Enhancement of catalytic activity in NH3-SCR reaction by promoting dispersibility of CuCe/TiO2-ZrO2 with ultrasonic treatment.

A series of CuCe-modified TiO2-ZrO2 catalysts synthesized by stepwise impregnation method and ultrasonic-assisted impregnation method were investigated to research the removal of NO in the simulated flue gas. Results showed that the CuCe/TiO2-ZrO2 catalyst prepared by ultrasonic-assisted impregnation method exhibited the superior NO conversion, in which higher than 85% NO was degraded at the temperature range of 250-400 °C and the highest NO conversion of 94% at 350 °C. It proves that ultrasonic treatment can markedly improve the performance of catalysts. The effect of ultrasonic enhancement on CuCe/TiO2-ZrO2 was comprehensively studied through being characterized by physicochemical characterization. Results reveal that the ultrasonic cavitation effect improves the distribution of active species and the synergistic interaction between Cu with Ce components (Cu+ + Ce4+ ↔ Cu2+ + Ce3+) on the catalysts significantly, thus resulting in better dispersibility as well as a higher ratio of Cu2+ and Ce3+ of the catalysts. Moreover, it was found that the CuCe/TiO2-ZrO2 catalyst prepared by the ultrasonic-assisted impregnation method represented a higher degree of ultrafine metal particles and evenness. The above results were described with the generalized dimension and singularity spectra in multifractal analysis and validated by the comparative test. Therefore, it can be concluded that ultrasonic treatment facilitates the particle size and distribution of active sites on the catalysts.

Introduction

In recent years, air pollution has attracted the increasing attention of society. It is well known that nitrogen oxide (NOx) has caused plenty of environmental problems, such as photochemical smog, ozone depletion, acid rain, and the greenhouse effect [1]. Hence, NOx emission needs to be limited nowadays, and selective catalytic reduction by NH3 (NH3-SCR) is proven to be the most widely-applied and efficacious denitration technology [2].

The multi-component interaction of multi-metallic catalysts is effective to improve the catalytic performance, bringing about many NH3-SCR catalysts consisted of various transition metal oxides or/and lanthanide metal oxides on different supports. As an important component of the commercial deNOx catalysts, TiO2 has multiple acidic sites and commendable SO2 resistance performance [3], but still confined for its high cost, low surface area, and low resistance to sintering [4]. ZrO2 is one of the most effective and commonly used dopants due to its low price and promotion role in surface area, thermal stability and catalytic performance [4], [5], it can effectually make up the defects of TiO2 catalyst in NH3-SCR reaction. Besides, the TiO2-ZrO2 binary oxide has been reported to exhibit multiple surface acidity by charge imbalance based on the generation of Ti-O-Zr bonding [6]. And TiO2-ZrO2 is an effective acid-base bifunctional catalyst, the formation of ZrTiO4 crystal as well as the amount of acid and base sites are the major factors affecting the catalytic activity [7]. Therefore, TiO2-ZrO2 support would be a better alternative to replace TiO2, and its surface area reaches a maximum when the molar ratio of TiO2 to ZrO2 is 1:1 [8], [9]. The redox ability of the active site is necessary to NH3-SCR reaction, so it is significant to select active site on catalysts. Because of the features such as low cost, non-toxicity, and high low-temperature activity, copper-based catalysts have been widely explored [10]. Furthermore, it has been confirmed that CeO2 plays a crucial role in the elimination of NOx, because it has favorable storage oxygen capacity and preferable redox performance associated with the electron transportation between Ce3+ and Ce4+ [11]. Besides, the existence of synergistic interaction between Cu with Ce species contributes to enhancing the catalytic performance of NH3-SCR catalysts [12]. Therefore, CuCe-modified TiO2-ZrO2 catalyst probably behaves better thermal stability, anti-toxicity, and catalytic property in NH3-SCR reaction.

The main factors affecting the stability and activity of Cu-based and Ce-based catalysts are the grain size and dispersibility degree of the active ingredient. It has been reported that the physical and chemical effects of ultrasonic cavitation can effectively control the growth of crystals, improve the dispersion of active species on the support and promote the interaction of metals with support [13], [14], [15], so the catalyst dealt with ultrasound will possess smaller average particle size, higher dispersion of active ingredients and stronger metal-support interaction. The catalysts dispersed with metal oxide crystals of various particle sizes promoted by ultrasound treatment need to be further explored by a depth theoretical analysis of experimental data to enhance the catalytic performance. According to the previous literatures [16], [17], [18], the particle size and distribution of active sites on the catalyst surface can be better described by the multifractal method with the generalized dimension and singularity spectra.

In the present study, a series of CuCe-modified TiO2-ZrO2 catalysts were prepared by stepwise impregnation method (SI) and ultrasonic-assisted impregnation method (UI). The research assessed the efficiency of different catalysts on NO removal in the simulated flue gas. The physicochemical properties of the catalysts were characterized through some essential characterizations, including FE-SEM, EDX, AFM, XRD, BET, FT-IR, XPS and TGA. What is more, the particle size and distribution of active sites on different catalysts were analyzed in the light of multifractal theory.

Experimental section

Catalyst preparation

The TiO2-ZrO2 (molar ratio of 1:1) supports were synthesized by the sol–gel method as follows. Firstly, an amount of tetrabutyl titanate was slowly added to ethanol absolute, and the mixed solution was named solution A. The desired amount of zirconyl chloride was thoroughly dissolved in ethanol absolute, and the mixed solution was called solution B. The deionized water, acetic acid and ethanol absolute (2:2:1 vol ratio) were added to the beaker successively to obtain solution C. Secondly, solution A was added dropwise into solution B and followed by dropwise addition of solution C with magnetic stirring at room temperature for 20 min until transparent sol was formed. Thirdly, the mixed solution was at 50 °C in a water bath (HH-S24S, Jintan Dadi Automation Instrument Factory, China) for 1 h, then it was aged at room temperature for 12 h till the hydrosol became transparent gel. Finally, the gel was dried at the oven (FN101-2SB, Changsha Lianshen Constant Temperature Instrument Factory, China), the temperature was increased from room temperature to 80 °C by 5 °C/min and incubated at 80 °C for 5 h, then further improved from 80 °C to 120 °C with same heating ramp rate and kept at 120 °C overnight. The obtained product was ground and calcined in a muffle furnace (SX-4–10, Changsha Zhonghua Electric Furnace Factory, China), the temperature was increased from room temperature to 500 °C by 10 °C/min and kept at 500 °C for 3 h.

The CuCe/TiO2-ZrO2 catalysts were prepared from stepwise impregnation method (SI) and ultrasonic-assisted impregnation method (UI), loaded the same amount of metal, and the loading quantity of copper and cerium were 20 wt% and 10 wt% for the support, respectively. In detail, a certain quantity of Cu(NO3)2·3H2O was dissolved in deionized water, and a required amount of TiO2-ZrO2 was added into the above solution and mixed, then the obtained solution at 80 °C under magnetic stirring for 1 h. Afterwards, the obtained product was aged at room temperature for 12 h, then dried at 105 °C overnight in an oven, and finally calcined at 450 °C for 4 h in the muffle furnace. The heating ramp rates were 5 °C/min and 10 °C/min, respectively. The above catalyst prepared using the stepwise impregnation method was denoted as Cu/TiO2-ZrO2-SI, and CuCe/TiO2-ZrO2-SI catalyst was prepared with the same method. The comparative CuCe/TiO2-ZrO2-UI catalyst prepared through ultrasonic-assisted impregnation method and steps as follows. First of all, the solution containing Cu(NO3)2·3H2O and TiO2-ZrO2 was prepared in advance, then mixed in a magnetic stirrer (79–1, Jintan Dadi Automation Instrument Factory, China) at 80 °C for 40 min and reprocessed in an ultrasonic cleaner (GTSONIC-D9, Guangdong Gute ultrasonic co., Ltd, China). Ultrasound with the frequency of 28 kHz and power of 240 W was used to treat samples at 20 °C for 20 min. Next, the above sample was aged at room temperature for 12 h, then dried at 105 °C overnight in an oven and calcined at 450 °C for 4 h in the muffle furnace. The heating ramp rates were 5 °C/min and 10 °C/min, respectively. Finally, the ground sample and corresponding amount of Ce(NO3)3·6H2O were dissolved in deionized water, and then through magnetic stirring, ultrasonication, evaporation, and calcination, successively.

Catalyst characterization

The Field emission scanning electron microscopy (FE-SEM) images were taken with a Zeiss sigma 300 field emission scanning electron microscope, equipped with an energy dispersive X-ray spectrometer (EDX) to verify the dispersion of elemental composition and the ratio of Cu, Ce, Ti, Zr and O in the catalysts. The Atomic force microscopy (AFM) images were recorded using a Bruker Dimension Icon Atomic force microscope operated with PeakForce Tapping mode. The X-ray diffraction (XRD) patterns of catalysts were recorded using a Bruker D8 Advanced diffractometer with Cu Kα radiation (40 kV 30 mA) in the 2θ range from 10° to 80° with step size 0.05° and time step 0.2 s. The Brunauer-Emmert-Teller (BET) surface area and Barrett-Joyner-Halenda (BJH) pore size distribution of the catalysts were determined by N2 adsorption/desorption isotherms at −196 °C on a Micromeritics TriStar Ⅱ 3020 analyzer. Before measurement, the samples were degassed at 180 °C for 6 h in vacuum. The Fourier Transform Infrared Spectroscopy (FT-IR) was conducted on a Nicolet Avatar 360 FT-IR spectrometer. The IR spectrometer was equipped with a DTGS KB detector. An average of 64 scans with the optical velocity of 0.6334 cm−1 and wavenumber resolution of 4 cm−1 was composed in the range of 400–4000 cm−1. The X-ray photoelectron spectroscopy (XPS) measurements were performed on a Kratos Axis Ultra DLD spectrometer equipped with Al Kα radiation source. The charging effects of all catalysts were compensated by calibrating the binding energy with C1s peak at 284.6 eV. The thermogravimetric analysis (TGA) was carried out on a Henven HCT-4 thermoanalyzer, where 10 mg of prepared sample was heated from room temperature to 800 °C at a heating rate of 10 °C/min with N2 flowing at 80 mL/min.

Catalytic performance test

The removal of NO in the simulated flue gas was studied with a bench-scale experimental system, as illustrated in Fig. 1. The catalyst activity tests were carried out in a fixed-bed reactor with a tubular furnace and contained a temperature control and gas mixing system. Here, 200 mg catalyst was fitted in a quartz tube of the tubular furnace and pretreated in high purified N2 stream at 200 °C for 1 h and then cooled to room temperature, after that, the simulated flue gas was switched on. The reactant gas mixture contained 500 ppm NO, 500 ppm NH3, 5% O2 and N2 balanced, was controlled by mass flow controllers (MFCs), and the total flow rate of the feeding gas was 500 mL/min. The temperature was varied from 150 °C to 400 °C in 50 °C per step, and an independent experiment was conducted using a new catalyst at each temperature point. During each catalytic test, the temperature reached the preset value and lasted for 10 min until the temperature reading was stable, then the temperature and corresponding NO concentration were recorded every 30 s in 5 min, and the average value as final data. The concentration of NO was continually monitored by a Testo350 flue gas analyzer (Testo Instruments International Trading (Shanghai) Co., Ltd, China). The NH3-SCR activity was determined by NO conversion, which was calculated by the following equation:where XNO represents NO conversion, CNO(in) denotes the NO concentration in the feed gas and CNO(out) is the NO concentration in the effluent stream.

Fig. 1

Schematic diagram of the experimental setup.

Multifractal theory and the algorithm

Multifractal analysis was applied to provide statistical properties for the catalysts in terms of their multifractal generalized dimension and singularity spectra in this study, and a brief description of the multifractal approach by the box counting algorithm is discussed as follows.

The size of electron microscope image of the prepared catalyst is assumed to be N × N. The image is covered by boxes with the size of ε × ε (ε = 20, 21, …, ) to cut into N(ε) boxes. If the metal particles counted in the ith box of size ε is stated as Mi(ε), the mass probability in the ith box is expressed as [19]

For the metal particles with multifractal property, the mass probability Pi(ε) scales with the size εwhere αi is called the Lipschitz–Hölder exponent or singularity strength, characterizing the density in the ith box.

If the number of boxes in the electron microscope image of the catalyst is recorded as Nα(ε) when singularity strength ranged from α to α + dα, thenwhere f(α) is the multifractal spectrum.

In order to calculate the regular and irregular multifractal spectrum, the partition function χq(ε) of the moment order q for Pi(ε) is selectedwhere Dq is the fractal dimension of q, called the multifractal generalized dimension.

The formulae used to calculate the multifractal generalized dimension Dq are as followswhen q < 0, the information about the small probability distribution for active metal particles is amplified; when q > 0, the information about the high probability distribution is amplified. Moreover, D0, D1 and D2 are capacity dimension, information dimension and correlation dimension, concerned with distribution range, size distribution and measurement interval, respectively.

The mass exponent τ(q) can be depicted as

The relationship between the multifractal generalized dimensions and mass exponent can be obtained according to Eqs. (5), (7)

The three equations below are key for the estimation of singularity spectra [18], [20], [21]

With Legendre transform, relationships among singularity strength α(q), multifractal spectrum f(α) and mass exponent τ(q) are given by

The singularity strength range Δα and can be described as

Where αmax and αmin are the maximum and minimum values of the singularity strength, respectively.

In general, the most popular multifractal parameters are Dq, α(q), f(α), Δα and Δf. Dq reflects the overall multifractal characteristics, while α(q) and f(α) reflect the local branch multifractal characteristics. Δα represents the difference and inhomogeneity of the whole fractal structure and Δf behaves the shape characteristics of the multifractal spectrum.

Results and discussion

Morphological analysis basing on FE-SEM and EDX results

FE-SEM characterizations were performed to investigate the morphology of all catalysts shown in Fig. 2. It suggests that TiO2-ZrO2 presents a block structure uniformly loaded with small particles, and larger particles deposit on Cu/TiO2-ZrO2-SI in the form of agglomerates. Meanwhile, more active ingredients are aggregated on CuCe/TiO2-ZrO2-SI, whereas CuCe/TiO2-ZrO2-UI exhibits high dispersion. The results indicate that ultrasonic treatment has a certain impact on the morphology and dispersibility of the catalysts. It is because that the random micro-environment with instantaneous high pressure and temperature produced by ultrasonic cavitation effect makes the active species to be fully in contact with the support, which improves the dispersion of active species on the support surface [22], [23]. The morphology of these catalysts was further characterized by EDX and the corresponding results are displayed in Fig. 3. It is observed that Cu species are well-distributed over CuCe/TiO2-ZrO2-UI, but aggregate on Cu/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-SI, which reflects that ultrasound contributes to the dispersion of active ingredients on the catalysts. In addition, the weight percentages and atomic percentages of various elements present in the catalysts are shown in Fig. S1 and Table S1. The data confirms that the weight proportion of Cu and Ce in CuCe/TiO2-ZrO2-UI are approximately 15% and 8%, respectively, which lower than CuCe/TiO2-ZrO2-SI and closer to the theoretical ratio (15.38 wt% and 7.69 wt%) of the catalysts in the experimental section. It is speculated that the ultrasonic cavitation effect promotes uniform distribution of active species, thus resulting in the weight percentages measured accurately.

Fig. 2

FE-SEM images of (a) TiO2-ZrO2 (b) Cu/TiO2-ZrO2-SI (c) CuCe/TiO2-ZrO2-SI (d) CuCe/TiO2-ZrO2-UI.

Fig. 3

EDX mappings of (a) Cu/TiO2-ZrO2-SI (b) CuCe/TiO2-ZrO2-SI (c) CuCe/TiO2-ZrO2-UI.

Textural analysis basing on XRD and N2 adsorption–desorption results

XRD was applied to determine the crystal structure of TiO2-ZrO2 support and the presentation forms of active species, as shown in Fig. 4. For TiO2-ZrO2 support, it only displays one diffraction peak at 2θ = 30.5° (PDF# 74-1504) due to the presence of ZrTiO4 phase. However, the peak becomes weaker after copper impregnation, indicating that there is a strong interaction between titanium-zirconium with copper, and CuO species have covered on the surface of TiO2-ZrO2 support [24]. In the pattern of Cu-loaded catalysts, the diffraction peaks locate at 32.5°, 35.6°, 38.7°, 48.8°, 53.4°, 58.2°, 61.6°, 66.3°, 68.0°, 72.3°, and 75.2° are ascribed to monoclinic structure CuO (PDF# 89-2529). It can be seen that the characteristic peak intensity of CuO over CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI are weaker than that of Cu/TiO2-ZrO2-SI, resulting from the strong interaction between copper with cerium on the catalysts [12]. For Ce-loaded catalysts, obvious characteristic peaks of CeO2 locate at 28.5°, 47.5° and 56.3° (PDF# 81-0792) can be observed. It is worth noting that the peak intensity of CuCe/TiO2-ZrO2-UI weakens compared with CuCe/TiO2-ZrO2-SI, meaning that UI method can aid Cu and Ce for better disperse on TiO2-ZrO2 support than SI method [24], [25]. The results further confirm that ultrasonic cavitation effect promotes the high dispersion of active ingredients on the catalyst.

Fig. 4

XRD patterns of TiO2-ZrO2, Cu/TiO2-ZrO2-SI, CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI catalysts.

To study the textural properties of different catalysts, the surface area and pore size distribution of the catalysts were determined by N2 adsorption–desorption isotherms and BJH pore size distributions, and the results are presented in Table 1 and Fig. S2. The BET surface area and pore structure of the catalysts are summarized in Table 1. It can be seen from Table 1 that the raw TiO2-ZrO2 exhibits the highest BET surface area (172 m2/g) and pore volume (0.14 cm3/g), the decreased BET surface area could be attributed to the blocking effect on pores arising from the impregnation of CuO and CeO2 species [26], [27], which is proved by the decreased total pore volume as shown in Table 1. For CuCe/TiO2-ZrO2 catalysts, CuCe/TiO2-ZrO2-UI behaves relatively higher BET surface area (86 m2/g) and pore volume (0.08 cm3/g), indicating that ultrasonic cavitation effect contributes to the formation of large surface area. It is possible to observe from Fig. S2 that TiO2-ZrO2, Cu/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI present typical type IV isotherms with type H2 hysteresis loop at p/p0 = 0.4–0.6, which is characteristic of mesoporous materials. The difference between hysteresis loops of CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI can be attributed to cavitation-induced evaporation [28]. Besides, there is an artifact between 4 and 5 nm in BJH pore size distribution curves of CuCe/TiO2-ZrO2-UI, which is caused by the tensile strength effect (i.e., cavitation) [29]. Therefore, ultrasonic treatment has an obvious effect on the catalyst texture.

Table 1

The BET surface area and pore structure of the catalysts.

Catalysts	BET surface area (m2/g)	Pore volume (cm3/g)	Average pore diameter (nm)
TiO2-ZrO2	172	0.14	2.91
Cu/TiO2-ZrO2-SI	117	0.10	2.88
CuCe/TiO2-ZrO2-SI	65	0.06	3.28
CuCe/TiO2-ZrO2-UI	86	0.08	3.05

Surface acidity analysis basing on FT-IR results

FTIR experiments were conducted to investigate the surface acid sites of the catalysts, and Fig. 5 displays FT-IR spectra of various catalysts. In the high wavenumber range, the band is observed at 3438 cm−1 of N–H stretching region, indicating that NH3 can be adsorbed steadily on Lewis acid sites and Brønsted acid sites [30]. It has been reported that the bands at 1100–1300 cm−1, 1500–1750 cm−1 are ascribed to NH3 adsorb on Lewis acid sites, while those at 1340–1560 cm−1, 1640–1850 cm−1 are assigned to NH4+ species adsorb on Brønsted acid sites [31], [32]. Compared with Cu/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-SI, the band at 1390 cm−1 of CuCe/TiO2-ZrO2-UI, decreases in intensity. On one hand, Ce loading results in a slight decline in the number of accessible Brønsted acid sites, maybe owing to the partial blocking on catalyst structure caused by CeO2 species [33]. On the other hand, the amount of NH4+ adsorb on Brønsted acid sites of CuCe/TiO2-ZrO2-UI decreases, which is mainly due to some bridged OH groups are substituted by Cu2+ [34], proved by the band associated with (Cu2+–OH) groups at 3646 cm−1 [35], [36]. It can be speculated that the ultrasonic cavitation effect would promote the form of Cu2+ species, which will be further discussed below.

Fig. 5

FT-IR spectra of TiO2-ZrO2, Cu/TiO2-ZrO2-SI, CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI catalysts.

Redox properties analysis basing on XPS results

The XPS technique was used to characterize the chemical states of elements, the types of oxygen species and the interaction of active ingredients over catalysts, and the corresponding results are presented in Fig. 6. For O 1 s spectra of the catalysts (Fig. 6(a)), exhibiting that there exist two binding energy peaks of oxygen for the catalysts. The two distinct peaks at low binding energy (529.8–530.1 eV) and the higher binding energy (531.2–531.6 eV) are regarded as the lattice oxygen (Oβ) and the surface chemisorbed oxygen (Oα), respectively [37]. Furthermore, due to its mobility stronger than Oβ, Oα has been reported to be the most active oxygen, promoting the oxidation of NO into NO2, and finally the NH3-SCR reaction enhanced via “fast SCR” reaction [38], [39]. It can be seen from Table S2 that the ratio of Oα increases markedly after the introduction of cerium. The reason is that the surface chemisorbed oxygen mainly exists in oxygen vacancies (Vo) (Vo + O2 + e− → adsorbed oxygen) [37] and Ce3+ will originate oxygen vacancies, charge imbalance and unsaturated chemical bonds [40]. Additionally, it also should be noticed that the ratio of Oα on CuCe/TiO2-ZrO2-UI (0.425) is visibly higher than that on CuCe/TiO2-ZrO2-SI (0.376). On one hand, it probably because ultrasonic treatment facilitates Ce3+ dissolves from CeO2 surface [41]. On the other hand, ultrasound creates more lattice defects and greater surface area of the catalysts [42], inducing more oxygen vacancies and higher surface chemisorbed oxygen concentration.

Fig. 6

XPS spectra of the catalysts: (a) O 1s, (b) Cu 2p, (c) Ce 3d.

Fig. 6(b) shows the Cu 2p spectra of the modified catalysts, which contains two main peaks detected at about 933.8 eV and 953.6 eV are assigned to Cu 2p3/2 and Cu 2p1/2. The presence of both Cu2+ and polycrystalline Cu2O species in these catalysts is confirmed from the existence of shake-up satellite peaks within 940–945 eV [43]. By the peak deconvolution, the binding energies at 935.5 eV and 955.4 eV are attributed to Cu2+, while the peak at low binding energy at 933.6 eV and 953.5 eV can be regarded as the Cu+. It can be concluded from these results that the Cu2+ and Cu+ species are coexistence on the modified catalysts. Compared to Cu/TiO2-ZrO2-SI, CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI exhibit the appearance of broader and weaker peaks, which because that the loading of cerium oxides stimulates the interaction of copper species with titanium, leading to the decreasing of outer electron cloud density around Cu species [44].

Fig. 6(c) illustrates the Ce 3d spectra of the CuCe/TiO2-ZrO2 catalysts. According to the previous literatures [39], [45], the spectrum of Ce 3d can be assigned to eight components, the peaks labeled as u' and v' are attributed to Ce3+, while the peaks denoted as u, u'', u''', v, v'' and v''' are represented for Ce4+. The experimental results express that Ce3+ and Ce4+ coexist on both sample surfaces. The coexistence of Ce3+ and Ce4+ will form Ce4+/Ce3+ redox couple on the surface, which is promotional for the NH3-SCR activity. According to the calculated results from the XPS peak areas presented in Table S2, the ratio of Ce3+ over CuCe/TiO2-ZrO2-UI (0.165) is larger than that over CuCe/TiO2-ZrO2-SI (0.127), while CuCe/TiO2-ZrO2-UI (0.543) demonstrates a slight decrease in Cu+ ratio compared with CuCe/TiO2-ZrO2-SI (0.554). It means that the increase of copper oxidation state and decrease of cerium oxidation state coexist and encouraged by ultrasonic treatment. There has occurred the synergistic interaction of copper with ceria, which is inferred as follows [26], [30]:Cu

It is mentioned above that Ce3+ is confirmed to generate oxygen vacancies, charge imbalance and unsaturated chemical bonds, which results in a higher proportion of surface chemisorbed oxygen and better denitration performance. So, it can be concluded that the ultrasonic effect would strengthen the synergistic interaction of Cu and Ce, thus increasing Cu2+ and Ce3+ contents and improving redox properties of the catalysts, which is consistent with the FT-IR results and the O 1s XPS spectra of different catalysts.

According to the Ti 2p and Zr 3d spectra of the catalysts (Fig. S3 in the Supporting Information), all the Ti 2p spectra exhibit doublet bands at about 464.4 eV (Ti 2p1/2) and 458.7 eV (Ti 2p3/2), Zr 3d spectra of these catalysts are fitted by two bands at about 184.5 eV (Zr 3d3/2) and 182.2 eV (Zr 3d5/2). The results signify that Ti and Zr both in the oxidation valence of + 4 and mainly react to form ZrTiO4 [46], which in concert with the XRD results.

Thermal stability analysis basing on TGA results

The thermal properties of the catalysts were confirmed by thermogravimetric analysis, and TG curves are shown in Fig. 7. According to the TGA results, the processes of weight loss are similar and can be divided into different steps for all samples. The initial weight loss between 30 °C and 200 °C (Zone Ⅰ) can be attributed to the elimination of physically adsorbed water [47], [48], [49]. The weight loss is relatively small over the temperature range of 450–600 °C (Zone III), it is likely that the loss of water occurring from dehydroxylation reactions [49]. The last one at the range from 600 °C to 800 °C (Zone Ⅳ) may be related to the structural arrangement of the zirconium oxide and the release of carbonate intermediates from the compound [50], [51]. It can be noticed that the weight loss of TiO2-ZrO2 (7.3%) sample is higher than Cu/TiO2-ZrO2-SI (5.5%), CuCe/TiO2-ZrO2-SI (3.6%) and CuCe/TiO2-ZrO2-UI (3.8%) samples, proving that the presence of copper ions makes the low weight loss of the catalysts [52]. Additionally, CuCe/TiO2-ZrO2-UI exhibits slightly lower thermal stability than CuCe/TiO2-ZrO2-SI, which results from the ultrasonic disruption [53], [54]. As reported in previous literature [55], [56], when layered double hydroxides or SAPO-18 loaded with Cu and Ce species, the weight loss exceeds 10% above 250 °C, which far more than the catalysts prepared in this study. The premium thermal stability can be ascribed to the strong interaction between loaded metals and TiO2-ZrO2 carriers. Moreover, the TG curves have no obvious change from 200 °C to 450 °C (Zone Ⅱ) in CuCe/TiO2-ZrO2 catalysts, which indicates that the catalysts are stable in the testing temperature range of the NH3-SCR reaction.

Fig. 7

TG curves of TiO2-ZrO2, Cu/TiO2-ZrO2-SI, CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI catalysts: (Ⅰ) Temperature interval of the elimination of physically adsorbed water; (Ⅱ) Temperature interval of NH3-SCR reaction; (III) Temperature interval of water occurring from dehydroxylation reactions; (Ⅳ) Temperature interval of the structural arrangement of the zirconium oxide and the release of carbonate intermediates from the compound.

Catalytic activity for NO conversion

To evaluate the denitration performance of all catalysts in NH3-SCR reaction, NO conversions over the raw and modified SCR catalysts under the simulated flue gas conditions with the temperature from 150 °C to 400 °C are presented in Fig. 8, each data point is the average of three determinations with the error bar representing the standard deviation. The experimental results demonstrate that the NO conversion of TiO2-ZrO2 support was very poor, which was below 50% during the entire reaction temperature range. The catalytic activity of Cu/TiO2-ZrO2-SI catalyst was slightly higher than TiO2-ZrO2 support and the NO conversion efficiency reached about 78% at 400 °C, but was still unsatisfactory. For CuCe/TiO2-ZrO2 catalyst prepared through stepwise impregnation method, it displayed a considerably enhanced catalytic activity, compared with TiO2-ZrO2 support and Cu/TiO2-ZrO2-SI catalyst. XNO on CuCe/TiO2-ZrO2-SI catalyst increased with increasing temperature, exceeding 85% at temperatures higher than 300 °C and closing to 40% NO conversion was achieved at 150 °C, indicating that the synergistic effect of copper and ceria promotes widening of the operating temperature window for the SCR reaction at temperatures below 300 °C [26]. As shown in Fig. 8, the sample prepared by ultrasonic-assisted impregnation method exhibited better catalytic activity than similar catalyst prepared by traditional stepwise impregnation method, higher than 85% of NO degradation and the highest NO conversion of 94% at the temperature range of 250–400 °C. Obviously, ultrasonic effect further improves NH3-SCR performance of the catalysts.

Fig. 8

NO conversion at different temperatures over the raw and modified SCR catalysts. Reaction conditions: 0.2 g samples, 500 ppm NO, 500 ppm NH3, 5%O2, balanced by N2 with a total flow rate of 500 mL/min.

These results show that the physicochemical characteristics and SCR performance of CuCe/TiO2-ZrO2 catalyst enhanced with the ultrasound treatment. By combining the characterization results above, we conclude that the higher catalytic activity of the CuCe/TiO2-ZrO2-UI catalyst is attributed to the highly dispersed active ingredients as well as good redox performance resulted from the synergistic interaction between Cu and Ce species.

Multifractal characteristics of size and distribution of metal particles on catalysts

In order to further characterize the particle size and distribution of active metal particles over the catalysts, a box counting algorithm was used to accomplish the multifractal analysis by considering the pixel mass distribution in digital SEM images. From the pixel mass distribution, FracLac, a plugin for ImageJ used for multifractal analysis of the electron microscope images, computes and returns data and graphics known as multifractal generalized dimension parameters and singularity spectra. Here, we used SEM images of TiO2-ZrO2, Cu/TiO2-ZrO2-SI, CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI to obtain multifractal parameters (Dq, α(q), f(α)) and relevant multifractal spectra.

Table 2 lists several parameters from the multifractal generalized dimension spectra, i.e., D0-D1, D1/D0, D0-D2 and ΔD. The difference D0-D1 and ratio D1/D0 are used to quantify the degree of dispersion of the metal particles on the catalysts, small value of D0-D1 and D1/D0 near 1 indicate a high degree of evenness [57]. For the CuCe/TiO2-ZrO2-UI, D0-D1 value is quite lower than other catalysts, and the D1/D0 value is closer to 1. Thus, the metal particles distribute evenly on the CuCe/TiO2-ZrO2-UI, which can be further proved by the dispersion of metal particles displayed in FE-SEM images and EDX mappings. The difference D0-D2 reflects the percentage of fine metal particles on the catalysts, as the value of D0-D2 decreases, so the catalyst expresses greater content of fine particles [58]. The D0-D2 value is much lower at CuCe/TiO2-ZrO2-UI compared to other metal-loaded catalysts. Besides, it can be seen in Fig. S4 that the particle size of CuCe/TiO2-ZrO2-UI distributes in the range between 30 and 100 nm, below the particle size distribution range of Cu/TiO2-ZrO2-SI (50–140 nm) and CuCe/TiO2-ZrO2-SI (60–180 nm). The crystallite size and particle size of CuCe/TiO2-ZrO2 catalyst with ultrasound treatment also smaller than raw catalyst (Table S3). Therefore, the overall content of ultrafine metal particles is much higher on the catalyst prepared by UI method, which results from the ultrasonic cavitation effect. In addition, ΔD is a useful parameter to describe the degree of surface differentiation, in which lower ΔD means lower complexity on the catalyst surface [59]. The calculated ΔD values of CuCe/TiO2-ZrO2 are found in the following order: CuCe/TiO2-ZrO2-UI (0.346) < CuCe/TiO2-ZrO2-SI (0.466), that is to say, the former owns a lower degree of complexity. It is reported that surface roughness, AFM images and fractal dimensions are related to each other [60]. Therefore, the complexity of the catalyst was measured by AFM, and the corresponding results shown in Fig. S5 and Fig. S6. The surface contour of TiO2-ZrO2 is favorable with the peaks concentrate between −0.3 nm to 0.3 nm and distribute uniformly, the peaks of Cu/TiO2-ZrO2-SI aggregate on the range of −0.2 nm to 0.2 nm, and the result is similar with CuCe/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2-UI. Moreover, according to the average roughness and root mean square roughness presented in Table S4, owing to the loading of metal oxide, Cu/TiO2-ZrO2-SI and CuCe/TiO2-ZrO2 catalysts become rougher than TiO2-ZrO2 support, but the average roughness and root mean square roughness of CuCe/TiO2-ZrO2-UI are still lower than CuCe/TiO2-ZrO2-SI. The above results represent that ultrasonic effect can reduce the surface complexity of the catalysts.

Table 2

Multifractal generalized dimension parameters for the catalysts investigated.

Catalysts	D0	D1	D2	D0-D1	D1/D0	D0-D2	Dmin	Dmax	ΔD
TiO2-ZrO2	1.651	1.628	1.620	0.023	0.986	0.031	1.603	1.968	0.365
Cu/TiO2-ZrO2-SI	1.767	1.743	1.732	0.024	0.986	0.035	1.711	2.121	0.410
CuCe/TiO2-ZrO2-SI	1.746	1.713	1.697	0.033	0.981	0.049	1.672	2.138	0.466
CuCe/TiO2-ZrO2-UI	1.751	1.738	1.738	0.013	0.993	0.013	1.722	2.068	0.346

The multifractal singularity spectra of metal particles distribution under different catalysts are illustrated in Fig. 9, which displays the shape and extension of the multifractal singularity spectrum f(α) versus α spectrum. For different catalysts, the spectra demonstrate asymmetry in various degrees and a left hook. The important parameters obtained from the multifractal singularity spectra are exhibited in Table 3. Δα is used to describe local variability characteristic of the catalysts, where lower Δα indicates more well-distributed [18]. Under ultrasonic modification, the change range of Δα is smaller than that of untreated catalyst, showing that ultrasonic vibration will promote the homogeneity over the catalyst surface. Δf reflects structural changes of fractal particles in different probability subsets, Δf > 0 implies the number of maximum probability subsets is more than that of the minimum probability subsets, or vice versa [61]. All of the Δf values are positive number thereby revealing higher probability subsets of active metal particles play dominant roles at all catalysts. In a word, according to some parameters and spectra of multifractal theory, it is further certified that catalyst surface features, such as particle size and distribution of active metal particles, will be better with the ultrasound modification.

Fig. 9

Multifractal singularity spectra for the metal particles distribution over the catalysts.

Table 3

Multifractal singularity parameters for the catalysts investigated.

Catalysts	αmin	f(αmin)	αmax	f(αmax)	Δα	Δf
TiO2-ZrO2	1.591	1.471	2.072	0.929	0.481	0.542
Cu/TiO2-ZrO2-SI	1.702	1.619	2.243	0.907	0.541	0.712
CuCe/TiO2-ZrO2-SI	1.665	1.598	2.265	0.873	0.600	0.725
CuCe/TiO2-ZrO2-UI	1.706	1.550	2.178	0.966	0.472	0.584

Conclusions

A series of CuCe-modified TiO2-ZrO2 catalysts were prepared through stepwise impregnation method and ultrasonic-assisted impregnation method for selective catalytic reduction by NH3 of NO in the simulated flue gas. Higher than 85% NO was degraded at the temperature range of 250–400 °C and the highest NO conversion of 94% at 350 °C were obtained on the CuCe/TiO2-ZrO2 catalyst prepared by the ultrasonic-assisted impregnation method, indicating that catalytic performance was enhanced transparently by ultrasonic treatment. To research the effect of ultrasound modification on CuCe/TiO2-ZrO2, different samples were investigated in detail by physicochemical characterization techniques. The results show that active metal dispersion and synergistic interaction between Cu and Ce species (Cu++Ce4+ ↔ Cu2++Ce3+) of the catalysts are promoted through the ultrasonic cavitation effect. Besides, by way of multiple analysis, it can be concluded that ultrasonic treatment facilitates the particle size and distribution of active sites on the catalysts, which described by the generalized dimension and singularity spectra and validated by the comparative test.

CRediT authorship contribution statement

Wei Zhang: Conceptualization, Methodology, Software, Resources, Data curation, Supervision, Project administration, Funding acquisition. Yunhao Tang: Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Visualization. Cheng Lu: Formal analysis, Investigation, Writing - original draft. Jiyao Zou: Writing - review & editing. Min Ruan: Validation, Funding acquisition. Yanshan Yin: Validation, Writing - review & editing, Funding acquisition. Mengxia Qing: Writing - review & editing. Quanbin Song: Writing - review & editing, Funding acquisition.

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.