Investigation of nitrous oxide decomposition over highly active and stable bimetallic CoFe-MOR zeolite catalyst: effective removal and mechanism study

Abstract

In this study, monometallic Co-mordenite (MOR) and bimetallic CoFe-MOR catalysts were prepared via simple wet ion exchange and tested for N₂O decomposition. Strong promotion effect of Fe on the activity and stability of Co ions in the zeolites was observed. To investigate the origin of this promotion effect, X-ray diffraction, H₂-temperature programmed reduction, UV-Vis spectroscopy, extended X-ray absorption fine structure analysis, and N₂O temperature-programmed desorption were used to characterize the bimetallic and monometallic catalysts. The characteristic results indicated that higher contents of Co ions located at β sites after Fe addition provided cooperation on N₂O splitting by two neighboring Co ions. Consequently, a greater amount of surface NO_x species were formed in situ and were more strongly bonded to the catalyst, facilitating the removal of O and increasing the activity. Moreover, extended X-ray absorption fine structure analysis indicated that β-type Co ions exhibited stronger coordination to framework oxygen after Fe addition, and higher exchange level was obtained in the bimetallistic CoFe-MOR. Both of them contribute to prevent the relocation of Co²⁺ ions to form cobalt oxides, thus, high activity was maintained. Consequently, the CoFe-MOR catalyst demonstrates a superior catalytic activity and a high durability in N₂O decomposition, showing great potential as a cost-effective catalyst for N₂O elimination in future applications.

---

1. Introduction

In recent years, extensive efforts have been devoted to developing effective technologies to control N₂O emissions from manufacturing plants since N₂O is considered to be responsible for the greenhouse effect and for contributing to ozone layer depletion.¹ The most economical one has been proposed to be catalytic decomposition of N₂O into environmentally friendly N₂ and O₂; several types of catalysts, for example, supported noble metals, mixed metal oxides, and metal-loaded zeolites have been proposed for N₂O catalytic decomposition.² Supported noble metal catalysts, such as Rh and Au, show high activities at 200–300 °C.^3,4 However, the high cost of noble metals limits their application. Although mixed metal oxides can efficiently decompose N₂O at the temperatures of 350–500 °C, unfortunately, their activities are severely inhibited by the presence of O₂ and NO,^5–7 which are usually unavoidable components in the tail-gas effluent emitted from e.g. the nitric acid plant. Good news is that many transition metal-exchanged zeolites, such as Fe,^8,9 Cu¹⁰ and Co zeolites,¹¹ show moderate activity in the temperature range of 450–600 °C and that N₂O decomposition over these zeolite catalysts is strongly promoted by NO.

Of all zeolite catalysts reported, iron-exchanged zeolites have probably attracted most attention due to their remarkable stability in the presence of H₂O.^8,9 This behavior strongly contrasts with other metal-zeolites (e.g. Cu-ZSM-5), which are greatly inhibited by H₂O, and suffers from serious deactivation under stream conditions.¹⁰ Lots of Fe-zeolites (ZSM-5, FER, BEA, FAU, etc.), although with different activity, have been examined and inevitably showed good durability under the simulated emission conditions; this good durability is attributed to the formation of framework Al–O–Fe species that stops the dealumination from the zeolite framework.¹² This feature makes the Fe-zeolites greatly attractive in their potential implementation in nitric acid plants. Consequently, extensive efforts have been devoted to prepare active Fe-zeolite catalysts, and several methods have been developed, including solid-state ion exchange,¹³ chemical vapor decomposition of FeCl₃,¹⁴ and isomorphous substitution.¹⁵ Compared with these methods, wet ion exchange (WIE) is simpler and more feasible for industrial applications, because it involves fewer steps, and the process parameters are easily controlled. However, the ion-exchange level is very low during WIE because of the long ion diffusion path and limited diffusion capacity of Fe ions, and catalysts with low Fe loading, i.e. low content of active sites, often result in insufficient activity under reaction conditions. Till now, none of the Fe-zeolites could show sufficiently high N₂O decomposition activities under realistic conditions below 500 °C, which restricts the application of this process in nitric acid plants with low-temperature tail-gases.¹

On the other hand, compared with Fe-zeolites, higher activities of Co-zeolites in N₂O decomposition under dry conditions are often reported.^16,17 However, in contrast with Fe-zeolites, moderate inhibition effect of H₂O is found on Co-zeolites, and they often suffer activity loss under rigorous reaction conditions (gas streams containing water at high temperatures).^18,19 Under streaming air, hydrated species of mono-atomic Co²⁺ on the zeolite exchange sites interact with un-exchanged protons remaining in the zeolites, leading to relocation of the Co cations and loss of dispersion and activity.²⁰ To achieve highly active and stable Co-zeolite catalysts, several methods have been proposed to improve the resistance of Co-zeolites to inhibition of H₂O. For example, van Bokhoven et al. has successfully prepared a Co-BEA zeolite with no adsorbing sites of H₂O molecules through calcinations of the as-synthesized BEA zeolite in an ammonium stream.^21,22 Our recent work has also demonstrated a full exchange of Co-MOR zeolites prepared by exchanging the MOR zeolites with a buffered cobalt solution at pH 8.²³ The current study indicates that bimetallic CoFe-MOR zeolites prepared via a simple wet ion-exchange method show high activity and good durability for N₂O decomposition, combining both the merits of Fe- and Co-zeolites. We have found that incorporation of Fe into Co-MOR catalysts significantly enhances the catalytic activity and stability of Co-species and strong synergistic effect existed between Fe and Co ions in the zeolites. X-Ray diffraction (XRD), H₂-temperature programmed reduction (H₂-TPR), UV-Vis spectroscopy, extended X-ray absorption fine structure analysis (EXAFS), and N₂O temperature-programmed desorption (N₂O-TPD) were used to characterize the bimetallic and monometallic catalysts in order to investigate the origin of this promotion effect. CoFe-MOR was able to decompose more than 90% of the N₂O at 420 °C under conditions simulating typical nitric acid emissions, e.g., 5000 ppm N₂O, 1000 ppm NO, 5% O₂, and 2% H₂O in a He balance. This high activity level continued for longer than 100 h under these strict conditions, showing great potential as a cost-effective catalyst for N₂O elimination in future applications.

2. Experimental

2.1 Catalyst preparation

Commercial H-MOR zeolite (Si/Al = 12; Sinopec Co.) was used as the parent zeolite. The Co-MOR catalyst was prepared by ion exchange of 3.0 g of H-MOR with 0.1 mol L⁻¹ Co(NO₃)₂ at room temperature for 24 h. After ion exchange, the sample was washed thoroughly with deionized water, dried at 80 °C, and calcined at 600 °C for 4 h. The calcined and un-calcined samples were denoted Co-MOR and Co-MOR-ex. Then, Fe was introduced into Co-MOR by ion exchange of 3.0 g of Co-MOR with 0.1 mol L⁻¹ Fe(NO₃)₃ at room temperature for 24 h. After thorough washing and drying, the solid was calcined and denoted CoFe-MOR. The un-calcined solid was denoted CoFe-MOR-ex.

The Fe-MOR catalyst was prepared by exchange of H-MOR with 0.1 mol L⁻¹ Fe(NO₃)₃ at room temperature. After ion exchange, the sample was washed thoroughly with deionized water, dried at 80 °C, and calcined at 600 °C for 4 h. Then, Co was introduced into Fe-MOR by ion exchange of 3.0 g of Fe-MOR with 0.1 mol L⁻¹ Co(NO₃)₂ at room temperature for 24 h. After thorough washing and drying, the product was calcined and denoted FeCo-MOR.

2.2 Material characterization

The Si, Al, Co, and Fe contents of these catalysts were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) using an Optima 2000 spectrometer. The X-ray diffraction patterns of all samples were measured on a Rigaku powder diffractometer (D/MAX-RB) using Cu Kα radiation (λ = 0.15418 nm) at a scanning rate of 4 ° min⁻¹ over the range 2θ = 10–80°. UV-Vis diffuse reflectance spectra (UV-Vis DRS) were recorded in air against BaSO₄ in the region of 12 500–24 000 cm⁻¹ on a Hitachi UV-3000 spectrometer.

Extended X-ray absorption fine structure (EXAFS) data of Co-MOR and CoFe-MOR were measured on beamline 4W1B at the Beijing Synchrotron Radiation Facility (BSRF). The data collection and analysis were similar to those in ref. 24 and 25 and a detailed description is given in the ESI.†

H₂-temperature programmed reduction (H₂-TPR) experiments were conducted on a Micromeritics Chemisorb 2720 apparatus. Details of the procedures are given in ref. 26. O₂-TPD and N₂O-TPD experiments were also conducted on the Micromeritics Chemisorb 2720 apparatus, using a procedure similar to that used for H₂-TPR. The difference was that prior to heating in a He stream to desorb adsorbed species, the catalyst was pretreated in O₂/He(10/90) or N₂O/He (2/98) at 573 K for 2 h and cooled to room temperature in the same atmosphere. The outlet gas composition was analyzed online using a quadrupole mass spectrometer (Omnistar Thermostar). Six masses characteristic of NO (3̲0̲), NO₂ (30, 4̲6̲), N₂O (28, 30, 4̲4̲), N₂ (2̲8̲), O₂ (3̲2̲), and He (4̲) were followed. The intensities of N₂ (28) and NO (30) were determined by solving a linear system of equations.

2.3 Activity measurements

N₂O decomposition experiments were performed using a fixed-bed flow microreactor at atmospheric pressure. In each run, 0.1 g of catalyst (0.25–0.425 mm) was placed into a quartz reactor (4 mm i.d.) and pretreated under a He stream at 600 °C for 1 h. After the reactor was cooled to 200 °C, the reactant gas mixture (5000 ppm of N₂O, with or without 5% O₂ or 700 ppm NO; He balance) was fed into the reactor. The total flow rate of the mixed gases was set at 60 mL min⁻¹ (GHSV = 30 000 h⁻¹). The outlet gas composition was analyzed online using a gas chromatograph (GC; Agilent 6820 series) equipped with a thermal conductivity detector and two serial columns: a Porapar Q column for the separation of N₂O and N₂/O₂, and a molecular sieve 5 Å column for the separation of N₂ and O₂. The steady-state N₂O conversion was calculated based on the peak area at a predetermined temperature, ascending from 200 to 600 °C at 25 °C intervals.

3. Results and discussion

3.1 Catalytic activity measurements

The cobalt and ferric loadings in the catalysts, in wt%, and atomic ratios are listed in Table 1. Co-MOR and CoFe-MOR have similar Co loadings (3.09% and 2.95%), indicating that incorporation of Fe (0.66%) into Co-MOR had no obvious effect on Co loading. However, almost no Co was loaded on the FeCo-MOR (0.12%), which mainly contains similar high content of Fe (2.63%) as in Fe-MOR (2.68%). This result suggests that no exchange sites remained available in Fe-MOR, of which the exchange level corresponded to 91%.

Table 1 Composition and characterization of catalysts

Catalyst	Co^a (%)	Fe^a (%)	Co/Al	Fe/Al	S BET/m^2 g^−1	Total exchange level^b (%)
a In wt%. b Full exchange level Co/Al = 0.5 or Fe/Al = 0.33.
Co-MOR	3.09	—	0.36	—	465	72
CoFe-MOR	2.95	0.66	0.34	0.07	450	89
FeCo-MOR	0.12	2.63	0.01	0.29	445	90
Fe-MOR	—	2.68	—	0.30	478	91

---

Fig. 1 shows the catalytic activities of these catalysts for direct N₂O decomposition under dry conditions. As shown, Co-MOR had much higher activity compared with Fe-MOR. This is in consistence with previous studies demonstrating that Co ions in zeolites possess higher activity than Fe ions for N₂O decomposition.^16,17 FeCo-MOR exhibited similar low activity as for Fe-MOR, which could be ascribed to its low Co content exchanged into the zeolite (Table 1). It is interesting to find that CoFe-MOR showed much higher activity compared with Co-MOR, which suggests that Fe has significant improving effect on the activity of Co ions in the zeolites, since they have similar Co content as revealed by the ICP results (Table 1). The improving effect of Fe is speculatively ascribed to the synergic effect between Co and Fe ions in the MOR zeolites, as reported over several zeolite supported bimetallic catalysts, such as Ru/Fe-FER,²⁷ Pt/Fe-FER,²⁸ Pd/Co-MOR¹⁴ and Pt/Co-ZSM-5.²⁹ These noble/transition bimetallic catalysts have been reported as very good catalysts for N₂O decomposition or NO_x reduction. However, the high cost of the noble metal limits their application. Moreover, the N₂O decomposition activity of CoFe-MOR is comparable with, or even better than, other zeolitic catalysts, including some noble/transition bimetallic zeolites (Table S1 of ESI†).^12,23,30,31 Such a catalyst (CoFe-MOR) with transition bimetals that can be prepared by the simple WIE method achieving a higher activity appears to be a promising catalyst candidate for future application.

Fig. 1 Conversion of N₂O over Co-MOR, CoFe-MOR, FeCo-MOR and Fe-MOR.

The high catalytic activity of CoFe-MOR for N₂O decomposition motivated us to investigate its performance under simulated conditions of typical nitric acid emission, including O₂, NO, and water vapor in the stream. O₂ had no obvious effect on N₂O decomposition, however, the presence of NO and water vapor showed similar negative effects on N₂O decomposition over CoFe-MOR. As shown in Fig. 2A, the T₉₀ for N₂O decomposition in the presence of NO and water shifted to higher temperatures by 10 and 20 °C, respectively. Moreover, it is found that no NO₂ formation is detected and NO concentration has no obvious changes during the whole reaction process, even in the presence of O₂. However, when O₂, NO, and water vapor were co-present in the stream, the total inhibition was not additive, but closed to the conversion in the presence of only one inhibition component: the T₉₀ for N₂O decomposition was achieved at 420 °C, representing an increase of 20 °C. This moderate inhibition suggests that the effects of O₂, NO, and H₂O molecules resulted primarily from their competitive adsorption on the active sites of the catalyst, and that these molecules may occupy the same kinds of active sites and thereby affect each other.^7,31 Therefore, CoFe-MOR performed well as a catalyst for N₂O decomposition under simulated emission conditions, achieving 50% and 90% N₂O conversion at 380 and 420 °C, respectively.

Fig. 2 Conversion of N₂O over CoFe-MOR (A) and Co-MOR (B) under different conditions: 0.1 g catalyst, 5000 ppm N₂O, 0 or 1000 ppm NO, 0 or 5% O₂, 0 or 2% H₂O and the balance He. GHSV = 30 000 h⁻¹; time on stream behavior of CoFe-MOR (C) and Co-MOR (D) for N₂O decomposition. Reaction conditions: 0.1 g of catalyst, 5000 ppm N₂O, 1000 ppm NO, 5% O₂, 2% H₂O and the balance He. GHSV = 30 000 h⁻¹.

In comparison, the presence of O₂, NO, and water vapor, especially water, had a greater inhibition effect on N₂O decomposition over Co-MOR. As shown in Fig. 2B, not only the N₂O decomposition was seriously inhibited in the presence of water, but also in the co-presence of O₂ and NO, the total inhibition increased significantly. This serious inhibition suggests that O₂, NO, and H₂O molecules have much stronger adsorption on the active sites of the Co-MOR catalyst, and their adsorption behavior may be different from that of CoFe-MOR, which will be discussed in the following section. Therefore, Co-MOR exhibited much lower activity under the simulated emission conditions, with 50% and 90% N₂O conversion occurring at 475 and 525 °C, respectively, which is ∼75–85 °C higher compared with CoFe-MOR.

We further tested the durability of CoFe-MOR catalysts under simulated conditions for a typical nitric acid manufacture. Time on stream profiles (Fig. 2C) showed that CoFe-MOR initially decomposed 71% and 92% of N₂O at 370 and 420 °C, respectively, and the conversion remained almost unchanged after reaction for more than 100 h. Remarkably, the high catalytic activity for N₂O decomposition and good durability of CoFe-MOR were comparable to or even superior to the most stable Fe zeolite catalyst.³¹ Thus, CoFe-MOR shows potential as a cost-effective catalyst for N₂O elimination from manufacturing plant exhaust. Moreover, this good durability of CoFe-MOR differed greatly from that of the Co-MOR catalyst. As shown in Fig. 2D, the initial N₂O conversion over Co-MOR suffered from activity loss of 10–20% under the same conditions after running for 100 h. From the above results, it could be concluded that incorporation of Fe not only improves the activity of Co-MOR under dry conditions, but also significantly improves the durability of Co ions in MOR under the simulated conditions. Similar promotion effect has been reported over Pt/Co-ZSM-5 used for NO_x reduction.²⁹ It is found that the introduction of Pt improved the catalytic activity and stability by changing the locations of the Co²⁺ ions in the zeolite channels. Moreover, our previous work has demonstrated that Co ions are the active sites for N₂O decomposition; however, Co ions at different exchange sites exhibit different activity. Therefore, we deduce that the addition of Fe promotes the activity of the catalyst by affecting the chemical states of Co²⁺ species in MOR. To obtain evidence to support this assumption, a series of characterization was carried out on CoFe-MOR and Co-MOR.

3.2 Cobalt species formed in the zeolites

As shown in Fig. S1 (ESI†), the XRD patterns of these mono- and bimetallic catalysts exhibited typical diffraction peaks corresponding to the characteristic MOR structure.³² In addition, no obvious diffraction peaks corresponding to crystallized Co₃O₄³³ were observed for Co-MOR or CoFe-MOR at cobalt loadings up to 3 wt%, suggesting good dispersion of cobalt species on the internal/external surface. The surface areas of the catalysts (450–465 m² g⁻¹) were similar and close to that of the parent H-MOR (475 m² g⁻¹), indicating that the structure of the parent zeolite was preserved after ion exchange and calcinations and the addition of Fe also did not change the specific surface areas significantly.

Fig. 3 shows the H₂-TPR patterns of Co-MOR, CoFe-MOR and Fe-MOR. For Co-MOR, an intense reduction peak was observed at the high temperature region (675–900 °C), which is ascribed to the reduction of mono-atomic Co²⁺ on ion exchange sites.^11,20 Moreover, a weak H₂ consumption evident at 200–400 °C indicated the presence of a small amount of Co₃O₄ particles on the external surface of the zeolite crystal.^34–36 For Fe-MOR, two weak H₂ consumption peaks centered at 650 and 710 °C were observed, and they are attributed to the reduction of isolated Fe ions and Fe oxides in the zeolites, respectively.³⁷ However, CoFe-MOR showed primary H₂ consumption at high temperatures, which implies that introduction of Fe makes Co exist predominantly as mono-atomic Co²⁺ on ion exchange sites, with much limited aggregation of Co₃O₄ particles compared with Co-MOR. This result indicates synergic effect between Co and Fe ions in the MOR zeolites and incorporation of Fe into Co-MOR greatly increased isolated Co²⁺ ion formation on the ion exchange sites.

Fig. 3 H₂-TPR profile of (a) Co-MOR, (b) CoFe-MOR and (c) Fe-MOR.

Moreover, it should be noted that not only CoFe-MOR exhibited more predominant mono-atomic Co²⁺ ion reduction at the high temperature region than that of Co-MOR, but there was a noticeable difference in their H₂ reduction peaks; the CoFe-MOR peak was 10 °C higher and much sharper. This different pattern of H₂ reduction of Co²⁺ ions suggests that the position and bonding of mono-atomic Co²⁺ ions may be affected by subsequent incorporation of Fe ions as a result of synergic interactions between Fe and Co. In our previous work we have found that Co(II) located at different sites of the zeolites with varied coordination states would exhibit greatly different activity for N₂O decomposition. Here, the difference in chemical status of mono-atomic Co²⁺ ions in Co-MOR and CoFe-MOR may be a reason for the higher activity and stability obtained for CoFe-MOR. Therefore, to further clarify their roles, UV-Vis DRS and Co K-edge EXAFS experiments, which are demonstrated to be an effective technique to quantitatively analyze the sitting and coordination of Co ions, were conducted on Co-MOR and CoFe-MOR.^38–40 Moreover, the spectra of Co-MOR-ex and CoFe-MOR-ex, as-prepared Co-MOR and CoFe-MOR after ion exchange without further calcinations, were also included to fully understand the migration and formation of Co²⁺ species, since the previous reports have demonstrated that the ion migration in the preparation process mainly occurred during the ion exchange and subsequent calcinations step.^36,41

3.3 Co ions sitting and bonding in the zeolites

3.3.1 UV-Vis DRS spectra of Co zeolite catalysts. The UV-Vis DRS spectra of the ion exchanged and calcined Co-MOR and CoFe-MOR samples are presented in Fig. 4. The strong absorbance at 12 500–24 000 cm⁻¹ indicates the presence of various Co²⁺ species in the Co-zeolite samples.^38–40 According to the literature, several characteristic UV-Vis adsorption bands were extracted by spectral deconvolution and attributed to specific isolated Co²⁺ ions located at the typical cationic sites (referred to as α, β, and γ) in the MOR zeolite channels and other cobalt oxides, such as CoO_x clusters and Co₃O₄ nanoparticles.^38–40 Their distributions were quantitatively analyzed by calculating the integrated intensities of their spectral bands according to the procedures described in ref. 38–40.

Fig. 4 UV-Vis spectra of Co-MOR-ex (A) Co-MOR (B), CoFe-MOR-ex (C) and CoFe-MOR (D).

The UV-Vis spectral results (Fig. 4 and Table 2) reveal that (i) the bands corresponding to isolated Co²⁺ ions of CoFe-MOR (89) are stronger than those of Co-MOR (79); (ii) for isolated Co²⁺ ions, a larger population of β-type Co ions formed in CoFe-MOR (84) than in Co-MOR (75); (iii) for Co-MOR-ex and CoFe-MOR-ex more α-type Co²⁺ ions and less cobalt oxides are formed compared with their calcined samples; (iv) after exchange with Fe, CoFe-MOR-ex contains higher content of isolated Co²⁺ ions compared with Co-MOR. These results inevitably indicate that Co ions sitting in MOR have migrated in the process of ion exchange with Fe and subsequent calcinations. Thus, CoFe-MOR contained not only a higher content of isolated Co²⁺ ions but also more Co ions at the β sites. These β-type Co ions, located in the interconnected small channels of MOR, exhibit stronger bonding to framework oxygen compared with α-type Co ions.^38–40 The increasing population of β-type Co ions in CoFe-MOR would significantly alter the accessibility and bonding properties of the Co ions, and this is likely to be very important in determining their catalytic activities.

Table 2 Percentage of subband areas derived by deconvoluting UV-Vis spectra and corresponding Co species loading

Catalyst	I α ^a	I β ^a	I γ ^a	I Co ^2+ ^a	I Co3O4 ^a	I CoO ^a	Coα^b	Coβ^b	Coγ^b	Co^2+^b	Co3O4^b	CoO^b
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
Band assignments: Coα: 14 800 cm^−1; Coβ: 15 900, 17 500, 19 200 and 21 100 cm^−1; Coγ: 20 150 and 22 050 cm^−1; Co3O4: 13 700 cm^−1; CoO: 19 600 cm^−1.a Calculated from the data in Fig. 4.b In wt%.
Co-MOR-ex	9	74	0	83	0	17	0.28	2.41	0.01	2.7	0	0.53
Co-MOR	3	75	1	79	21	0	0.1	2.32	0.04	2.46	0.66	0
CoFe-MOR-ex	5	68	11	84	6	10	0.16	2.03	0.32	2.51	0.17	0.3
CoFe-MOR	3	84	2	89	2	9	0.08	2.48	0.05	2.61	0.06	0.28

---

3.3.2 EXAFS of Co zeolite catalysts. The k³-weighted and Fourier transform spectra of Co K-edge EXAFS for exchanged and calcined Co-MOR and CoFe-MOR are given in Fig. 5. The structural parameters of the EXAFS fitting are presented in Table S2 (ESI†). The fitting results show that the first coordination shell of the adsorbed Co consisted of 4.1 oxygen atoms at a distance of 2.09 Å and 1.2 oxygen atoms at a distance of 2.87 Å for Co-MOR (Table 3).⁴² The Co–O distance of 2.09 Å, which is typical for a Co–O bond, represents isolated Co²⁺ ions located in the cationic-exchange positions of MOR, while the distance of 2.87 Å suggests that 1.2 distant oxygens formed the next neighboring shell. This result is consistent with Drozdová's studies, which found that α-type Co ions in MOR were strongly coordinated to three or four framework oxygens with a Co–O bond distance of 2.06 Å, and β-type Co ions coordinated with two oxygens with a distant attraction of 2.85 Å in addition to four framework oxygens at a distance of 2.06 Å.⁴⁰ The number of coordinated oxygens (5.3) for Co-MOR suggests that it contained β-type Co ions (typically four direct and two distant oxygens) in addition to other components such as α-type Co ions (typically, three or four direct oxygens). Moreover, the Co–Co coordination with 0.5 atoms at a distance of 3.12 Å indicates the presence of a small amount of CoO_x clusters or Co₃O₄ particles.

Fig. 5 (A) k³-weighted observed (—) and fitted ([dash dash, graph caption]) Co K-edge EXAFS spectra of (a) Co-MOR-ex, (b) Co-MOR, (c) CoFe-MOR-ex, (d) CoFe-MOR and (e) CoO. (B) Cobalt RSFs (radial structural functions) for (a) Co-MOR-ex, (b) Co-MOR, (c) CoFe-MOR-ex, (d) CoFe-MOR and (e) CoO. The peak positions are uncorrected for phase shifts.

Table 3 Summary of Co K-edge EXAFS results for Co-MOR and CoFe-MOR catalyst^a

Catalyst	Co–O(1)			Co–O(2)			Co–Co			Res.
	CN	R/Å	σ ^2	CN	R/Å	σ ^2	CN	R/Å	σ ^2
a The amplitude reduction factor (S0^2) was fixed to 0.9 in the fitting process.^42
Co-MOR-ex	5.2	2.08	0.008							13.5
Co-MOR	4.1	2.09	0.009	1.2	2.87	0.01	0.5	3.12	0.01	7.0
CoFe-MOR-ex	6.0	2.09	0.015							13.2
CoFe-MOR	4.6	2.09	0.007							8.3

---

However, before calcinations the Co-MOR-ex sample exhibits much different coordination and bonding property compared with Co-MOR. For Co-MOR-ex, the best fit was obtained with a model consisting only the first Co–O coordination shell of 5.6 oxygen atoms at a distance of 2.08 Å. The absence of Co–Co coordination excludes the presence of large cobalt oxides and suggests mainly isolated Co(II) ions in Co-MOR-ex. This result is in line with the UV-Vis results that after ion exchange the Co ions in MOR are mainly the mixture of α-type Co ions and β-type Co ions; after calcinations, parts of α-type Co ions migrated to β sites and even formed Co₃O₄ particles in the calcinated Co-MOR catalyst. However, after exchange with Fe, the Co ions present again as typical isolated Co ions with coordination to 6.0 oxygen atoms at the distance of 2.08 in CoFe-MOR-ex. This result suggests that synergistic effect exists between Fe and Co, which makes the Co ions in oxides relocated into isolated ions sites with six coordinated H₂O or OH. This synergistic effect has also been found between Co²⁺ and Pd²⁺, the improved thermodynamic affinity of the isolated Pd²⁺ to the zeolite was found in the presence of isolated Co²⁺ due to electrons effect.⁴³ Similar effect has also been found in Ir/In-zeolite,⁴⁴ in which the incorporated In and Ir cations act freely with zeolites to reach intimate sitting in the zeolite cavities.

After calcinations, the first Co–O coordination number of CoFe-MOR decreased to 4.6 oxygen atoms at a distance of 2.09 Å. The UV-Vis results indicate that Co ions are mainly formed as β-type Co ions in CoFe-MOR. However, the EXAFS results of CoFe-MOR do not show the typical β-type Co ions coordination (four direct and two distant oxygens). While the higher coordination number (4.6) at the first shell implies small number of α-type Co ions and suggests stronger bonding to the framework oxygens. Combining the strong bonding property and UV-Vis results probably implies that the Co ions in CoFe-MOR are coordinated directly with more than four, maybe five, oxygens in the β sites. They are stronger coordinated to framework oxygens compared with four direct and one distant coordination for β sites in Co-MOR due to the presence of Fe ions. Combining the characteristic results of UV-Vis and EXAFS, a schematic diagram was drawn to outline the formation of various Co species in the CoFe-MOR catalyst and their transformation in sitting and bonding property during the Co exchange, calcinations, Fe exchange and consequent calcinations steps (shown in the Graphical Abstract).

3.4 Ferric species formed in the zeolites

The sitting of Fe ions in the exchanged or calcinated CoFe-MOR could not be detected due to the low content of Fe in the zeolites. However, according to our characteristic results, Fe exchanged into the zeolites would form as isolated Fe³⁺ ions on the ion exchange sites.⁴⁵ Moreover, according to the rule proposed by Mortier⁴⁶ and Takaishi et al.,⁴⁷ exchanged cations are first located at the most released site, i.e. the α site in MOR. Taking into account the low content of Fe exchanged to the zeolites and the little β sites remaining in Co-MOR, it is reasonable to conclude that Fe ions would locate mostly at the α sites. Consequently, the location and coordination of Co in the zeolites are affected by Fe ions that are intimately located, with the zeolite acting as a solid solvent.^48,49

3.5 Role of Fe in improving the activity and stability of Co(II) in MOR

3.5.1 Mechanism of Fe in improving the activity. In regard to the N₂O decomposition mechanism, different reaction pathways have been proposed.^50–52 In general, N₂O chemisorption first takes place on the active site of the catalyst, followed by the N₂ release and the formation of chemisorbed oxygen on the active site (R₁). The chemisorbed oxygen can then be desorbed as O₂ (rate determining step) via recombining two chemisorbed oxygens (R₂) and/or reacting with gaseous N₂O (R₃). Alternatively, parts of N₂O could be splitted by two neighboring active sites to form NO and the in situ formed NO reacts with chemisorbed oxygen to form N₂ and O₂ (R₄). It is demonstrated that this NO assisted N₂O decomposition (R₄) would greatly increase the activity and such mechanism has been reported to be prevailing in several zeolite systems, such as Fe-FER⁵³ and Co-MOR.²³ Our previous work has found that in MOR, β-type Co ions are the most active sites for this NO assisted N₂O decomposition due to the formation of two unique β-type Co ions in Co⋯Co pairs for cooperation on the N₂O splitting. Coincidentally, the above characteristic results have shown increasing population of β-type Co ions in CoFe-MOR, and this is consistent with the observed increase in activity compared with Co-MOR. Thus, we could deduce that the promotion effect of Fe addition is mainly attributed to redistribution of more Co ions at active β sites, i.e. more active sites for NO assisted N₂O decomposition. To verify such decomposition mechanism and to reveal the effects of Fe addition, a series of N₂O-TPD experiments were conducted on CoFe-MOR and Co-MOR.

N₂O + * → N₂ + * −O (R1)

2 * −O → O₂ + 2* (R2)

N₂O + * −O → N₂ + O₂ + * (R3)

N₂O + * → * −NO + N; * −NO + * −O → * −NO₂ + *; * −NO₂ + * −O → * −NO + O₂ + * (R4)

As shown in Fig. 6, a much greater amount of NO_x species desorbed from CoFe-MOR than from Co-MOR after treatment with N₂O at 300 °C. More significantly, the surface NO_x species were more strongly bonded over CoFe-MOR, as indicated by a substantial shift in the desorption temperature of NO to nearly 400 and 550 °C; for Co-MOR, the desorption temperature was centered at 350 °C. This could be directly correlated with the established role of surface NO_x species in the N₂O decomposition.^54,55 Thus, CoFe-MOR exhibited higher activity with a greater amount of NO taking part in the decomposition. Moreover, it should be noted that a considerable amount of O₂ started to desorb from CoFe-MOR at a much lower temperature (280 °C) than that of Co-MOR (320 °C). This temperature agrees well with their activity performance for N₂O decomposition (Fig. 1), in which N₂O starts to decompose at 280 and 315 °C for CoFe-MOR and Co-MOR, respectively. This result combined with the NO formed clearly proves that more NO is formed over CoFe-MOR, and further facilitates the removal of adsorbed atomic oxygen, i.e. the rate-determining step for N₂O decomposition, thus enhancing the activity. One may suspect that this O₂ species has come from chemisorbed oxygen reacting with N₂O (R₃) or even deposited chemisorbed O₂ (R₂). However, according to the literature, NO is much more effective in scavenging atom oxygen than N₂O as a result of higher activity obtained for NO-assisted N₂O decomposition than decomposition of N₂O alone.^54,55 Furthermore, the O₂-TPD profiles of CoFe-MOR and Co-MOR pretreated with O₂ (as shown in Fig. 7) clearly show that there is no obvious O₂ desorption below 400 °C; however the N₂O decomposition was already evident on both of them at this temperature region: CoFe-MOR and Co-MOR give 90% and 50% N₂O conversion at 400 °C, respectively (as shown in Fig. 1). This result excludes the possibility of desorption of chemisorbed O₂ in this temperature region and O₂ removal through recombination of two atomic oxygens could hardly happen, leaving R₄, i.e. NO assistant N₂O decomposition, as the most possible reaction mechanism over both CoFe-MOR and Co-MOR. This result also supports our analyses and activity results as shown above, and we conclude that the mechanism of Fe in improving the activity could be attributed to more Co ions located at active β sites, providing more active sites for this NO assisted N₂O decomposition.

Fig. 6 N₂O-TPD in He over CoFe-MOR (A) and Co-MOR (B).

Fig. 7 O₂-TPD profile of (a) Co-MOR, (b) CoFe-MOR and (c) Fe-MOR.

3.5.2 Mechanism of Fe in improving the stability. The high activity of Co-zeolites for N₂O decomposition has been generally recognized, however, they suffer from serious activity loss under the streaming conditions. On the other hand, Fe-zeolites possess super stability although their activities are lower than Co-zeolites. To combine the merits of both Fe- and Co- zeolites, we attempted to prepare the bimetallic CoFe-MOR zeolites. Fortunately, we have found that this CoFe-MOR catalyst shows not only high activity, but also good durability for N₂O decomposition under the simulated conditions (Fig. 1 and 2). It is known that there are two main reasons for the activity loss of Co²⁺ ions. One is formation of cobalt oxides and loss of dispersion upon interacting isolated Co²⁺ cations with un-exchanged protons remaining in the zeolites.²⁰ The other is dealumination, i.e. removal of the tetrahedral Al³⁺ ion from the zeolite lattice and destruction of the zeolite structure. As shown in Fig. 2D, Co-MOR suffers greatly from activity loss under the simulated conditions, however, after Fe addition, the activity of CoFe-MOR remained almost unchanged for at least 100 h. The super stability has been reported on lots of Fe-zeolites, and found mainly attributed to the formation of framework Al–O–Fe species that stops the dealumination from the zeolite framework.¹² CoFe-MOR has a considerable amount of Fe (Fe/Al ratio of 0.07), therefore the formation of framework Al–O–Fe species may play a role in stopping the dealumination from the zeolite framework. However, this role could not be major since the Fe/Al ratio is quite low compared to the whole zeolites.

On the other hand, the fitting of EXAFS spectra over CoFe-MOR suggests that Co ions in CoFe-MOR were coordinated directly with more than four, maybe five, oxygens at β sites. They were more strongly coordinated to framework oxygen compared with the four direct and one distant oxygen for β sites in Co-MOR due to the presence of Fe ions. This is consistent with the H₂-TPR results, which showed that after incorporation of Fe, the reduction of Co ions shifted to higher temperatures because of their stronger coordination environment. This stronger coordination would not only restrict the Co ions stabilized at the less accessible β sites but also render them much lesser reactive to water vapor or other inhibitory molecules. Consequently, these Co ions are more stable under the reaction situation and thus much higher activity stability of CoFe-MOR was obtained than Co-MOR under the same stream conditions. However, for Co-MOR, under stream conditions, Co²⁺ ions may hydrolyze, and the hydrolyzed Co²⁺ species would further interact with the nearest Brönsted acid on the zeolite matrix, leading to the formation of small oxide crystallites and causing irreversible activity loss, which is a general phenomenon for Co exchanged zeolite.^23,40,41 In our previous work, we demonstrated that a higher exchange level with a few un-exchanged protons remaining in the zeolite greatly improved the stability of Co-MOR. In the case of CoFe-MOR, the total exchange level (89%) is also higher than that of Co-MOR (72%). The NH₃-TPD results of Co-MOR and CoFe-MOR (Fig. S2, ESI†) also support that CoFe-MOR has a much lower content of Brönsted acid sites, mainly caused by un-exchanged protons compared with Co-MOR. Thus, we conclude that the role of Fe in improving the activity stability of Co²⁺ ions is mainly attributed to additional Co ions in β sites with much stronger coordination to framework oxygen and higher exchange level with less un-exchanged protons after Fe addition, which would inevitably prevent the relocation of Co²⁺ ions to form cobalt oxides.

The Co species formed in CoFe-MOR and Co-MOR after the stability test was also investigated by H₂-TPR. As shown in Fig. 8(b), the Co-MOR catalyst showed one strong reduction peak in the range of 500–600 °C, indicating the relocation of Co species from highly dispersed mono-atomic Co²⁺ into CoO_x clusters in the zeolite channels. On the other hand, for CoFe-MOR (Fig. 8(a)), a large quantity of isolated Co ions was still located at exchange sites after reaction for 100 h under simulated emission conditions, proving that the redistribution of Co²⁺ ions into cobalt clusters was mostly stopped after Fe incorporation.

Fig. 8 H₂-TPR profile of used catalysts: (a) CoFe-MOR and (b) Co-MOR. After reaction for 100 h under the conditions of 5000 ppm N₂O, 1000 ppm NO, 5% O₂, 2% H₂O and the balance He. GHSV = 30,000 h⁻¹.

4. Conclusions

Strong synergic effect between Co and Fe ions was found in the MOR zeolites, and the activity and stability of Co ions in the zeolites were greatly enhanced after Fe addition. Bimetallic CoFe-MOR catalyst was able to decompose more than 90% of the N₂O at 420 °C under conditions simulating typical nitric acid emissions, e.g., 5000 ppm N₂O, 1000 ppm NO, 5% O₂, and 2% H₂O in a He balance. This high activity level continued for longer than 100 h under these strict conditions, showing great potential as a cost-effective catalyst for N₂O elimination in future applications. Based on the characteristic results, the excellent activity of bimetallic CoFe-MOR was attributed primarily to the higher content of Co ions located at active β sites, which provide cooperation on N₂O splitting by two neighboring Co ions. Thus, a greater amount of strongly bonded surface NO_x species were formed in situ during N₂O decomposition, facilitating the removal of O, i.e., the rate-determining step for N₂O decomposition. Moreover, the EXAFS results indicate that these β-type Co ions exhibited stronger coordination to framework oxygen and a higher exchange level was obtained after Fe addition, which prevent relocation of Co²⁺ ions into CoO_x clusters and maintain the high activity.

Acknowledgements

This work was financially supported by National Natural Science Funds for Distinguished Young Scholar (20725723) and National Basic Research Program of China (2010CB732300). We thank Beijing Synchrotron Radiation Facility (BSRF, China) for providing the beam time and help in EXAFS analysis.

References

1. J. Pérez-Ramírez, F. Kapteijn, K. Schöffel and J. A. Moulijn, Appl. Catal., B, 2003, 44, 117–151.
2. F. Kapteijn, J. Rodríguez-Mirasol and J. A. Moulijn, Appl. Catal., B, 1996, 9, 25–64.
3. G. Centi, S. Perathoner, F. Vazzana, M. Marella, M. Tomaselli and M. Mantegazza, Adv. Environ. Res., 2000, 4, 325–338.
4. L. Yan, X. Zhang, T. Ren, H. Zhang, X. Wang and J. Suo, Chem. Commun., 2002, 860–861.
5. P. Stelmachowski, G. Maniak, A. Kotarba and Z. Sojka, Catal. Commun., 2009, 10, 1062–1065.
6. Z. P. Xu and H. C. Zeng, J. Mater. Chem., 1998, 8, 2499–2506.
7. L. Obalová, K. Jirátová, F. Kovanda, K. Pacultová, Z. Lacný and Z. Mikulová, Appl. Catal., B, 2005, 60, 289–297.
8. J. Pérez-Ramírez, F. Kapteijn, G. Mul and J. A. Moulijn, Appl. Catal., B, 2002, 35, 227–234.
9. J. Pérez-Ramírez, A. G. Hevia Miguel and S. Abelló, Chem. Commun., 2008, 5351–5353.
10. P. J. Smeets, B. F. Sels, R. M. van Teeffelen, H. Leeman, E. J. M. Hensen and R. A. Schoonheydt, J. Catal., 2008, 256, 183–191.
11. P. J. Smeets, Q. G. Meng, S. Corthals, H. Leeman and R. A. Schoonheydt, Appl. Catal., B, 2008, 84, 505–513.
12. L. D. Li, Q. Shen, J. J. Yu, Z. P. Hao, Z. P. Xu and G. Q. Max Lu, Environ. Sci. Technol., 2007, 41, 7901–7906.
13. K. Sun, H. Xia, E. Hensen, R. van Santen and C. Li, J. Catal., 2006, 238, 186–195.
14. G. D. Pirngruber, J. Catal., 2003, 219, 456–463.
15. J. Pérez-Ramírez, J. C. Groen, A. Brückner, M. S. Kumar, U. Bentrup, M. N. Debbagh and L. A. Villaescusa, J. Catal., 2005, 232, 318–334.
16. J. Pérez-Ramírez, F. Kapteijn, G. Mul and J. A. Moulijn, Kinet. Catal., 2003, 44(5), 639–647.
17. F. Kapteijn, G. Marbán, J. Rodriguez-Mirasol and J. A. Moulijn, J. Catal., 1997, 167, 256–265.
18. M. C. Campa, V. Indovina and D. Pietrogiacomi, Appl. Catal., B, 2009, 91, 347–354.
19. L. Gutierrez and E. A. Lombardo, Appl. Catal., A, 2009, 360, 107–119.
20. A. V. Boix, S. G. Aspromonte and E. E. Miró, Appl. Catal., A, 2008, 341, 26–34.
21. L. Capek, J. Dedecek and B. Wichterlova, J. Catal., 2004, 227, 352–366.
22. J. A. van Bokhoven, D. C. Koningsberger, P. Kunkeler, H. van Bekkum and A. P. M. Kentgens, J. Am. Chem. Soc., 2000, 122, 12842–12847.
23. X. Y. Zhang, Q. Shen, C. He, Y. F. Wang, J. Cheng and Z. P. Hao, J. Hazard. Mater., 2011, 192, 1756–1765.
24. G. Z. He, G. Pan, M. Y. Zhang and G. A. Waychunas, Environ. Sci. Technol., 2011, 45, 1873–1879.
25. G. Z. He, G. Pan, M. Y. Zhang and G. A. Waychunas, J. Phys. Chem. C, 2009, 113, 21679–21686.
26. L. D. Li, Q. Shen, J. J. Li, Z. P. Hao, Z. P. Xu and G. Q. Max Lu, Environ. Sci. Technol., 2007, 41, 7901–7906.
27. J. A. Z. Pieterse, G. Mul, I. Melian-Cabrera and R. W. van den Brink, Catal. Lett., 2005, 99, 41–44.
28. D. Kaucký, K. Jíša, A. Vondrová, J. Nováková and Z. Sobalík, J. Catal., 2006, 242, 270–277.
29. A. Boix, E. E. Miró, E. A. Lombardo, M. A. Bañares, R. Mariscal and J. L. G. Fierro, J. Catal., 2003, 217, 186–194.
30. Q. Shen, L. D. Li, Z. P. Hao and Z. P. Xu, Appl. Catal., B, 2008, 84, 734–741.
31. J. Pérez-Ramírez, F. Kapteijn, G. Mul and J. A. Moulijn, J. Catal., 2002, 208, 211–223.
32. V. Sebastian, S. Irusta, R. Mallada and J. Santamaría, Appl. Catal., A, 2009, 366, 242–251.
33. T. Tsoncheva, L. vanova, J. Rosenholm and M. Linden, Appl. Catal., B, 2009, 89, 365–374.
34. X. Wang, H.-Y. Chen and W. M. H. Sachtler, Appl. Catal., B, 2000, 26, L227–L239.
35. C. Chupin, A. C. van Veen, M. Konduru, J. Despres and C. Mirodatos, J. Catal., 2006, 241, 103–114.
36. A. Martinez-Hernandez and G. A. Fuentes, Appl. Catal., B, 2005, 57, 167–174.
37. A. Guzmán-Vargas, G. Delahay and B. Coq, Appl. Catal., B, 2003, 42, 369–379.
38. J. Dědeček, L. Čapek, D. Kaucký, Z. Sobalík and B. Wichterlová, J. Catal., 2002, 211, 198–207.
39. J. Dědeček, D. Kaucký and B. Wichterlová, Microporous Mesoporous Mater., 2000, 35–36, 483–494.
40. L. Drozdová, J. Dědeček, D. Kaucký and B. Wichterlová, J. Phys. Chem. B, 2002, 106, 2240–2248.
41. L. D. Li, Q. Shen, J. J. Li, Z. P. Hao, Z. P. Xu and G. Q. Max Lu, Appl. Catal., A, 2008, 344, 131–141.
42. D. C. Cronauer, G. Jacobs, L. Linganiso, A. J. Kropf, J. W. Elam, S. T. Christensen, C. L. Marshall and B. H. Davis, Catal. Lett., 2011, 141, 968–976.
43. M. Ogura, S. Kage, T. Shimojo, J. Oba, M. Hayashi, M. Matsukata and E. Kikuchi, J. Catal., 2002, 211, 75–84.
44. E. Kikuchi and M. Ogura, Catal. Surv. Jpn., 1997, 1, 227–237.
45. X. Y. Zhang, Q. Shen, C. He, C. Y. Ma, J. Cheng and Z. P. Hao, Catal. Commun., 2012, 18, 151–155.
46. W. J. Mortier, Compilation of Extra Framework Sites in Zeolites, Butterworth, Stoneham, 1982.
47. T. Takaishi, M. Kato and K. Itabashi, Zeolites, 1995, 15, 21–32.
48. G. Yang, J. Q. Zhuang, Y. Wang, D. H. Zhou, M. Q. Yang, X. C. Liu, X. W. Han and X. H. Bao, J. Mol. Struct., 2005, 737, 271–276.
49. M. Ogura, S. Kage, T. Shimojo, J. Oba, M. Hayashi, M. Matsukata and E. Kikuchi, J. Catal., 2002, 211, 75–84.
50. T. Yamashita and A. Vannice, J. Catal., 1996, 161, 254–262.
51. N. Hansen, A. Heyden, A. T. Bell and F. J. Keil, J. Catal., 2007, 248, 213–225.
52. S. Suarez, M. Yates, A. L. Petre, J. A. Martin, P. Avila and J. Blanco, Appl. Catal., B, 2006, 64, 302–311.
53. K. Jíša, J. Nováková, M. Schwarze, A. Vondrová, S. Sklenák and Z. Sobalik, J. Catal., 2009, 262, 27–34.
54. J. Nováková and Z. Sobalík, Catal. Lett., 2005, 105, 169–177.
55. J. Nováková and Z. Sobalík, Catal. Lett., 2006, 111, 195–202.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c2cy00445c

---