Investigation of Reaction Mechanism of NO-C3H6-CO-O2 Reaction over NiFe2O4 Catalyst.

To elucidate the reaction mechanism of NO-C3H6-CO-O2 over NiFe2O4, we investigated the dynamics of the adsorbed and gaseous species during the reaction using operando Fourier transform infrared (FTIR). The NO reduction activity dependent on the C3H6 and CO concentrations suggested that NO is reduced by C3H6 under three-way catalytic conditions. From FTIR measurements and kinetic analysis, it was clarified that the acetate species reacted with NO-O2 to form N2 via NCO, and that the rate-limiting step of NO reduction was the reaction between CH3COO- and NO-O2. The NO reduction mechanism of the three-way catalyst on NiFe2O4 is different to that on platinum-group metal catalysts, on which NO reduction proceeds through N-O cleavage.

Introduction

Platinum-group metals (PGMs), such as Pt, Pd, and Rh, are used as active components in automotive three-way catalysts (TWCs) to purify nitrogen oxides (NOx), unburned hydrocarbons (HC), and carbon monoxide (CO) in gasoline vehicle exhausts.[1,2] Recently, the demand for PGMs has become increasing significant because of exponential increases in passenger cars in developing countries and increasingly stringent emission regulations in developed countries. However, due to the high costs and limited resources of PGMs, PGM-free automotive catalysts are highly desired.

There are few reports of PGM-free catalysts being examined under three-way catalytic conditions. Kameyama et al. reported a carburized Fe–Ce catalyst, which showed comparable activity to Pt/CeO2 under three-way catalytic conditions.[3] Kang et al. proposed a Cu–Ni bimetallic catalyst, which had higher activity than a monometal Cu catalyst.[4] Some mixed oxides, for example, spinel oxides[5] and perovskite oxides,[6,7] were also reported as PGM-free TWCs. Although these catalysts exhibited moderate activities for HC and CO oxidation, the activities for NO reduction were poor, that is, the NO conversions were less than 10% at 500 °C[5−7] or were measured at a low weight hourly space velocity (WHSV = 30 000 mL g–1 h–1).[3] Recently, we reported that NiFe2O4 exhibits high NO reduction activity under TWC conditions and at a high WHSV (206 000 mL g–1 h–1);[8,9] however, the NO reduction activities of base metal oxides are still lower than those of PGM catalysts. As there is a possibility that the TWC reaction mechanism over base metal oxide catalysts is different to that over PGM catalysts, a different strategy might be required for the activity improvement of NiFe2O4 catalyst.

The NO reduction mechanism over PGM catalysts has been extensively studied with NO–CO as a model TWC reaction. Many reports have demonstrated that NO reduction proceeds with NO dissociation on a PGM catalyst by using spectroscopic techniques[10−14] and theoretical calculations.[15,16] Reduction of NO requires reduced M0 (M = metal) sites. CO removes atomic oxygen from the catalyst surface to create reduced M0 sites and is converted to CO2. N–O cleavage easily occurs on the reduced catalytic sites due to the high extent of electron back donation into the molecular antibonding π* orbital of NO,[17] which leaves atomic nitrogen and oxygen on the surface. N2 is produced by coupling two nitrogen atoms on the surface. The remaining oxygen atoms are removed by CO, and the reduced catalytic sites are restored. There are some reports that NCO is an intermediate species during NO reduction.[18−22] Surface nitrides, which originate from NO dissociation, react with CO molecules to form NCO species. Finally, NCO is converted by NO to form N2 and CO2. NCO is also considered as a spectator.[22−24] In the presence of O2, NO reduction proceeds as described above, although the reaction rate is decreased.[11,13] The reaction mechanism of NO reduction using a hydrocarbon as a reductant under stoichiometric conditions has rarely been reported. Halkides et al. reported that C3H6 indirectly reduced NO to N2 by removing adsorbed oxygen to restore the catalytic active sites on Rh/TiO2.[25]

As for base metal oxides, the NO reduction mechanism in NO–CO has also been explored by transient experiments[26−30] and theoretical studies.[31] NO reduction is usually accompanied by N–O bond dissociation[26−30] on base metal oxides as well as on PGMs. It has been reported that N–O dissociation occurs at the oxygen vacancy or metallic sites for CeO2,[26] Cu/CeO2,[29] NiO/CeO2,[27,28] and CuO-V2O5/γ-Al2O3.[30] Therefore, enhancing the reducibility of catalysts, that is, the ease of formation of oxygen vacancies or metallic sites, is an essential factor for improving NO reduction activity in NO–CO.[6,32,33] However, this strategy did not work well in the presence of HCs.[6]

Base metal catalysts are widely used in NO-selective catalytic reduction by hydrocarbons (HC-SCR) in an oxygen-rich atmosphere, such as Cu/Al2O3[33−35] and Ag/Al2O3.[36−38] The reaction mechanism of HC-SCR has been extensively studied and it was reported that NOx and/or partially oxidized hydrocarbon species are the intermediate species.[35,38−42] However, the NO reduction mechanism for NO–HC–O2 in stoichiometric amounts, much less NO–HC–CO–O2, has never been reported. In this article, we investigated the NO reduction mechanism over NiFe2O4 in NO–C3H6–CO–O2 for further development of PGM-free TWCs. To the best of our knowledge, this is the first report of the elucidation of the NO reduction mechanism over metal oxide catalysts under stoichiometric TWC conditions.

Results and Discussion

Figure shows the temperature dependence of NO conversion over NiFe2O4 in NO–C3H6–CO–O2, NO–C3H6–O2, and NO–CO–O2. NO conversion in NO–C3H6–O2 was the same as that in NO–C3H6–CO–O2. The C3H6 and CO conversions during each reaction are also shown in Figure S2. C3H6 conversion was also similar in NO–C3H6–CO–O2 and NO–C3H6–O2 (Figure S2A). However, NiFe2O4 exhibited higher NO reduction activity in NO–CO–O2 than that in NO–C3H6–CO–O2. These results suggest that C3H6 is involved in NO reduction over NiFe2O4 in the three-way catalytic reaction. Furthermore, we explored the effect of C3H6 and CO concentrations on NO conversion at 325 °C, at which the NO conversion was around 14%. Figure shows the NO conversion as a function of C3H6 and CO concentrations. The stoichiometry was balanced by adjusting the O2 concentration, which only slightly affected the NO reduction activity (Figure S3). When the C3H6 concentration was decreased from 1200 to 500 ppm, NO conversion also slightly decreased. Further decreasing the C3H6 concentration resulted in a significant increase in NO conversion. The NO conversion in NO–CO–O2 (C3H6 = 0 ppm) was three times higher than that under TWC conditions. This result indicates that CO is a better reducing agent for NO than C3H6, and that C3H6 might have poisoned specific catalytic sites. Actually, NO conversion was almost independent of CO concentration even in the absence of CO (CO = 0 ppm). These results indicate that C3H6 is responsible for NO reduction under TWC conditions. As shown in Figure S2B, in C3H6–CO–O2 and CO–O2, CO conversion decreased in the presence of C3H6. Therefore, C3H6 suppresses the oxidation of CO by NO or O2.

Figure 1

Temperature dependence of NO conversion over NiFe2O4 in NO–C3H6–CO–O2, NO–C3H6–O2, and NO–CO–O2.

Figure 2

Dependence of NO conversion on (A) C3H6 concentration and (B) CO concentration under stoichiometric conditions over NiFe2O4 at 325 °C. The stoichiometry was balanced by adjusting the O2 concentration. The broken lines show NO conversion under standard TWC conditions.

Figure shows the IR spectra of the adsorbed species on the NiFe2O4 surface under various reaction conditions. For NO–O2 (h), bands at 1600, 1569, 1540, 1245 cm–1, and 1210 cm–1 were observed. These bands can be assigned to the two components of the split ν3 vibration of nitrate;[43,44] bridging nitrate (1600 and 1210 cm–1), bidentate nitrate (1569 and 1245 cm–1), and monodentate nitrate (1540 cm–1). As shown in spectrum (a), two strong bands at around 1549 and 1440 cm–1 assignable to νas(COO) and νs(COO) of chelating bidentate acetate[45−47] were observed for C3H6–O2. Bands at 1360 and 1310 cm–1 may be assigned to δ(CH3) of acetate[48] and νas(COO) of carbonate,[49] respectively. In the high wavenumber region, bands at 2953 and 2873 cm–1 derived from ν(CH3) of acetate were observed.[48,50] The doublet band at 2180 and 2117 cm–1 can be assigned to the R and P branches of weakly adsorbed CO. The IR spectrum for NO–C3H6–O2 (b) was almost the same as that for C3H6–O2 (a) except for a band at 2193 cm–1 assignable to NCO.[51,52] For NO–CO–O2 (f), NCO was not observed. It has been reported that NCO is an intermediate in NO-selective catalytic reduction by hydrocarbons (HC-SCR)[35,38,40] in an O2-rich atmosphere and NO–CO.[18−21] The IR spectrum for NO–C3H6–CO–O2 (c) was almost the same as that for NO–C3H6–O2 (b). The main adsorbed species on NiFe2O4 was acetate species. The bands assignable to NOx and CO were hardly observed for NO–C3H6–CO–O2.

Figure 3

IR spectra at (A) 3000–1100 cm–1 and (B) 2300–1700 cm–1 on NiFe2O4 at 325 °C in a flow of (a) C3H6–O2, (b) NO–C3H6–O2, (c–e) NO–C3H6–CO–O2, (f) NO–CO–O2, (g) NO, and (h) NO–O2. The concentrations of NO and CO were fixed at 1000 and 4000 ppm, respectively. The concentration ratios of C3H6/O2 were (a–c) 1000/6000 ppm, (d) 250/3125 ppm, and (e) 125/2563 ppm.

We also investigated the influence of C3H6 and CO concentrations on the adsorbed species on the NiFe2O4 catalysts (Figure b–f). The IR spectrum hardly changed regardless of the presence of CO (Figure b,c). This agrees with the result of the activity test, that is, NO reduction activity is almost independent of CO concentration. The IR spectrum negligibly changed when the concentration of C3H6 decreased to 250 ppm (d), at which NO reduction activity also hardly changed. Further decreasing the C3H6 concentration resulted in an enhancement in NO reduction activity and the appearance of a band at 1777 cm–1. The IR spectrum for NO–CO–O2 (f) was clearly different to that for NO–C3H6–CO–O2 (c) and showed a sharp band at 1777 cm–1. This band is attributable to NO adsorbed on Fe(III) and appeared when the catalyst had been pretreated with CO.[53] In fact, this band did not appear for NO–O2 (h) or NO alone (g). The appearance of the band at 1777 cm–1 suggested that NO adsorbed on coordinatively unsaturated Fe sites. Theoretical calculations showed that NO–CO reactions over Fe2O3 clusters proceed by undergoing compositional changes between Fe2O2 and Fe2O3.[31] In this report, CO was oxidized by O in Fe2O3 clusters and the clusters were reduced to Fe2O2. Subsequently, NO adsorbed on the less coordinated Fe sites. Accordingly, the presence of Fe(III)–NO suggests that NO reacted with CO through this redox mechanism. On the basis of the above results, it is suggested that NO–CO reactions proceed at low C3H6 concentrations whereas NO–C3H6 reactions dominantly proceed in the presence of some C3H6. These results are consistent with those of the activity dependence on C3H6 and CO concentrations, that is, NO reduction activity is high in the absence of C3H6, but it is low and independent of CO concentration in the presence of some C3H6.

Figure shows the temperature dependence of the IR spectra for NO–C3H6–O2. The bands of partially oxidized hydrocarbon were observed at 1638, 1557, 1433, and 1358 cm–1 at various temperatures. As the reaction temperature increased, the intensity of partially oxidized hydrocarbon species decreased, and the bands almost disappeared above 400 °C. The band at 2180 cm–1 assignable to the NCO species appeared above 250 °C, which is the light-off temperature of NO conversion, suggesting that NCO is an intermediate species of NO reduction over NiFe2O4. In the presence of H2O (10 vol %), the intensity of the NCO band nearly diminished completely (Figure S4), which suggests that the NCO species hydrolyzed, as reported elsewhere.[54]

Figure 4

Temperature dependence of IR spectra in a flow of NO–C3H6–O2 over NiFe2O4.

The reactivity of the adsorbed species was examined based on transient changes in the IR spectra. For C3H6–O2 (Figure S5), acetate bands (1550 and 1440 cm–1) appeared simultaneously with a negative band (ca. 3680 cm–1), which is assignable to a hydroxyl group.[46,55] Consumption of the OH group along with the appearance of the acetate species suggest that the OH group was removed or substituted with mainly acetate.[45,56]Figure A shows the IR spectrum for NO–O2 over NiFe2O4 at 300 °C after treatment with C3H6–O2. The intensity of the acetate band decreased for NO–O2. At the same time, the NCO band at 2177 cm–1 appeared and the intensity subsequently decreased (Figure B), which is similar to the dynamics of the intermediate species concentration for sequential reactions. The band attributed to NOx was hardly observed. For a flow of O2 alone, NCO was not observed, although the band intensity of CH3COO– decreased (Figure S6). These results suggest a sequential reaction from CH3COO–, via NCO, and finally to gaseous species in flowing NO–O2. Figure shows the effluent gas composition from the IR cell measured by a NOx/CO/CO2 analyzer and a mass spectrometer. When NO–O2 was supplied over the CH3COO– preadsorbed catalyst, the signal of m/e = 28 increased. Because CO concentration was negligible, this signal was attributed to N2. Therefore, CH3COO– reacted with NO–O2 to form NCO, and NCO was finally converted to N2 over NiFe2O4.

Figure 5

Dynamic changes of (A) IR spectra and (B) IR intensities of each band as a function of time in a flow of NO–O2 at 300 °C over NiFe2O4, on which acetate species were preadsorbed.

Figure 6

Effluent gas compositions of the in situ IR cell in a flow of NO and O2 over NiFe2O4 with preadsorbed CH3COO– species.

In a similar manner, the dynamics of the adsorbed NOx species in C3H6–O2 was examined. The intensity of the bands assigned to nitrate species (1600, 1566, 1540, 1245, and 1213 cm–1) increased and that of the hydroxyl group decreased in a flow of NO–O2 (Figure S7), which suggests that NOx also reacted or substituted with surface −OH groups.[57] After exposure of NiFe2O4 to NO–O2, C3H6–O2 was fed to the catalyst. As shown in Figure A, after switching NO–O2 to C3H6–O2, the intensity of the bands attributed to NOx rapidly decreased and the bands for CH3COO– appeared. NCO was not observed in this procedure and gaseous NO was formed (Figure B). The amount of NOx was calculated from the IR band area of nitrate (1095–1355 cm–1) using an adsorption coefficient of 15.9 cm–1 cm2/μmol. The amounts of desorbed NOx (3.0 μmol; measured by the NOx analyzer) and adsorbed NOx (3.0 μmol; measured by IR) were almost the same. Therefore, NOx desorbed as a gaseous species and N2 was not produced in flowing C3H6–O2 over the NOx-treated catalyst. From the above results, it can be concluded that the activation of C3H6 to partially oxidized species is an important step for NO reduction, and NCO is the intermediate species during NO reduction.

Figure 7

Dynamic changes in (A) the IR spectra and (B) effluent gas compositions from the IR cell as a function of time in a flow of C3H6–O2 at 300 °C over NiFe2O4, on which NOx species were preadsorbed.

We carried out kinetic analyses to explore the reaction between CH3COO– and NO–O2 during the TWC reaction. The initial rate of CH3COO– consumption, d[CH3COO–]/dt (Figure S8), in NO–O2 was calculated from the transient reaction of acetate with NO–O2. The concentration of CH3COO– was estimated using the Lambert–Beer equation,[35,58]where A, ε, C, W, and S represent the absorbance of the IR spectra at 1430 cm–1, the extinction coefficient of CH3COO–, the concentration of CH3COO–, the weight of the catalyst, and the geometric area of the IR disk, respectively. The extinction coefficient of CH3COO– (0.124 cm–1 cm2/μmol) was determined from the IR intensity of CH3COO– when CH3COOH was injected into the IR disk. The amount of adsorbed CH3COO– was determined from CH3COO– reacting with O2 at 500 °C and the analysis of the amount of CO and CO2 in the effluent gas. Initial rates of CH3COO– consumption at each temperature (260, 275, 300, and 325 °C) were investigated in the same manner as described above. Figure displays the Arrhenius plots for the CH3COO– consumption rates measured in transient reactions by Fourier transform infrared (FTIR) as well as for the reaction rates of NO reduction, C3H6 oxidation, and CO oxidation measured in the flow reactor under TWC conditions. Table shows the reaction rates and activation energies of each reaction. The activation energy of CH3COO– consumption was almost the same as those of NO reduction and C3H6 oxidation during the TWC reaction. The consumption rate of CH3COO– was of the same order as the oxidation rate of C3H6 under the TWC conditions. These results demonstrate that the reaction between CH3COO– and NO–O2 occurs during the TWC reaction and is the rate-limiting step for NO reduction. The activation energy of CO oxidation was entirely different to that of NO reduction and C3H6 oxidation under the TWC conditions, which supports the previous argument that CO does not contribute to NO reduction.

Figure 8

Arrhenius plots for NO reduction, C3H6 oxidation under TWC conditions, and CH3COO– consumption in a flow of NO–O2.

Table 1

Reaction Rates and Activation Energies of Each Reaction

reaction	rate (300 °C) (μmol g–1 s–1)	activation energy (kJ mol–1)
CH3COO– consumption (in situ IR)	3.8	87
NO reduction (flow reactor)	0.38	84
C3H6 oxidation (flow reactor)	2.0	92
CO oxidation (flow reactor)	3.6	146

It has been reported that CH3COO– and HCOO– are intermediate species during NO reduction for HC-SCR.[35,38,40] To investigate the reactivity of CH3COO– and HCOO–, we conducted the transient reaction of adsorbed CH3COO– and HCOO– in NO–O2. After injection of CH3COOH or HCOOH (2 μL) and adsorption of CH3COO– or HCOO– on the catalyst disk, NO–O2 was fed to the catalyst. Consumption of adsorbed CH3COO– and formation of NCO were observed (Figure S9). The consumption rate of CH3COO– in NO–O2 (3.7 μmol s–1) is close to that of C3H6–O2 treated catalysts (3.8 μmol s–1). However, after adsorption of HCOOH, NCO was not formed, although the intensity of the bands assigned to HCOO– decreased (Figure S10). On the basis of the stoichiometry, CH3COO– reacts with 3.5 mol of NO, whereas HCOO– only reacts with 0.5 mol of NO. Accordingly, it is suggested that the contribution of CH3COO– to NO reduction is major but that of HCOO– is negligible. The reaction mechanism of the transformation from CH3COO– to NCO has been considered in previous reports as follows: the reaction between CH3COO– and NO2 gives the aci-anion of nitromethane,[59] which easily decomposes to NCO species (Scheme S1).[59,60]

The consumption rates of CH3COO– under NO, O2 alone, or NO2 were also investigated (Figures S11, S6, and S12, respectively, and Figure S8 and Table S1). The consumption rate of CH3COO– under NO (0.38 μmol s–1) was much lower than that under NO–O2 (3.8 μmol s–1). In a flow of O2, the CH3COO– consumption rate (2.9 μmol s–1) was moderate, and was three quarters of that in a flow of NO–O2. The ratio of the CH3COO– consumption rate in O2 to that in NO–O2 (2.9/3.8 = 0.76) is similar to the selectivity for the reaction of C3H6 with O2 in C3H6–NO–O2 (0.76) roughly estimated from activity tests (detailed in Supporting Information). It is suggested that the selectivity of the reaction of C3H6 with NO or O2 under TWC conditions was determined by that of CH3COO– with NOx or O2, which is a critical factor for determining the NO reduction activity. In a flow of NO, the consumption rate of CH3COO– under NO (0.38 μmol s–1) was much lower than that under NO–O2 (3.8 μmol s–1), and the bands assignable to Fe(III)–NO (1820, 1777 cm–1)[54] and NCO (2180 cm–1) appeared (Figure S11). The band of NCO continued to increase in a flow of NO, unlike that in a flow of NO–O2. The formation of N2 was observed in a flow of NO–O2 (Figure ) and NO (Figure S11), and the amount of N2 was higher in NO–O2 than that in NO. These results suggest O2 was not necessary for the decomposition of NCO to N2, however, O2 promoted the reaction.[59] The consumption rate of CH3COO– under NO2 (1.1 μmol s–1) was higher than that under NO (0.38 μmol s–1), showing that NO2 reacted with CH3COO– more easily than did NO. We also investigated the CH3COO– consumption rate dependent on the O2 partial pressures in NO–O2 mixtures (Figure S13). The consumption rate declined with decreasing O2 partial pressures and dropped steeply without O2. The above results suggest the role of O2 in the reaction of CH3COO– was the oxidization of NO to NO2, which easily reacted with CH3COO– to lead to the formation of N2. Furthermore, the effect of H2O (10 vol %) on the CH3COO– consumption rate was examined (Figure S14). The addition of H2O deteriorated the consumption rate by a factor of 0.35 (Table S1) and nearly diminished the intensity of the NCO band.

On the basis of the above results, the overall reaction pathway is proposed, as shown in Scheme . C3H6 is responsible for NO reduction under NO–C3H6–CO–O2 reaction. CH3COO–, derived from the partial oxidation of C3H6, reacts with NO–O2 and NO converts to N2 via NCO. The reaction between CH3COO– and NO–O2 is the rate-limiting step. Accordingly, the reactivity of CH3COO– is an essential factor for determining the NO reduction activity under TWC conditions. The reaction mechanism of NO-selective catalytic reduction by hydrocarbons (HC-SCR) in an O2-rich atmosphere was reported as follows:[35,38,40−42,60,61] NO and the hydrocarbon are oxidized to NOx and partially oxidized the hydrocarbon species (acetate, formate, enolate, or acrylate), respectively. NOx and the partially oxidized hydrocarbon species react together and convert to N2 via NCO and/or CN species. The mechanism of NO reduction in NO–HC–CO–O2 on NiFe2O4 is entirely different to that on PGM catalysts[10−16,25] and is similar to that under HC-SCR on base metal oxides. NO reduction on PGM catalysts is accompanied with N–O dissociation and NO does not directly interact with the reductant (CO or C3H6). Figures A and S13 show the IR spectra for NiFe2O4 and Rh (1 wt %)/Al2O3, respectively, under TWC conditions. The bands derived from NO (1896 cm–1; Rh(NO)δ+)[62−64] and CO (2015 and 2091 cm–1; RhI(CO)2)[62−64] were observed on Rh/Al2O3. However, the bands attributed to NO and CO on NiFe2O4 were not observed and partially oxidized hydrocarbon species were predominantly adsorbed. The Rh/Al2O3 catalyst maintained its metallic sites to adsorb NO in the presence of O2, which are effective for NO dissociation. However, the surface of the NiFe2O4 catalyst was mostly covered by oxygen during the reaction. Therefore, the reaction pathway via oxygenates similar to SCR under oxygen-rich conditions may be preferable for base metal oxide catalysts. Even in the presence of C3H6, adsorption of NO and CO was observed on Rh/Al2O3, but was not observed on NiFe2O4. This might be why the reaction between NO and CO did not occur on NiFe2O4. The coverage of adsorbates should be one of the essential factors for the determination of the reaction mechanism under TWC conditions.

Scheme 1

Reaction Mechanism on NiFe2O4 under TWC Conditions

Although the NO reduction mechanism during the TWC reaction on NiFe2O4 (Scheme ) is similar to that during NO-selective catalytic reduction (under O2-rich conditions), NiFe2O4 exhibited low NO reduction activity under O2-rich conditions (λ > 1).[8] The density of the adsorbed CH3COO– on NiFe2O4 in C3H6–O2 was 3.02 nm–2, which was calculated from the amount of adsorbed CH3COO– and the Brunauer–Emmett–Teller surface area. This value indicates that the CH3COO– species almost completely covered the NiFe2O4 surface (detailed in Supporting Information).[65] Further, the basic OH group was removed or exchanged upon adsorption of CH3COO– and NOx (Figures S5 and S7), suggesting that both were adsorbed on the same sites (Mcus–O sites).[65] NOx species hardly adsorbed on NiFe2O4 during steady-state TWC reactions because of competitive adsorption of hydrocarbon oxygenates, which may adsorb more strongly on NiFe2O4. Accordingly, CH3COO– reacted with gaseous NO or weakly adsorbed NOx species. The unbalanced adsorption of the C3H6- and NO-derived species may be the reason for the low selectivity of CH3COO– as a reductant of NOx. In other words, adsorption of NOx in the presence of C3H6 may enhance the NOx reduction activity during the TWC reaction on NiFe2O4 catalyst.

Conclusions

The TWC reaction mechanism on NiFe2O4 was investigated using in situ and operando FTIR measurements, and was elucidated to be as follows: CH3COO–, which originates from C3H6–O2, reacts with NO–O2 to form N2 via NCO on NiFe2O4 under three-way catalytic conditions. The reaction between CH3COO– and NO–O2 is the rate-limiting step during NO reduction. The results of the present work give us a distinct strategy for the further development of base metal oxide TWCs, such as controlling the adsorption energy of the CH3COO– species.

Methods

NiFe2O4 catalysts were prepared by a reverse strike coprecipitation method described elsewhere.[8] Fe(NO3)3·9H2O (purity 99%, Kishida Chemical Co., Ltd, Japan) and Ni(NO3)2·6H2O (purity 98%, Kishida Chemical Co., Ltd, Japan) in an atomic ratio of Fe/Ni = 2:1 were dissolved in 100 mL of distilled water. An aqueous solution of Fe and Ni nitrates was added dropwise into a NaOH solution (1 M) to obtain a brown precipitate. The precipitate was filtered and washed with hot water several times until the pH of the filtrate became neutral to completely remove NaOH in the precipitate. The sample was dried overnight at 80 °C and calcined at 500 °C for 3 h. Synchrotron X-ray diffraction (XRD) at the BL5S2 of Aichi SR was performed to confirm that the synthesized NiFe2O4 formed a single cubic spinel phase (Figure S1). The surface area of the synthesized NiFe2O4 was 44.4 m2 g–1, which was determined by its N2 adsorption isotherm (MicrotracBEL BELSorp 28SA).

Activity tests were performed by means of a fixed-bed flow reactor at atmospheric pressure. The catalyst (17.5 mg) was put inside a Pyrex glass tube with an internal diameter of 4 mm. The catalytic test was carried out at a total flow rate of 60 mL min–1 (WHSV was 206 000 mL h–1 g–1, gas hourly space velocity was roughly 140 000 h–1). Prior to the activity test, the catalyst was pretreated in a flow of O2 at 400 °C for 15 min. The catalytic run was carried out from 200 to 500 °C in 50 °C increments. The steady-state activity was measured using a NOx analyzer and a nondispersive infrared CO/CO2 analyzer (Horiba VIA-3100). The activity tests were performed in NO–C3H6–CO–O2, NO–C3H6–O2, and NO–CO–O2 under stoichiometric conditions (λ = 1, described below) and Ar was used as a balance (NO, C3H6, and CO were fixed at 1000, 1000, and 4000 ppm, respectively. The concentration of O2 was 6000 ppm (for the NO–C3H6–CO–O2 reaction), 4000 ppm (for the NO–C3H6–O2 reaction), or 1500 ppm (for the NO–CO–O2 reaction) to balance the stoichiometric conditions.) The dependence of NO reduction activity on C3H6/CO concentration was investigated under stoichiometric conditions (λ = 1) by adjusting the O2 concentration. λ is defined as below.Thermal stability (calcination at 1000 °C for 10 h) and water tolerance were examined under TWC conditions in our previous work.[8]

IR measurements were performed using a JASCO FTIR/6100 instrument equipped with an in situ quartz cell connected to a gas-flow reactor. The sample was pressed into a 0.13 g self-supporting wafer and mounted into the cell with CaF windows. Most of the spectra were measured at 300 °C. Prior to IR measurements, the catalyst was heated in flowing O2/Ar at 400 °C for 15 min, cooled to the desired temperature under Ar, and then the background spectra were measured. Subsequently, various gas mixtures were fed at a flow rate of 60 mL/min. The concentrations of NO, C3H6, CO, and O2 in the gas mixtures were the same as those for the activity tests. Transient IR measurements were performed to study the dynamics of adsorbed species on the catalyst surface. In the case of the acetate species, C3H6–O2 (C3H6 1000 ppm and O2 4000 ppm) was fed to the catalyst for 30 min. After purging with Ar for 10 min, NO–O2 (NO 1000 ppm and O2 4000 ppm) was fed to the C3H6–O2 treated catalyst. In a similar manner, the dynamics of the NOx species was measured in a flow of C3H6–O2 after pretreatment of the catalyst in a flow of NO–O2. The FTIR cell was connected to a NOx/CO/CO2 analyzer (Horiba VIA-3100) and a quadrupole mass spectrometer (MicrotracBEL BELMass) to analyze the effluent gas composition. In particular, N2 and CO can be distinguished by the combination of the mass spectrometer (mass-to-charge ratio, m/e, is 28) and the CO analyzer.