1 Introduction

Nitrogen oxides are one of the most common contributors to atmospheric pollution. The majority of these emissions are caused by the fossil fuels combustions, such as in coal-fired power plants or the diesel engines of cars and trucks (Luo et al. 2022; Shan et al. 2016), which has caused many environmental issues (Zhu et al. 2022; Qi et al. 2020; Awual et al. 2019; Awual 2016). Selective catalytic reduction (SCR) of NOx with ammonia is currently the most imperative, successful, and commercially applicable method for NOx removal (Twigg 2007; Pu et al. 2022). Iron-based oxides are recognized as a kind of efficient SCR catalysts in industrial applications because of their low toxicity, high reactivity and satisfactory N2 selectivity (Han et al. 2019a, b). Unfortunately, the SCR reactivity of pure iron oxide-based catalysts at low temperatures (LT) is not satisfied owing to the lack of acid sites and redox capabilities (Mu et al. 2020). Numerous attentions have been attached on regulating the crystalline phase /facet structures and nanostructures of iron oxides, or doping heteroatoms into the oxide matrix so that improve the oxidizing ability and acidity/redox properties of catalysts (Han et al. 2019a, b). Nevertheless, how to produce Fe-based SCR catalyst with simpler method and low-cost is still challenging (Han et al. 2019a, b).

Zinc slag (ZS), which is also called as lead-zinc slag, or ferro-silicate slag (FSG) (Morrison et al. 2003), is the primary solid waste produced during the zinc ore smelting processes. In recent years, the global metal zinc consumption has exceeded 13 million tons per year, which is inevitably accompanied by the increase in ZS discharge (Song et al. 2019) and the issue of ZS disposal has not been properly solved. Recently, Nath et al. (Nath 2020) reported the preparation geopolymer paving blocks using ZS/fly ash blends as raw materials. Xia et al. (Xia et al. 2019) investigated the solidification/stabilization of ZS by preparing of geopolymer composites. There are also some other similar reported works, which all highlight the harmless utilization potential of ZS in construction and building materials areas (Zhang et al. 2022; Pan et al. 2019). Nevertheless, it is still wasteful to use ZS as building materials because large amounts of useful elements such as Fe and Mn in ZS are not sufficiently used with high added value. Indeed, ZS contains approximately 35 wt% Fe2O3 and a small amount of other metal oxides (Nath 2020; Alex et al. 2013; Prasad et al. 2016). Hence, if the active components in zinc slag could be extracted and enriched producing nanocomposite material(Awual 2016, 2017; Awual et al. 2017a, b; Awual 2019), zinc slag would become potential raw to prepare environment remediation materials with high added value, e.g. composite Fe-based catalyst for the treatment of wastewater or waste gas (Zhang et al. 2013; Chen et al. 2023; Li et al. 2021; Awual et al. 2017a, b, 2019).

Herein, ZS was utilized as the sole raw material to prepare novel low-cost and high-performance iron oxide-based deNOx catalyst by extracting and restructuring of the endogenous constituents of ZS. This idea is based on following considerations:(1) ZS is a kind of solid waste, but it contains a considerable high iron content (greater than 35 wt%), which can be extracted and enriched under acidic condition. (2) Except for iron species, some other acid soluble elements such as Al and Mn also exist in ZS, which can be extracted simultaneously during the solation process. These elements can be utilized as heteroatoms doped in FeOx catalyst matrix to promote nanocomposite material producing. (3) This method allows the preparation of structure controllable and analyzable iron oxide-based deNOx catalyst, which has weighty significance for zinc slag utilization in large-scale.

In the present work, zinc slag was dissolved by different concentrations of HNO3 to control the extraction rate of different elements, and followed by precipitation and calcination to prepare heteroatoms doped iron oxide-based SCR catalysts. The changes in the pore structure, crystalline structure, surface texture, redox capability, chemical components, and microtopography under different acid concentrations are systematically evaluated via various spectroscopy methods, and the changes are also compared with their variation in SCR reactivities. Moreover, a pure Fe2O3 was also prepared using chemical reagent under the same condition as comparison, so that better protrude the superiority and advancement of the iron oxide-based catalyst derived from low-cost zinc slag waste.

2 Materials and methods

2.1 Catalyst synthesis

Raw zinc slag (ZS) was provided by Shanxi Shangluo Lead-zinc smelter, which was directly used without further purification. Nitric acid (65 wt%), analytic pure reagents including ion nitrate nonahydrate, aluminum nitrate nonahydrate, silica sol, magnesium nitrate hexahydrate calcium chloride, zinc nitrate hexahydrate, and manganese nitrate hydrate, and ammonium hydroxide (25 wt%) were provided by Sinopharm Chemical Reagent Co., Ltd.

ZS-derived catalysts (Ha-FeOx/yZS) were synthesized by the sol-gel method. Firstly, raw was dried and ground to pass 0.154 mm sieve, and then dissolved 10 g of raw material in y mol/L (y = 0.5, 2.0, 4.0, 6.0) of HNO3 solution at 75 °C for 6 h to extract metal ions. The mass ratio of solid/liquid was 1:30. Subsequently, ammonia hydroxide was added dropwise into the mixture solution for adjusting the pH value to approximately 9 and transforming the mixture solution from sol to gel via stirring for several hours. Then, the obtained gel was filtered and separated, and washed for 3–5 times ensuring the pH of the gel was around 7. Thereafter, the gel was dried in an oven (105 °C, 12 h) before being calcined in a muffle furnace (400 °C, 5 h). The FeOx sample was also prepared via sol-gel method under the same condition to study the effect of heteroatoms on the structure and morphology of the samples.

2.2 Catalytic experiments

The performance testing experiments were carried out in a quartz reactor (inner diameter: 6 mm). In each run, 0.4 g of catalyst was put into the reactor with pretreatment at 450 °C for half an hour. After cooling the reactor to 80 °C, the gas mixture was injected into the reaction cell at 168 mL/min (NO = NH3 = 750 ppm; 5 vol% O2; GHSV = 20,000 h− 1 balanced by N2). The outlet gas concentration was recorded every 25 °C interval for 30 min from 125 to 450 °C to assess the NO conversion of the catalysts. The NOx conversion and N2 selectivity are calculated by equations below (Pu et al. 2022):

$$\text{N}\text{O} \,\text{c}\text{o}\text{n}\text{v}\text{e}\text{r}\text{s}\text{i}\text{o}\text{n} =\left(1-\frac{{\left[{\text{NO}}_{x}\right]}_{\text{out}}}{{\left[{\text{NO}}_{x}\right]}_{\text{in}}}\right)\times 100\%$$
(1)
$${\text{N}}_{2}\,\text{s}\text{e}\text{l}\text{e}\text{c}\text{t}\text{i}\text{v}\text{i}\text{t}\text{y}=\left(1-\frac{{2\left[{\text{N}}_{2}\text{O}\right]}_{\text{out}}}{{\left[\text{NO}\right]}_{\text{in}}+{\left[{\text{NH}}_{3}\right]}_{\text{in}}-{\left[\text{NO}\right]}_{\text{out}}-{\left[{\text{NH}}_{3}\right]}_{\text{out}}}\right)\times 100\%$$
(2)

The \({\left[{\text{NO}}_{x}\right]}_{\text{in}}\) is inlet \({\text{NO}}_{x}\) concentrations; \({\left[{\text{NO}}_{x}\right]}_{\text{out}}\) is outlet \({\text{NO}}_{x}\) concentrations; while \({\left[{\text{NH}}_{3}\right]}_{\text{in/out}}\) is inlet or outlet NH3 concentrations, respectively.

2.3 Characterizations

The chemical components of the samples were quantitatively analyzed by X-ray fluorescence (XRF) using an ARL PERFORM’X Analyzer. The crystalline structure of the samples was detected by Rigaku D/Max 2200 X-ray diffractometer (XRD) at 40 kV and 26 mA. Raman spectrum was recorded on the LabRAM HR Evolution spectrometer using a 532 nm laser beam. SEM and TEM images were observed by a Sigma-300 scanning electron microscope (Zeiss, Germany) and a JEOL JEM-2100 F field-emission electron microscopy at 200 kV. Micrometrics APSP 2460 analyzer was utilized to recorded the gas adsorption isotherm for further calculating the BET surface areas as well as the pore size distributions. Thermo SCIENTIFIC ESCALAB Xi + equipment was used for analyzing the surface state of the catalysts.

H2-TPR and NH3-TPD experiments were carried out on a completely automatic chemical adsorption device (Auto Chem TM II 2920, Micromeritics, USA). For H2-TPR: catalysts were preheated in a helium flow (300 °C, 1 h) and then heated to 800 °C at 10 °C/min in 10% H2/He (50 mL/min). For NH3-TPD: catalysts were placed in a reaction tube and firstly heated with helium flow (300 °C, 1 h). After the catalyst cooled to 50 °C, a flow of 10% NH3/He (50 mL/min) was fed in until the equilibrium of adsorption, and then heated to 800 °C at 10 °C/min.

VERTEX 70v was utilized for observing the in-situ DRIFTS spectra. Catalysts were first processed in N2 (350 °C, 30 min) flow to remove impurity substances and then cooled to 250 °C. Then, a gas mixture (NH3 or NO + O2) was introduced into catalyst surface for 30 min, purged with N2, and reacted with the flow of NH3 or NO + O2. The IR curves were recorded at intervals and analyzed in order to comprehend the SCR catalytic mechanism.

3 Results and discussion

3.1 SCR performance

Figure 1a compares the catalytic performance of ZS, FeOx and Ha-FeOx/2ZS samples. Owing to the presence of abundant iron and other metal species, ZS sample exhibits SCR reactivity, where the NO conversion slowly increased in 125–425 °C and then decreased. However, the highest NO conversion of ZS is only approximately 78% at a high temperature of 375 °C, and it is almost inactive at LT (less than 250 °C), suggesting that ZS is not suitable directly utilized for NOx removal, which requires a further treatment. Meanwhile, FeOx sample exhibits approximately 65% NO conversion at around 300 °C and the change of conversion rate with temperature is similar to the ZS. Notably, following by recombining the active contents from zinc slag, the NO conversion over Ha-FeOx/2ZS is obviously increased which means that the new synthesized catalyst could maintain excellent catalytic performance (> 90%) in the temperature range of 190–340 °C.

Fig. 1
figure 1

a Catalytic activity of ZS, FeOx and Ha-FeOx/2ZS catalysts b Catalytic activity of Ha-FeOx/yZS catalysts c N2 selectivity of Ha-FeOx/yZS catalyst d stability of Ha-FeOx/2ZS at 250 °C (NO = NH3 = 750 ppm; O2 = 5 vol%; balance N2; GHSV = 20,000 h-1)

Full size image

As shown in Fig. 1b, the ZS-derived catalysts prepared under different nitric acid conditions all show better SCR performance than that of zinc slag, especially at low temperatures. When ZS was treated under low acid concentration, the NO conversion of the Ha-FeOx/0.5ZS was 90% at 225 °C, and up to 96% at 275–300 °C. Although the NO conversion of Ha-FeOx/0.5ZS gradually decreases with the rising of the temperature, it remains at a high value of approximately 88% at 350 °C, higher than that of raw ZS. Furthermore, when ZS was treated under higher acid concentrations, the NO conversion is further enhanced, whereas the temperature window showed an increasing trend first and then declined. Ha-FeOx/2ZS exhibits the widest reaction temperature window. Its NO conversion reached up to 82% and N2 selectivity was nearly 100% at 170 °C (Fig. 1c). The conversion ratio further increased above 90% at elevated temperatures (190–325 °C). According to Figure S1, Ha-FeOx/2ZS also performed well at different GHSV. Although the GHSV went up to 80,000 h − 1, the activity was only decreased from 98.4 to 89.2% at 250 °C. Moreover, the ZS-derived catalysts had excellent LT SCR reactivity and wide working temperature ranges when comparing with other reported catalysts (Table 1).

Table 1 The reactivity of various iron-based catalysts
Full size table

3.2 Physical properties

Table S1 summaries the chemical components of raw ZS and ZS-derived catalysts. It is clear that ZS is predominately made up of Fe, Si, Al and Ca species, Especially, the Fe species account for half of the original amount. Other metal species such as Mn, Zn, Mg, etc. are less than 5%. However, the amount of iron species in the Ha-FeOx/yZS catalyst significantly grew from 49.34% to 73.36% with the increase of acidic concentration. In contrast, the amount of Si species indicated a decreasing trend because of the insoluble quartz existing. In a word, after the acid extraction and alkali precipitation treatment (Fig. 2a), the main elemental compositions of the obtained catalysts are similar to that of ZS, but they are essentially different among crystalline structure, microtopography, and physicochemical property.

Fig. 2
figure 2

a The procedure for the preparation of Ha-FeOx/yZS catalysts via sol-gel method; FESEM image of b Raw ZS, c FeOx and d Ha-FeOx/2ZS

Full size image

As shown in Fig. 2b, the raw ZS exhibits a condensed bulk morphology with some cracks that formed during the slag collection process. Figure 2c displays the microstructure of pure FeOx, and we can see that the iron oxides are well crystalized with a particle size of approximately 30 nm, and these regular multilateral particles are packing closely. However, the Ha-FeOx/yZS catalyst as shown in Fig. 2d exhibits an amorphous morphology, and the catalyst particles are accumulated by smaller nanoparticles with particle size of approximately 5–10 nm. Moreover, the interconnected accumulation of these refined nano-sized particles naturally forms abundant pore structures, which would facilitate the SCR reaction.

Figure 3a displayed the XRD patterns of different catalysts, where we could see that ZS exhibited a broad diffraction peak coupled with some sharp ones. In reported literatures, the broad hump peak at approximately 20°–40° 2θ was regarded as amorphous aluminosilicate (Chen et al. 2019), while these sharp peaks at 35.4, 43.0 and 53.4 are ascribed to (311), (400) and (422) surfaces of quartz (JCPDS NO.46-1045), crystalline magnetite (JCPDS NO.74–0748) and mullite (JCPDS NO.31–0267), respectively.

Fig. 3
figure 3

a XRD pattern and b Raman spectra of ZS, Ha-FeOx/2ZS and FeOx samples

Full size image

Differently, as shown in Fig. 3a and Figure S2, all the samples treated with different acid concentrations exhibited two broad hump peaks centered at approximately 35° and 62°, which are the characteristic peaks of amorphous Fe2O3 doped with heteroatoms. Zhang et al.(Zhang et al. 2021a, b) and Sun et al.(Sun et al. 2017) demonstrated that the doping ions lead to the formation of poorly crystalline Fe2O3. As a comparison, we can see that the pure FeOx prepared from chemical reagent under the same condition displays clear diffraction peaks, where the peaks at 24.1°, 33.1°, 35.6°, 40.8°, 49.4°, 54.0°, 57.4°, 62.4°, 64.0°, and 71.9° 2θ are consistent well with (012), (104), (110), (113), (024), (116), (122), (214), (300) and (1010) crystalline surfaces of hematite Fe2O3. These results suggest that the impurity elements from ZS would greatly affect the crystal growth of Fe2O3. Indeed, previous works have also shown that the doping of cations such as Si, Al, Ti, and Ce into the FeOx matrix would significantly induce the deformation of FeOx unit cell and suppresses crystal growth, thereby reducing the particle size and results in amorphous phase product (Sun et al. 2017). Notably, poor crystallized FeOx was more efficient in SCR reaction because it had better pore structure and more exposed active sites (Zhang et al. 2021a, b).

Raman spectra of the ZS, Ha-FeOx/2ZS and FeOx were shown in Fig. 3b. ZS had three mainly large peaks at 370, 520 and 670 cm− 1, which were considered as the magnetite phase in raw ZS (Chernyshova et al. 2007; Bersani et al. 1999). For pure FeOx, peaks at 225 (A1g), 245 (Eg), 411 (Eg), 504(A1g), and 611 (Eg) cm− 1 ascribed to α-Fe2O3 were clearly observed (Faria et al. 2005). What is worth highlighting, however, is that the peak related to the α-Fe2O3 shows significantly weaker and wider for Ha-FeOx/2ZS catalyst with an emerging shoulder peak at 611 cm− 1 (Chernyshova et al. 2007), which reveals the presence of the poorly crystallized Fe2O3 nanoparticles, as demonstrated in Figs. 2d and 3a.

From Fig. 4a, ZS was composed of nano particles, which should be the aggregation of oxides such as magnetite, quartz and mullite as revealed by XRD. Moreover, owing to the high content of iron species, the crystalline Fe3O4 can be observed in raw ZS as shown in Fig. 4d. Figure 4b and e exhibited the TEM images of pure FeOx, where we can see that the FeOx is well crystalized Fe2O3 structure with particle size of approximately 30 nm. Figure 4c and f displayed the TEM results of Ha-FeOx/2ZS, where we can see that the Ha-FeOx/2ZS has consisted of nanoparticles at 5–10 nm, smaller than that of pure FeOx. Moreover, the high-resolution image as illustrated in Fig. 4f confirms that the nano particles are mainly amorphous, and the amorphous structure also indicates its excellent SCR reactivity(Hasan et al. 2023a, b; Zhang et al. 2021a; Zhang et al. 2021b). Furthermore, the EDS mapping of ZS-derived catalyst as shown in Fig. 4g–n demonstrated that except for Fe, there are also some heteroatoms exist in the catalyst, and these elements are uniformly distributed in catalyst matrix, which should be the key factor responding for the as-prepared catalyst’s unique structure and catalytic behavior.

Fig. 4
figure 4

TEM results of a, d Raw ZS, b, e Pure FeOx, c, f Ha-FeOx/2ZS; the elements mapping of Ha-FeOx/2ZS for Fe, O, Si, Al, Mn, Zn, and Ca (h-n)

Full size image

Figure 5 and Figure S3 display the N2 adsorption-desorption isotherms and pore size distributions of the samples, while the parameters are shown in Table 2. As shown in Fig. 5a, the plot of ZS was a type IV isotherm with a wide hysteresis loop at mid-high relative pressure ranges of 0.42–0.98 (P/P0) (Chen et al. 2022), and the N2 uptake amount was very limited, suggesting its poor developed pore structures, where the pore size distribution was dispersive, and the BET surface area was only 28 m2/g. In contrast, as illustrated in Fig. 5a and Figure S3a, all the resultant catalysts obtained under different acid conditions exhibit type IV isotherms, and the N2 adsorption amounts are significantly larger than raw ZS, suggesting their well-developed hierarchical pore structure characteristics (Chen et al. 2019). As shown in Fig. 5b and Figure S3b, the ZS-derived catalysts are composed of both micropore and mesopore centered at approximately 1.4 and 3.8 nm, respectively, and the mesopores are predominant, which would promote the transformation of the reactants during the reaction, thereby facilitating the SCR reactivity. Moreover, as a comparison, we can see from Fig. 5a; Table 2 that the pure FeOx exhibits much poorer pore structure than that of ZS-derived catalyst. These results further confirm that the endogenous heteroatoms in ZS can not only regulate the crystalline structure of the ZS-derived catalyst, but also significantly optimize its pore structures, which responses to its excellent SCR performance (Chen et al. 2011; Salman et al. 2023).

Fig. 5
figure 5

a N2 adsorption-desorption isotherms b Differential pore size distributions of ZS, FeOx, Ha-FeOx/2ZS catalysts

Full size image
Table 2 Structural characteristics of catalysts
Full size table

3.2.1 Surface chemical states analysis

XPS spectra was measured for further understanding the composition and state of the catalysts’ surface elements. As illustrated in Fig. 6a and Figure S4, the O 1s XPS spectra was split into three independent peaks located at 529.3–530.2, 530.9–532.0 eV and 533.0-533.4 eV. The band at lower binding energy was lattice oxygen O′ species (O2−), and the higher represented for chemisorbed oxygen O″ species, including the surface oxygen vacancy (O22−, O2−) and adsorbed water, respectively (Wei et al. 2020; Liu et al. 2010; Liu et al. 2021a, b). O″ was more active in the oxidizing process because they have higher mobilities than lattice oxygen species (O′) (Kang et al. 2021); besides, the adsorbed chemisorbed oxygen O″ species can effectively promote the oxidation of NO (Fang et al. 2022), and the redox cycle of active substances, thereby facilitating the SCR reactivity of samples (Liu et al. 2021a, b; Sheng et al. 2018). In Table 3; Fig. 6a, all the ZS-derived catalysts exhibit high chemisorbed oxygen O″ species above 72%, which are much higher than that of pure FeOx (28.04%). The biggest difference between ZS-derived iron oxide-based catalysts and pure FeOx is their chemical components. As shown in Table 3, except for iron species, a small amount of impurity cations such as Si, Al, and Mn are also present in ZS-derived catalysts’ surface. Combining with above XRF, XRD, BET, Raman, and microtopography results, it can be certified that the impurity atoms that simultaneously extracted during the acid treatment of ZS can effectively inhibit the growth of iron oxide, which not only improve pore structure of iron oxide-based catalysts, but also increase its adsorbed oxygen content, thereby improving its SCR reactivity.

Fig. 6
figure 6

XPS patterns of a O 1s; b Fe 2p of the ZS, Ha-FeOx/2ZS and FeOx catalysts

Full size image
Table 3 Atomic concentrations and relative ratios of the main elements
Full size table

Furthermore, these heteroatoms can also improve the redox property of iron oxide-based catalyst. In general, higher Fe2+ content in iron oxide-based catalyst indicates better redox capability (Zhang et al. 2020; Yuan et al. 2022; Wang et al. 2020). However, for a pure FeOx catalyst, as shown in Fig. 6b and Table 3, owing to its well-developed crystalline structure, it is mainly composed of Fe3+, where the Fe2+ proportion is only approximately 16.92%. This should be one of the crucial underlying causes that pure FeOx catalyst exbibits poor SCR reactivity. In contrast, the Fe2+ contents in ZS-derived catalysts are 2 to 3 times larger than that of pure FeOx, which should be attributed the influence of impure heteroatoms. Indeed, Sun et al. (Sun et al. 2017) analyzed the influence of doping with heteroatoms on the Fe2+ content of iron oxide-based catalysts. Their results confirmed that the strong interaction among the doped cations and Fe species would regulate the inherent electron cloud of iron atoms, which promotes the formation of more Fe2+ species, thereby facilitating the redox property of samples, and further enhance SCR reactivity.

3.2.2 H2-TPR

Herein, H2-TPR experiments were performed to better understand the redox properties of samples. Based on previous studies, the reduction of pure Fe2O3 with H2 could take place in the following order: Fe2O3→Fe3O4→FeO→Fe with a hydrogen molar ratio of 1:2:6 (Zhang et al. 2021a, b; Liu et al. 2021a, b). According to this ratio, the H2-TPR profiles of pure FeOx and ZS-derived catalysts were divided into three steps with increasing temperatures as shown in Fig. 7, Figure S5, and Table S2. In Fig. 7, the reduction temperature of pure Fe2O3 to Fe3O4 was centered at approximately 318–350 °C (Zhang et al. 2021a, b). However, we can clearly see that the same reduction process in ZS-derived catalyst (Ha-FeOx/2ZS) was centered at approximately 265–325 °C, lower than pure Fe2O3. Moreover, the following reduction steps show similar characteristic, i.e., the reduction of iron species in ZS-derived catalysts requires obvious lower temperature than that of well crystalized pure FeOx, firmly conformed the intensive interaction between Fe and impurity atoms in ZS-derived catalysts, which improves the reducibility of iron species, thereby facilitating its LT SCR reactivity. Besides, in Figure S5 and Table S2, ZS sample exhibits only one broad peak at 370–715 °C, and no reduction peak can be observed at lowers temperature, suggestion that the iron species in ZS mainly are mainly Fe3O4.

Fig. 7
figure 7

H2-TPR patterns of FeOx and Ha-FeOx/2ZS samples

Full size image

3.2.3 NH3-TPD

The surface acid sites and acid species of the catalyst were important for promoting the NO conversion. Because sufficient NH3 adsorption could ensure further SCR reaction. (Sun et al. 2017; Liu et al. 2017a, b). Therefore, the NH3-TPD experiment was conducted to get more acidity information of ZS-derived catalysts. Generally, the strength and quantity of the catalyst’s acid sites can be intuitively reflected by the NH3 desorption temperature and amounts. At lower temperatures (50–250 °C), the desorption peak was related to the ammonia species physically adsorbed on the material surface and chemisorbed on the weak acid sites (Wang et al. 2020). The peak observed at approximately 350 °C was due to the adsorption of NH4+ on medium acid sites. The peak at high temperatures of 400–800 °C was assigned to the desorption of chemical adsorbed ammonia at strong acidic sites (Liu et al. 2017a, b, 2021a, b). In Figure S6, ZS shows three NH3 desorption peaks at approximately 110, 370 and 530 °C with total acid amount of 1.26 mmol/g. All the ZS-derived catalysts exbibit similar TPD curves showing two main NH3 desorption peaks at around 120 and 350 °C. The NH3 desorption amounts of the resultant catalysts (> 1.65 mmol/g) are significantly higher than that of raw ZS (1.26 mmol/g), revealing that the acid extraction and alkali precipitation treatment of ZS can effectively improve its surface acidity. Moreover, as shown in Fig. 8, the acid amount of Ha-FeOx/2ZS is roughly four times more than that of pure FeOx. Indeed, Zhang et al. (Zhang et al. 2021a, b) found that Nb species introduced into the iron oxide matrix effectively enhances its acid amounts; Sun et al. (Sun et al. 2017) also confirmed that the doping of Ti, Al, or Ce into FeOx can observably boost the surface acidity of resultant catalysts. Therefore, it can be speculated that the higher acid amount of ZS-derived catalysts than that of raw ZS is attributed to the structure regulation of ZS during the acid extraction and alkali precipitation treatment process, while the better surface acidity of the ZS-derived catalysts than pure FeOx is ascribed to the contribution of heteroatoms which also acquired from the zinc slag.

Fig. 8
figure 8

NH3-TPD patterns of pure FeOx and Ha-FeOx/2ZS catalyst samples

Full size image

3.3 In-situ FTIR study over Ha-FeOx/2ZS catalyst

In-situ DRIFT experiments of NH3 adsorption on Ha-FeOx/2ZS at 250 °C were carried out for further understanding the impact of the surface acidity on catalytic reaction. As illustrated in Fig. 9a, when Ha-FeOx/2ZS was persistently exposed to 750 ppm of NH3/N2 flow for 30 min, several IR bands of the adsorbed ammonia species on the catalyst surface emerged in the region of 1000–4000 cm− 1, and the peak intensities were enhanced with increasing exposure time. Among them, the bands at 1620, 1281, and 1107 cm− 1 were identified as the symmetric and asymmetric bending vibration modes of ammonia species absorbed on Lewis acid sites, and the bands at 1691, 1644, and 1392–1512 cm− 1 were originated from the coordinated NH4+ on Brønsted acid sites (Guo et al. 2022), while the peaks centered at 3365 and 3267 cm− 1 is related to the vibration of N-H bonds (Sheng et al. 2018; Liu et al. 2020; Zhu et al. 2021). The peaks emerged at 1344, and 1520–1549 cm− 1 were assigned to the amide (-NH2) species(Fang et al. 2020). This phenomenon shows that the sufficient acid sites of ZS-derived catalyst ensure abundant adsorbed ammonia species, and they can also be activated simultaneously to improve the NH3-SCR performance.

Fig. 9
figure 9

In-situ DRIFT spectra of Ha-FeOx/2ZS catalyst exposed to a 750 ppm NH3 b 750 ppm NO with 5 vol% O2 c 750 ppm NO with 5 vol % O2 after NH3 adsorption, and d 750 ppm NH3 after NO + O2 adsorption at 250 °C

Full size image

Figure 9b shows the in-situ DRIFTs spectra of NO + O2 co-adsorption on the as-prepared Ha-FeOx/2ZS catalyst at 250 °C. The nitrates species for gaseous NO2 (1606 cm− 1), nitrites (1346 cm− 1), monodentate nitrates (1292 and 1564 cm− 1) and bidentate nitrates (1514 cm− 1) can be observed with increasing time intervals (Xue et al. 2021). The intensity of monodentate nitrates (1292 and 1564 cm− 1) substantially increased after 5 min, suggesting that the NO can be adsorbed and oxidized on the Ha-FeOx/2ZS surface.

Figure 9c shows the in-situ DRIFTS spectra of the interaction between NO + O2 and pre-adsorbed NH3 species at 250 °C over the Ha-FeOx/2ZS catalyst. For Lewis acid sites, the band at 1107 cm− 1 decreased a little after the introduction of NO + O2, and the bands at 1281 cm− 1 and 1620 cm− 1 showed a decrease trend and vanished in 15 min, indicating that the coordinated NH3 species bound to Lewis acid sites were reduced. The band at 1303 cm− 1 (monodentate nitrates) began to accumulate, which suggested that monodentate nitrates and NH3 reacted through the Langmuir-Hinshelwood (L-H) mechanism at Lewis acid sites. The peak at 1640 cm− 1 related to gaseous NO2 accumulated gradually at the same time. According to an Eley-Rideal (E-R) mechanism, the gaseous NO2 would interact with the absorbed NH3 species and then accumulate on the catalyst surface. Moreover, the peaks of NH3 adsorbed on the Brønsted acid sites (1392–1512 cm− 1) vanished after 20 min, illustrating that Brønsted acid site also took part in the SCR reaction. The bands at 1514 and 1362 cm− 1 were covered by newly formed peaks and showed a marked increase, which might be due to the accumulation of the nitrate species (Pu et al. 2022). These results illustrate that the NH3-SCR reaction on the Ha-FeOx/2ZS proceeded through both the L-H and E-R mechanisms.

Figure 9d shows the in-situ DRIFTs spectra of the interaction between NH3 and pre-adsorbed NO + O2 at 250 °C. After feeding NH3, only the band at 1514 cm− 1 belonging to bidentate nitrates was unchanged, while the bands at 1624, 1564, 1346 and 1292 cm− 1 attributed to gaseous NO2, nitrites and monodentate nitrates rapidly disappeared proving that bidentate nitrates were not the active intermediates, whereas the gaseous NO2, nitrites and monodentate nitrates reacted with NH3 actively. Notably, the NH3 adsorbed on Lewis acid sites at 3334, 3269 and 1610 cm− 1 and the Brønsted sites at 1453 cm− 1 can be rapidly observed at the first 1 min when the NO + O2 pre-adsorbed catalysts were exposed to NH3. Compared with Fig. 9c, the rapid reaction between the adsorbed nitrate species and the adsorbed NH3 species firmly suggests that the NO reduction over Ha-FeOx/2ZS easily proceeded via the L-H mechanism, which is the dominant reaction route for NH3-SCR reaction at low temperature over ZS-derived catalysts.

According to above-mentioned results, a possible SCR deNOx mechanism over Ha-FeOx/2ZS catalysts was proposed as displayed Fig. 10. Firstly, NH3 can be absorbed and activated to be reactive -NH2 and NH4+ intermediates at Lewis and Brønste acid sites, respectively (Eqs. (3) and (4)). Meanwhile, the heteroatoms in the catalyst improved the amount of oxygen vacancies, promoting the oxidation of physically adsorbed NO by Mn+ and forming nitrites and nitrates species via Eqs. (6), (7) and (8), which are beneficial for the reaction in Eqs. (10), (11) and (12) (Zhang et al. 2021a, b; Song et al. 2020; Chen et al. 2021). Finally, the activated NH3 and NOx species are reacted to be N2 and H2O (9)–(12). Moreover, owing to the strong redox interaction between Fe with heteroatoms as shown Eqs. (14), which promotes the reducibility of Fe species at low temperatures, thereby further facilitating its low temperature reactivity.

Fig. 10
figure 10

Scheme represents for the NH3-SCR mechanism over the Ha-FeOx/2ZS catalyst

Full size image
$$\text{NH}_3\left( \text{g} \right) \to \text{NH}_3\left( \text{a} \right) \to \text{L} - \text{NH}_2\left( \text{a} \right)$$
(3)
$$\text{NH}_3\left( \text{g} \right)\, + \,\text{B} - \text{H} \to \text{NH}_4^ + \left( \text{a} \right)\,\left( \text{B} \right)$$
(4)
$$\text{NO}\left( \text{g} \right) \to \text{NO}\left( \text{a} \right)$$
(5)
$$\text{Fe}^{3 + }\, + \,\text{NO}\left( \text{a} \right) \to \text{Fe}^{2 + } - \text{O} - \text{NO}$$
(6)
$$\text{Fe}^{2 + } - \text{O} - \text{NO} + \,1/2{\text{O}_2}\left( \text{a} \right) \to \text{Fe}^{2 + } - \text{O} - \text{NO}_2$$
(7)
$$\text{Fe}^{2 + } - \text{O} - \text{NO} + \,1/2{\text{O}_2}\left( \text{a} \right) \to \text{Fe}^{2 + } = {\left( \text{O} \right)_2} - \text{NO}$$
(8)
$$\text{NH}_2\left( \text{a} \right) + \,\text{NO}\left( \text{g} \right) \to {\text{N}_2} + \,{\text{H}_2}\text{O}$$
(9)
$$\eqalign{{\text{Fe}^{2 + }} - {\text{O}} - {\text{NO}}{\mkern 1mu} + {\mkern 1mu} {\rm{N}}{{\rm{H}}_3}\left( {\rm{a}} \right) \to {{\rm{M}}^{\left( {n - 1} \right) + }} }$$
(10)
$$\eqalign{- {\rm{O}} - {\rm{NO}} - {\rm{N}}{{\rm{H}}_3} \to {\rm{F}}{{\rm{e}}^{2 + }}}$$
$$\eqalign{- {\rm{OH}}{\mkern 1mu} + {\mkern 1mu} {{\rm{N}}_2} + {\mkern 1mu} {{\rm{H}}_2}{\rm{O}} }$$
$$\eqalign{{\rm{F}}{{\rm{e}}^{2 + }} - {\rm{O}} - {\rm{N}}{{\rm{O}}_2} + {\mkern 1mu} {\rm{N}}{{\rm{H}}_3}\left( {\rm{a}} \right) \to {{\rm{M}}^{\left( {n - 1} \right) + }}}$$
(11)
$$\eqalign{- {\rm{O}} - {\rm{N}}{{\rm{O}}_2} - {\rm{N}}{{\rm{H}}_3} \to {\rm{F}}{{\rm{e}}^{2 + }}}$$
$$\eqalign{- {\rm{OH}}{\mkern 1mu} + {\mkern 1mu} {{\rm{N}}_2}{\rm{O}}{\mkern 1mu} + {\mkern 1mu} {{\rm{H}}_2}{\rm{O}} \cr}$$
$$\eqalign{ & {\rm{F}}{{\rm{e}}^{2 + }} = {\left( {\rm{O}} \right)_2} = {\rm{N}}{{\rm{O}}_2} + {\mkern 1mu} {\rm{N}}{{\rm{H}}_3}\left( {\rm{a}} \right) \to {\rm{F}}{{\rm{e}}^{2 + }} \cr & = {\left( {\rm{O}} \right)_2} = {\rm{N}}{{\rm{O}}_2} - {\rm{N}}{{\rm{H}}_3} \cr} $$
(12)
$$\text{Fe}^{2 + } - \text{OH}\, + \,{\text{O}_2}\, \to 2{\rm{ }}\text{Fe}^{3 + } = \text{O}\, + \,{\text{H}_2}\text{O}$$
(13)
$$\text{Fe}^{3 + } + \,{\text{M}^{n + }}\,\text{Fe}^{2 + } + \,{\text{M}^{\left( {n - 1} \right) + }}$$
(14)

Notes: L: Lewis acid sites; B: Brønste acid sites; M: non-ferrous metals

4 Conclusions

In this study, hazardous zinc slag was high value-added utilized to prepare a new type of low-cost and efficient SCR catalyst via a facile sol-gel method. The prepared Ha-FeOx/yZS catalyst exhibited greater than 90% of NO removal efficiency and N2 selectivity, which was obviously superior to the reported pure FeOx, Fe9Ti1Ox and γ-Fe2O3 catalysts. In addition, compared with α-Fe2O3, Fe/Beta and Mn(0.2)-FeOx reported (> 300 °C), the Ha-FeOx/yZS showed better low temperature activity. The characterizations showed that the intensive interactions among Fe and heteroatoms inhibited the crystal growth, leading to the prepared catalyst possessing amorphous morphology with large BET surface area, plenty of adsorbed oxygen species, and abundant acid sites. These characteristics were beneficial for the electron transfer between iron and other metal ions, thus promoting the oxidation-reduction cycle of SCR reaction. In this case, not only the high-content specie Fe in zinc slag was utilized as raw material to prepare the iron oxide-based catalyst, but the low-content species such as Si, Al, Mn, and Zn were also resourced as doped heteroatoms in the iron oxide matrix, and their synergy effect afforded the resultant catalyst exhibiting excellent reactivity. It is a new and efficient method for zinc slag treatment with high added-value, and shows promising application prospects in NOx removal by NH3-SCR, thereby reaching the goal of treating gas waste with solid waste.