent that both dealumination and aggregation of Cu2+ions are related to the presence of steam [10,21].However, the hydrothermal aging is normally performed above 750°C,where the thermal aging(TA)would also impose a big influence on the zeolite framework[22].The systematic investigation on the thermal aging effect on Cu-SSZ-13 is still lacking,but without which,the effect of steam on the deactivation of Cu-SSZ-13 during hydrothermal aging cannot be well understood.

This study aims to investigate the effect of steam on the deactivation of Cu-SSZ-13,by comparing the difference between samples after thermal and hydrothermal aging at the same temperature(750.800, and 850 °C), respectively.The deactivation schemes of Cu-SSZ-13 by thermal aging are given.The effect of the steam on the catalysts degradation during hydrothermal aging is discussed.

2.Experimental

2.1.Catalyst preparation

The Na-SSZ-13 was synthesized according to the procedure reported in our previous report [23].After synthesis, Na-SSZ-13 zeolite was calcined at 600 °C in air for 6 h to remove the organic template.

The Cu-exchanged catalysts were obtainedviaa two-step ionexchange.Firstly, the calcined Na-SSZ-13 was ion-exchanged in a 5 mol·L-1(NH4)2SO4solution for 2.5 h at 80 °C to obtain NH4-SSZ-13.2 g of NH4-SSZ-13 was dispersed in 200 ml 0.008 mol·L-1CuSO4solution.The pH of the copper solution was adjusted to 3-4 by HNO3.Subsequently, ion-exchange was processed at 80 °C for 1 h.After each step, the Cu exchanged samples were filtered,dried, and then calcined in static air at 600 °C for 6 h, obtaining the Cu-SSZ-13 samples.The Cu loading content and the Si/Al ratio are 2.38% (mass) and 6, respectively, as determined by ICP-OES.The Cu/Al ratio of Cu-SSZ-13 is 0.17.The Na content of Cu-SSZ-13 is determined by ICP-OES, which is around 0.01% (mass).The low content indicates that the Na has little influence on the hydrothermal stability of Cu-SSZ-13.

The hydrothermal aging (HTA) was performed in synthetic air containing 10% H2O at 750, 800, and 850 °C for 16 h on the fresh samples, respectively.The thermal aging (TA) was carried out in muffle furnace at 750, 800, and 850 °C, respectively.In the text,the fresh sample is named as Cu-SSZ-13,and the HTA(TA)samples are named as Cu-HTA (TA)-X, where X represents the HTA (TA)temperature.

2.2.NH3-SCR performance test

The activities of the catalysts for NH3-SCR were tested in a fixed-bed quartz reactor (6 mm in diameter) in steady flow operation.50 mg of the catalyst diluted by 300 mg of quartz sands(calcined at 800 °C for 6 h before usage) was loaded in the reactor.Before the test, the catalysts were treated at 600 °C in 5% O2/N2flow for 1 h.The gas flow was kept at 600 ml·min-1during the reaction, corresponding to the gas hourly space velocity of 360000 h-1.The reactant feed contains 0.05% NO, 0.05% NH3,6% (vol) H2O, 5% (vol) O2, and N2as balance.For each run of testing, the reaction temperature was increased from 150 °C to 550 °C stepwise in every 50 °C.At each step, the concentration of the products (NO and NO2) was monitored online by Testo 350.The NOx(NOx= NO + NO2) conversion was calculated with Eq.(1):

2.3.Characterization

The Cu,Si,Al contents of the samples were determinedviaICPOES (VISTA-MPX, Varian).The N2physisorption was carried out at-196 °C on a Micromeritics ASAP 2020 instrument to determine the specific surface area and pore volume of the catalysts.Prior to the analysis, the sample was pre-degassed in vacuum at 300 °C for 6 h.The crystallization structure was measuredviaXRD (Rigaku, D/max-γ b-type X-ray) using monochromatic Cu Kα radiation (40 kV and 100 mA), with a scan speed of 2 (°)·min-1.

Thermo-gravimetric analysis (TGA) was performed on Mettler-Toledo TGA2 in a flow of air (50 ml·min-1), and the heating rate was 10°C·min-1from 100 to 1000°C.All the samples were treated in a flow of air (50 ml·min-1) at 100 °C for 1 h before the tests.

27Al MAS NMR were conducted on Varian Infinity plus 300 WB spectrometer utilizing a 4 mm triple resonance probe operating with the resonance frequencies of 78.13 MHz.Spectra of27Al NMR were acquired by using calibrated27Al π/20 pulses of 0.5 us, a 40 kHz spectral window, a spinning speed of 8 kHz, and a 3 s pulse delay.Al(NO3)3aqueous solution (1 mol·L-1) was used for27Al MAS NMR spectroscopy as references.The amount and volume of the samples in NMR test were kept constant, thus the intensities of the resonances can be compared.For the27Al spectra,the resonance peaks were deconvolutedviaLorentzian/Gaussian,with the peak position not fixed for the best fit.

EPR spectra in the X-band were recorded with a CW spectrometer JES-FA200,with a microwave power of 1 mW modulation frequency of 100 kHz.The EPR signal of isolated Cu2+ions were recorded at -150 °C with the magnetic field being swept from 2000 to 4000 G with a sweep time of 5 min.

The H2-TPR measurements were performed on an AutoChem 2920 apparatus with a TCD detector.75 mg of sample was pretreated at 550°C in the 5%O2/N2for 1 h before the test.After cooling down to room temperature,the sample was exposed to 5%H2/N2, with the temperature increasing to 1000 °C at a rate of 10 °C·min-1.

Transmission electron microscopy (TEM) was performed on Tecnai G2 20 at an acceleration voltage of 200 kV.The samples were primarily dispersed in ethanol and followed by ultrasonication.

3.Results

3.1.SCR performance prior to and post HTA and TA treatment

Fig.1. NOx conversion during standard NH3-SCR as a function of the temperature(150-550 °C)of the catalysts.Reaction conditions: 0.05% NO, 0.05% NH3, 5% O2, 6%H2O, balanced with N2; flow rate: 600 ml·min-1, GHSV: 360,000 h-1.

Fig.1 shows the SCR activities of the fresh,HTA,and TA samples.The Cu-SSZ-13 exhibited excellent SCR performance, showing the NOxconversion of around 45%at 200°C and keeping 100%conversion in 250-550 °C.The deNOxactivities of the HTA samples decrease with the increase of HTA temperature.The NOxconversion remains in 80%-90% in 300-550 °C for Cu-HTA-750 sample,while decreases to 65%-80% for Cu-HTA-800 sample in the same temperature range.After HTA at 850 °C, the Cu-SSZ-13 suffers a serious deactivation, showing the maximum NOxconversion of 15% at 350 °C.

The Cu-SSZ-13 also suffers degradations after TA at 750 °C and 800 °C, while the TA samples show around 10%-15% higher NOxconversion than the counter HTA samples.Besides, similar to the Cu-HTA-850,the Cu-TA-850 sample showed a severe deactivation,with the maximum NOxconversion of around 15%at 350°C.It indicates that both the HTA and TA result in the deactivation of Cu-SSZ-13, while the HTA leads to a more severe degradation.

3.2.Characterizations of the structure prior to and post HTA and TA treatment

3.2.1.N2-physisorption and XRD results

To investigate the effect of HTA on the pore structure of the Cu-SSZ-13 catalysts, the N2-physisorption measurement was carried out, and the results are shown in Table 1.The Cu-SSZ-13 has a BET surface area of 482 m2·g-1and a pore volume of 0.26 cm3·g-1.90% of the pore volume is contributed by micropores.Different declining extent of micropore volume was observed after HTA at 750(38%),800(50%),and 850°C(83%)compared to the fresh sample, demonstrating the degradation of the pore system upon HTA.The micropore volume also decreases with the increment of TA temperature, with 29%, 42%, and 83% loss after TA at 750 °C,800 °C and 850 °C, respectively, as compared to Cu-SSZ-13.The more severe micropore volume loss in samples after HTA at 750 °C and 800 °C than counter TA samples implies that the presence of steam accelerated the degradation of the structure of SSZ-13.The significant loss of micropore volume in Cu-HTA-850 and Cu-TA-850 samples suggests that the structure is severely damaged in both samples.

Table 1SBET, Smicro, Spore, Smeso, and Vmicro for fresh and aged Cu-SSZ-13 samples

The XRD patterns of Cu-SSZ-13,HTA,and TA samples are shown in Fig.2.The Cu-SSZ-13 showed typical CHA diffraction peaks[24],while the intensities of the peaks decrease with the aging temperature increasing from 750 °C to 800 °C, implying the structure degradation of the HTA samples.The diffraction peaks disappeared after HTA at 850°C,suggesting that the structure is completely collapsed.Also,compared to Cu-SSZ-13,the intensities of the diffraction peaks decrease for Cu-TA-750 and Cu-TA-800 samples.After TA at 850 °C, the Cu-SSZ-13 changes to amorphous phase, since no diffraction peaks were found.Those observations suggest that the TA causes the structure damage of Cu-SSZ-13.

Fig.2. XRD patterns of the fresh, HTA, and TA catalyst samples.

3.2.2.TGA measurements

The TGA technique is able to record the mass changes of the tested samples with the temperature increasing,which would provide more information of the structure damage of Cu-SSZ-13 during TA.

Fig.3(a) presents the TGA results of H-SSZ-13 conducted in air.An apparent weight loss in 700-850°C with a DTG peaks centered at 782 °C was observed, assigned to the dealumination in zeolite(Fig.4, see below) upon TA at high temperatures [25,26].It indicates that the dealumination results in the structure degradation of H-SSZ-13 above 700 °C.To further examine the universality of this phenomenon, the TGA measurement on another NH4-SSZ-13,with a higher Si/Al ratio (13) purchased from ACS material, was also carried out (Fig.S1 in Supplementary Material).An apparent weight loss in the high temperature ranges(750-1000°C)was also observed in this commercial material.The above observations suggest that the structure degradation by dealumination upon TA is an intrinsic feature of SSZ-13.

Also, the weight loss in 700-850 °C due to the dealumination was found in Cu-SSZ-13, shown in Fig.3(b).In addition, another weight loss peak above 850°C was also observed.This peak should be related to Cu species in Cu-SSZ-13, revealing that the presence of Cu induces another weight loss upon TA.

3.2.3.27Al MAS NMR results

27Al MAS NMR is highly sensitive to the subtle changes in the local Al environment of the catalysts after aging [27-29].Therefore, this technique was used to further characterize the variation of framework Al after HTA and TA at different temperatures.

Fig.4 shows the27Al MAS NMR results of fresh, HTA, and TA samples.Cu-SSZ-13 showed two resonance peaks at 59 and 0,arising from the tetrahedrally coordinated framework aluminum(TFAl) and extra-framework aluminum (EFAl), respectively[29,30].A peak at 32.5 was also observed after HTA and TA at 750 and 800 °C, which is assigned to the pentahedral Al [21].

Fig.3. TGA results of the H-SSZ-13 (a) and Cu-SSZ-13 (b) conducted in air.

The intensity of peak at 59 decreases after HTA and TA at 750°C and 800 °C, which should be attributed to the dealumination.To acquire the quantitative degrees of the dealumination after HTA and TA,the TFAl signal in each sample was integrated and normalized,using the Cu-SSZ-13 as a reference.It shows that 78%and 85%of TFAl remain in Cu-HTA-750 and Cu-TA-750 samples, respectively.Also, a higher content of TFAl is maintained in Cu-TA-800(76%) than Cu-HTA-800 (66%).It indicates that the presence of steam aggravates the dealumination during HTA.

The resonance peak at 59 remains in Cu-HTA-850 and Cu-TA-850, while the peak width is widened, compared to the samples aged at 750 °C and 800 °C.Such broadening illustrates that the Al sites are in a multiple,less uniform bonding environments,suggesting that the Cu-SSZ-13 underwent significant structure changes upon aging at 850°C[31].The XRD results in Fig.2 shows that Cu-SSZ-13 is degraded to amorphous phase after HTA and TA at 850 °C, indicating the long-range order of CHA structure, constructed by Si and Al, is completely collapsed.Consequently, the broad27Al NMR peak at 59 implies the formation of short-range ordered Al after HTA and TA at 850 °C.

3.3.Characterizations of the copper sites prior to and post HTA and TA treatment

3.3.1.Electron paramagnetic resonance (EPR) results

The isolated Cu2+ions are EPR active, while some other copper species,e.g.,CuO, are EPR silent [10,32].Therefore, the changes of Cu2+ions content upon aging can be probed by EPR measurements.The hydrated Cu-SSZ-13 was used in EPR measurement to determine the changes of total content of Cu2+ions in Cu-SSZ-13 samples after hydrothermal/thermal treatment.

The EPR spectra for the samples are given in Fig.5(a).The signal line-shapes do not change upon HTA and TA at 750 °C and 800°C,compared to Cu-SSZ-13, which give the same g value of 2.060,showing that the Cu2+ions are in a similar environment [5,33].In addition, the intensities of the EPR signal peaks decrease with the increasing HTA and TA temperature.Since the intensity of EPR signal represents the Cu2+ions content in each sample[21,34],the isolated Cu2+ions content is normalized by comparing the EPR intensity of each sample,using the Cu-SSZ-13 as reference,as shown in Fig.5(b).It shows that 80% and 62% of isolated Cu2+ions remain in Cu-HTA-750 and Cu-HTA-800,respectively.In comparison, 85% and 70% are maintained in Cu-SSZ-13 after TA at 750 °C and 800 °C, respectively.

Cu-HTA-850 and Cu-TA-850 also give EPR signals, while shows totally different line-shapes compared to Cu-SSZ-13.Another g factor at 2.031 was observed in both samples(Fig.5(a)),indicating the formation of CuAl2O4upon treated at 850 °C [35].The content of Cu2+ions in Cu-HTA-850 and Cu-TA-850 were not given based on the EPR results, as both Cu2+and CuAl2O4give EPR signals [35].

3.3.2.Temperature-programmed reduction by hydrogen (H2-TPR)results

The EPR (Fig.5) results demonstrate the transformation of isolated Cu2+to other copper species,e.g.,CuO, after HTA and TA.H2-TPR measurements can distinguish those different cupric species, in terms of their different reduction temperatures, providing clues of the changes of isolated Cu2+ions during HTA and TA[4,12,36].The results are shown in Fig.6.

Fig.4. Solid state 27Al MAS NMR spectra of the fresh, HTA, and TA catalysts.The amounts of TFAl in the fresh and aged samples were normalized and shown above each spectrum, using the TFAl signal area of the Cu-SSZ-13 sample as a reference.

Fig.5. EPR spectra(a)and the normalized Cu2+ions concentration(b)of the fresh and aged catalyst samples.The Cu2+ions concentration in each sample was normalized by double integrating the EPR spectra, using the isolated Cu2+ ions content in Cu-SSZ-13 as the reference.

Fig.6. (a) H2-TPR spectra of fresh,HTA, and TA samples.(b) Evolutions of different types of Cu ions with aging temperature deduced from deconvolution of H2-TPR curves.The content of different types of Cu ions in each sample are normalized, using the Cu2+-2Al and Cu(OH)+ content in Cu-SSZ-13 as reference, respectively.

The H2-TPR profile of Cu-SSZ-13 shows reduction peaks in different temperature regimes,i.e., 250-450 and 780-920°C, respectively, as shown in Fig.6(a).The peak below 500 °C should be attributed to the reduction of Cu2+to Cu+(Cu2++?H2=Cu++H2O)and CuO(if present)to Cu0(CuO+H2=Cu0+H2O);the peak above 500 C is ascribed to the reduction of Cu+to Cu0(Cu++?H2=Cu0+H2-O)[23,37].The equal H2consumption areas of the peaks below and above 500 °C(Fig.S5) indicates no CuO existing in Cu-SSZ-13,and the cupric sites are exclusively isolated Cu2+ions.Furthermore,the peak below 500 °C can be divided into two peaks centered at around 275 °C and 380 °C, respectively.The peak at 275 °C can be assigned to the reduction of Cu(OH)+and the peak at 380 °C is attributed to Cu2+-2Al, since the Cu2+-2Al has a higher stability than Cu(OH)+[38-40].The area of the peak at 380°C is larger than that at 275°C,demonstrating that the Cu2+-2Al is dominate in Cu-SSZ-13.

After HTA at 750°C and 800°C,the reduction peak below 500°C is widened to 200-450°C,which can be divided into three peaks at 230 °C, 275 °C, and 356 °C, respectively.The unequal area of the peak below and above 500 °C indicates the formation of CuO after HTA (Fig.S5).Therefore, the peak at 230 °C is assigned to the reduction of CuO clusters [23,30].The evolution of the Cu2+-2Al and Cu(OH)+with the increase of HTA temperature are shown in Fig.6(b).52% and 69% of Cu(OH)+decreases after HTA at 750 °C and 800 °C, respectively, while Cu2+-Al only shows minor decline(around 5% and 23% after HTA at 750 °C and 800 °C, respectively),as compared to Cu-SSZ-13,consistent with the NH3-DRIFTs results(Fig.S4).It indicates that the Cu(OH)+suffers a more severe agglomeration than Cu2+-2Al upon HTA at 750°C and 800°C.Upon HTA at 850°C,only two peaks in 200-300°C and 400-550°C were observed.The first peak can be attributed to the reduction of CuO[23,30].The second peak should be associated with the reduction of CuAl2O4(Fig.6(a)), consistent with the works by Maet al.[41]and Kwaket al.[42], who reported that the reduction of CuAl2O4occurs in 460-550 °C and around 560 °C, respectively.

Two reduction peaks below and above 500°C were observed in Cu-SSZ-13 after TA at 750 and 800 °C.The peak below 500 °C is broadened after TA, which can be divided into three peaks at 230 °C, 275 °C, and 356 °C, respectively.The appearance of the reduction peak at 230°C suggests that the Cu2+ions agglomeration to CuO also occurs during TA.Similar to the HTA samples,Cu(OH)+show a more severe loss than Cu2+-2Al after TA(Fig.6(b)),indicating the lower TA stability of Cu(OH)+than Cu2+-2Al.It should be pointed that less decline of Cu(OH)+and Cu2+-2Al was observed upon TA at 750 and 800°C than upon HTA at counter temperatures,indicating that the presence of steam facilitates the agglomeration of both types of Cu2+ions.

Cu-TA-850 and Cu-HTA-850 display simar H2-TPR profiles,showing two reduction peaks in 200-300 °C and 400-550 °C,assigned to the reduction of CuO and CuAl2O4, respectively[23,30,41].It suggests that the steam might only has little influence on the formation of CuO and CuAl2O4upon HTA at 850 °C.

3.3.3.TEM images

H2-TPR results in Fig.6 illustrate that CuO particles are formed in Cu-SSZ-13 after HTA and TA at 750 and 800 °C, while both CuO and CuAl2O4are formed upon treated at 850 °C.The size,location,and distribution of those particles are visible in transmission electron microscope (TEM) images, which are shown in Figs.7 and 8.Also, the particles size distribution of each sample is given in the right of the counter image.

No particles were observed in the Cu-SSZ-13 in Fig.7(a),because of the cupric ions are atomically dispersed (Fig.6).After the HTA at 750°C and 800°C,CuO nanoparticles are formed in both samples, with a mean size of around 3.2 and 3.5 nm, respectively.Uniformly dispersed nanoparticles are also found in Cu-HTA-850,while the particle distribution can be distinguished by two regions,peaking at around 3.8 nm and 8.5 nm,respectively.The Al inserted into CuO crystal to form CuAl2O4would enlarge the size of CuO particles; thus, the large particles with the size of around 8.5 nm should be CuAl2O4.

The TA also causes the formation of CuO nanoparticles, which can be clearly observed in the Cu-TA-750 and Cu-TA-850 samples in Fig.8,with a mean size of 2.2 and 2.4 nm,respectively.The particles in Cu-TA-850 are distributed in two ranges,i.e., 2-4 and 6-9 nm, ascribed to the CuO and CuAl2O4particles, respectively.

4.Discussion

4.1.Deactivation of Cu-SSZ-13 by TA

The Cu-SSZ-13 suffers activities loss upon TA (Fig.1).Since the isolated Cu2+ions are the active sites catalyzing SCR reactions[43-45],the decline of the activity should be related to the transformation of Cu2+ions to other SCR inactive sites,i.e., CuO and CuAl2O4,upon TA,as shown in EPR(Fig.5),H2-TPR(Fig.6),and TEM(Fig.7)results.Only CuO phase is formed in the Cu-TA-750 and Cu-TA-800 samples, while both CuO and CuAl2O4are found in Cu-TA-850.It indicates that the agglomeration of Cu2+ions in Cu-SSZ-13 might follow similar pathway upon TA at 750°C and 800°C,but different from that upon aging at 850 °C.

The TEM results show that the mean particles size of CuO in Cu-TA-750 and Cu-TA-800 samples is 2.2 and 2.4 nm, respectively.Based on calculation result (details shown in SI), the two kinds of CuO particles consist of 200 and 260 of Cu atoms, respectively.Besides, the cupric sites are atomically dispersed in the fresh Cu-SSZ-13 (Fig.8(a)), and each unit CHA cell contain 0.82 of Cu atom(details of calculation shown in SI).Therefore, the formation of a CuO particle with a size of around 2.2-2.4 nm requires the aggregation of copper ions locating in the 240-320 different cages in Cu-SSZ-13.It indicates that the cupric species must be moveable upon TA.Generally,the growth of metal nanoparticles due to the movement of metal atoms upon TA can be understood in terms of two operative mechanisms: Ostwald ripening (OR) or particle migration and coalescence(PMC)[46,47].Ripening involves interparticle transportation of mobile species, with larger particles growing at the expense of smaller particles due to differences in surface energy [46].Particle migration involves the Brownian motion of nanoparticles leading to coalescence when particles come in close proximity to each other[46].In our case, all the cupric sites are in an atomic level in Cu-SSZ-13 (Fig.8).Besides, the CuO nanoparticles size is in the same ranges in TA samples (Fig.9).Therefore,the formation of CuO nanoparticles is more likely following the PMC pathway,i.e., the moveable cupric ions, driven by Brownian motion by heating, are collided and then agglomerated to CuO particles.

15 and 30% of isolated Cu2+ions are agglomerated to CuO after TA at 750 °C and 800 °C, respectively (Fig.5), most of which are derived from the aggregation of Cu(OH)+(Fig.6(b)).Based on the argument above, the formation of CuO should be primarily attributed to the collision of moveable Cu(OH)+ions following the Eq.(2).

Besides, there are still some Cu2+-2Al ions that aggravated to CuO upon TA at 750°C and 800°C.It is known that the high binding energy between the Cu2+ions and TFAl render the Cu2+-2Al an extremely high stability [10].Thus, the agglomeration of Cu2+-2Al ions should be related to the dealumination (Fig.4) [23], which leads the Cu2+-2Al to become instable.

The structure of Cu-SSZ-13 changes to amorphous phase upon TA at 850 °C.Besides, the TFAl content shows a sharp decline(66% to 0%) with the increment of TA temperature from 800 °C to 850 °C, accompanied with the formation of CuAl2O4(Fig.2 and Fig.7).It seems that the interaction of Cu2+ions and TFAl forming CuAl2O4increases dealumination, resulting in the structure collapse.The comparison of27Al NMR results of H-SSZ-13 and Cu-SSZ-13 after TA at 850 in Fig.S6 lend a further support to this hypothesis, which shows that large amount of TFAl still remains in H-SSZ-13 after TA at 850°C,but no framework Al was observed in Cu-TA-850.Actually,it has been widely reported that the metal loading facilitates the collapse of zeolite, such as Zn/zeolite-A[48,49], Co/zeolite-A [48], Na/zeolite-A [50], Mg/zeolite-A [51],and Mn/zeolite-B [52], where the metal atoms and zeolite recrystallize to form high-density ceramics or spinel phase upon TA at high temperature.It is proposed that the TFAl stay in a strongly disordered geometry at extra high temperature; the metal ions would act as nuclei to draw the TFAl to form binary spinel, resulting in the dealumination [49].In this sense, it can be concluded that the affinity of Cu2+ions and TFAl accelerates the detachment of TFAlviaformation of CuAl2O4upon TA at 850 °C, resulting in the collapse of the CHA structure and causing the compete deactivation of Cu-SSZ-13.

Fig.7. The TEM images of the fresh(a),HTA-750(b),HTA-800(c),HTA-850(d)samples.The CuO size distribution of the fresh and HTA Cu-SSZ-13 were shown in the right of counter image, with at least 150 particles were counted.The scale of Cu-HTA-750 and Cu-HTA-850 images is 100 nm, which is 50 nm for Cu-HTA-800.

4.2.The effect of steam on the deactivation of Cu-SSZ-13

The Cu-HTA-750 and Cu-HTA-800 samples exhibit around 10%-15%lower SCR activities than the counter TA samples(Fig.1),illustrating that the presence of steam accelerates the deactivation of Cu-SSZ-13.Again, since the Cu2+ions are the catalyzing site for SCR reactions [5,33,53-55], the more severe loss of the activity for HTA samples should be related to the larger content of Cu2+ions agglomerated to CuO, compared to the TA samples (Fig.5).

Fig.8. The TEM images of the TA-750(a),TA-800(b),and TA-850(c)samples.The CuO size distribution of the fresh and HTA Cu-SSZ-13 were shown in the right of counter image, with at least 150 particles were counted.The scale of images is 50 nm.

Fig.9. Schemes for Cu ions agglomeration during TA (a) and HTA (b) at 750 and 800 °C.

The size of uniformly dispersed CuO nanoparticles in Cu-SSZ-13 HTA at 750 °C and 800 °C are around 3.2 and 3.5 nm, containing around 615 and 805 of Cu atoms, much larger than that in the counter TA samples (size of 2.2 and 2.4 nm, containing 200 and 260 of atoms,respectively).It indicates that the presence of steam improves the mobility of Cu2+ions, leading to a more severe collision and aggregation of Cu2+ions to CuO.This can be explained by two aspects.The first is that the dealumination during HTA is aggravated by steamviahydrolytic severance of bridging oxygens in Br?nsted acid sites (Al-OH-Si) which are vulnerable to H2O attack [9,14], as confirmed by27Al MAS NMR results in Fig.4.It renders more Cu2+-2Al loss the bondage provided by TFAl and become moveable, facilitating the agglomeration (Fig.6(b)).On the other hand, the Cu2+ions would be hydrolyzed by steam to Cu(OH)2intermediates [10], which own extra mobility other than the Brownian movement provided by heating.This proposal is shown in Fig.9.

Cu-TA-850 and Cu-HTA-850 shows simar SCR activity (Fig.1).Also, the framework integrity and the state of Cu in both samples are almost identical (Fig.2, Fig.5, and Fig.6), indicating that the deactivation of Cu-SSZ-13 upon treated at 850°C is mainly caused by TA.Therefore,the deactivation of Cu-SSZ-13 upon HTA at 850°C should be primarily attributed to the formation of CuAl2O4by TA,and the steam might have little influence.

5.Conclusions

The effect HTA and TA on Cu-SSZ-13 is compared.Upon TA at 750 °C and 800 °C, the dealumination and agglomeration of Cu2+to CuO result in the activity loss of Cu-SSZ-13.The formation of CuO by TA is mainly attributed to the collision and agglomeration of the Cu2+ions, probably driven by Brownian motion.The presence of steam in the feed aggravates the Cu2+ions agglomeration upon HTA at 750 °C and 800°C, which might be attributed to that the steam accelerates dealumination and hydrolysis of Cu2+ions to Cu(OH)2intermediates.The similar characteristic features of Cu-HTA-850 and Cu-TA-850 samples demonstrates that the catalyst deactivation upon treatment at 850 °C is primary caused by TA.The formation of CuAl2O4is believed accelerating the structure collapse as well as the resultant deactivation of Cu-SSZ-13 upon HTA and TA at 850 °C.

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 Key Research and Development Program of China (2018YFC0214103), the National Natural Science Foundation of China(22006044),and the Scientific Research Funds of Huaqiao University (605-50Y200270001).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.04.039.