Rod‐Shaped Fe2O3 as an Efficient Catalyst for the Selective Reduction of Nitrogen Oxide by Ammonia

Nanorust: Fe2O3 nanomaterials with controllable crystal phase and morphology have been successfully fabricated. The γ-Fe2O3 nanorods that are enclosed by the reactive {110} and {100} facets are highly active and distinctively stable for selective catalytic reduction of NO with NH3. The picture shows the shape and surfaces and the surface atomic configurations of the preferentially exposed {110} and {001} planes. Selective catalytic reduction (SCR) of nitrogen oxides (NOx) with NH3 is nowadays the most promising technology for eliminating nitrogen oxides that are emitted from combustion of fossil fuels.1 Vanadia-based catalysts are commonly used for this reaction in stationary sources, such as coal-fired power plants and energy use in industry. This mature system displays adequate activity typically at 300–400 °C, but it is still not satisfactory with respect to the volatility and toxicity of VOx and the easy deactivation.1–3 Attracted by the inherently environmentally benign character and the prominent thermal stability, ferric oxides have long been explored to catalyze selective reduction of NOx with ammonia,3–8 but no significant progress has been achieved so far. Ferric oxides commonly suffer from the insufficient activity at low temperatures and the severe deactivation induced by H2O and SO2 that are permanently and abundantly present in the exhaust. On the other hand, studies on crystal phase and shape control of Fe2O3 nanomaterials are of great interest in materials science and are actively being pursued in recent years.9–14 However, most of the synthetic strategies involve high-temperature treatments, yielding undesired mixtures of ferric oxide polymorphs owing to diverse phase transitions.9 To date, there still remains a challenge to effectively tailor the shape and the crystal phase of Fe2O3 materials at nanometer level. Herein, we show that γ-Fe2O3 nanorods, which preferentially expose the reactive facets by crystal-phase and morphology control through a solution-based approach, efficiently catalyze selective reduction of NO with NH3. We started with shape-controlled synthesis of β-FeOOH nanorods by precipitation of ferric chloride with sodium carbonate in aqueous solution containing poly(ethylene glycol) (PEG) at 120 °C. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) micrographs show that the obtained β-FeOOH has a rod-like structure with a diameter of 30–50 nm and a length of 350–500 nm (Figure S1 in the Supporting Information). When viewed along the [10] (Figure 1 a) and [100] directions (Figure 1 b) by means of high-resolution TEM (HRTEM), it can be seen that the {010} planes of the nanorod are dominantly exposed. Together with the square-shaped cross section (Supporting Information, Figure S1), the β-FeOOH nanorod is determined to be a rectangular block that is enclosed by two {100} flat planes, two {010} side planes, and two {001} end planes (Figure 1 c). a) A low-magnification TEM image of a single β-FeOOH nanorod viewed along the [10] direction.The insert HRTEM image indicates that the nanorod exposes the {010} planes. b) Another HRTEM image viewed along the [100] orientation, also showing the preferential exposure of the {010} planes. c) The real shape of the β-FeOOH nanorod. γ-Fe2O3 nanorods were then obtained by refluxing the β-FeOOH precursor in PEG at 200 °C. An X-ray powder diffraction (XRD) pattern (Supporting Information, Figure S2) shows that the diffraction lines of the (400) and (440) planes are intensified significantly, whereas those of the (422) and (511) planes are weakened considerably compared to the standard powder reflection pattern, which is indicative of anisotropic growth of the {110} and {100} planes. SEM and TEM observations reveal that the γ-Fe2O3 nanorods have an average diameter of 40 nm and a mean length of 400 nm (Supporting Information, Figure S3). Moreover, the γ-Fe2O3 nanorods have mesopores of about 22 nm in diameter that have open structures and are isolated from each other, yielding a high surface area of 120 m2 g−1. Iron oxyhydroxide can be easily dehydrated into iron oxide, but the crystal phase varies largely depending on the rearrangement of Fe3+ and O2− ions.15 β-FeOOH has the least dense crystalline structure among its polymorphs and usually dehydrates to α-Fe2O3 upon heating.16 We obtain, as expected, α-Fe2O3 nanorods by calcining the β-FeOOH precursor at 500 °C in air (Supporting Information, Figure S4). When heated to reflux in PEG, the β-FeOOH precursor is surprisingly converted into γ-Fe2O3 nanorods owing to the moderate water removal manner that benefits the crystallization of maghemite. Ferric oxide has four polymorphs, and the phase transition is strongly dependent on the size of the particles, especially at the nanometer level.9, 17, 18 Therefore, we examined the crystal-phase stability of the γ-Fe2O3 nanorods by an in situ XRD experiment (Supporting Information, Figure S5). The characteristic (311) and (400) diffraction lines of γ-Fe2O3 remain stable up to 600 °C with only slightly weakened intensities. Moreover, the rod shape is also well maintained even if the γ-Fe2O3 nanorods were calcined at 550 °C for 5 h in air. All of these results show the exceptionally high crystal-phase and shape stabilities of the γ-Fe2O3 nanorods that are a consequence of the unique morphological feature. The crystallographic nature of the γ-Fe2O3 nanorods was further examined by HRTEM. Figure 2 a shows a single γ-Fe2O3 nanorod growing along the [10] direction. The interplanar distances of 0.253, 0.297, and 0.417 nm on the enlarged image (Figure 2 b) correspond to the lattice fringes of the {113}, {110}, and {001} planes, respectively. The square-shaped cross-section viewed near the [110] orientation is constructed by (10) and (001) as side planes (Figure 2 c). When viewed along the [001] direction, the nanorod dominantly exposes the {110} planes (Figure 2 d,e). Taking all the analyses into account, the nanorod is constructed with two {110} as end planes, and two {110} and two {001} as side planes (Figure 2 f). It is noteworthy that these surfaces are terminated simultaneously by Fe3+ and O2− sites.19–21 There are equal numbers of octahedral and tetrahedral Fe3+ ions on the {110} plane (4.05 atoms per nm2), but only octahedral Fe3+ ions on the {001} plane (5.72 atoms per nm2) (Figure 2 g and Supporting Information, Figure S3). Consequently, the surface of the γ-Fe2O3 nanorod is relatively rich in octahedral Fe3+ ions. More importantly, the length of the FeO bond (0.2091 nm) is greater than that in the tetrahedrally coordinated environment (0.1837 nm),20 making it more reducible and reactive. a) Low- and b) high-magnification TEM images of a single γ-Fe2O3 nanorod, indicating the dominant exposure of the {110} and {001} planes. c) A square-shaped cross-section of the single rod near the [110] orientation, which is constructed by two side (10) and (001) planes with a lattice angle of 90°. d) Low- and e) high-magnification TEM images of another γ-Fe2O3 nanorod viewed along the [001] direction where the (220) planes are preferentially exposed. f) The real shape of the nanorod and the surface atomic configurations of the preferentially exposed {110} and {001} planes. g) Atomic illustration of a cubic symmetry (space group Fdm) and the coordination patterns of tetrahedral and octahedral Fe3+ ions. Figure 3 a shows the catalytic performance of the γ-Fe2O3 nanorods in SCR of NO with NH3. The light-off temperature (the temperature at which the conversion of NO reaches 50 %) is as low as 170 °C, and 100 % NO conversion is achieved at 220 °C. The temperature window for 80 % NO conversion ranges from 200 to 400 °C and the selectivity towards N2 is maintained at about 98 % over the whole temperature range (Supporting Information, Figure S6). This salient performance is of utmost importance in NOx removal at low to medium temperatures. When compared to the industrially used VOx-based catalysts,22, 23 the less costly and non-toxic Fe2O3 nanorods offer fairly compatible or even slightly better efficiency under similar reaction conditions (Supporting Information, Figure S7). As the exhaust usually contains a large amount of H2O (2–15 vol %) and certain amounts of SO2 (δ=30–2000 ppm),2, 24 we further tested the Fe2O3 nanorods with a feed stream containing 10 vol % H2O and δ=100 ppm SO2. Because of the adsorption of water on the catalyst surface, the light-off temperature increases to 260 °C and the temperature region for 80 % NO conversion shifts to 300–500 °C (Figure 3 a). Nevertheless, the pronounced activity and the available temperature window still fulfill the criteria for practical stationary NOx removal.2 a) NO conversion as a function of temperature with (◊) and without (○) water (10 vol %) and SO2 (δ=100 ppm) in the feed gas of δ=1000 ppm NO/δ=1000 ppm NH3/3.0 vol % O2/He (72 000 mL g−1 h−1). b) Stability tests at 350 °C for 150 h. c) TEM images of the used catalyst. The γ-Fe2O3 nanorods are very robust both in a transient process with frequent variation in temperature and in a long-term steady operation at a constant temperature. Upon periodically fluctuating the reaction temperature over the range 150–450 °C, the conversion of NO and the selectivity of N2 are highly reproducible; the temperature region for 80 % NO conversion remains at 200–400 °C, and the size and shape of the used catalyst remains unchanged (Supporting Information, Figure S8). When used for a long-term test at 350 °C, a typical temperature in practical applications, NO conversion is maintained at 100 % for 150 h (Figure 3 b). The addition of H2O and SO2 slightly lowers NO conversion to about 90 %, but the removal of SO2 from the reaction gas immediately raises NO conversion to 93 %. After cutting off the supply of water, the conversion of NO is rapidly restored to 100 % and is kept at this level for the rest of the operation. This reaction pattern demonstrates that the γ-Fe2O3 nanorods are highly sulfur-resistant and the inhibition of water is reversible, thus differing from the traditional iron oxides that are severely and irreversibly deactivated by H2O and SO2.5, 6 TEM analyses of the catalyst after the stability test confirm that both the morphology and the exposed facets are almost the same as the fresh one (Figure 3 c). Electron energy-loss spectroscopy (EELS) reaffirmed that the oxidation state of iron and the coordination of oxygen in the catalyst keep unchanged before and after the reaction (Table S1 in the Supporting Information), indicating the stable surface coordination environment. Activation of NH3 on the catalyst surface is generally viewed as the primary step in SCR of NO, as it is experimentally indentified on VOx-based catalysts.25–27 Temperature-programmed desorption (TPD) of NH3 on the γ-Fe2O3 nanorods shows an intense desorption peak of N2 at 250–430 °C (Figure 4 a). As there is no oxygen in the feed gas, the generation of nitrogen originates from the oxidation of adsorbed ammonia by the lattice oxygen of γ-Fe2O3. This is also supported by the TPD profile of NH3/O2 co-adsorption where the pre-adsorbed molecular oxygen has negligible impact on the desorption pattern of N2 (Figure 4 b). In fact, the Fe2O3 nanorods catalyze NH3 oxidation very efficiently; the conversion of NH3 and the selectivity towards N2 are as high as 90 % even at 350 °C (Supporting Information, Figure S9), evidencing the facile activation of ammonia. On the other hand, the adsorption of NO is strongly affected by the presence of oxygen molecule. NO is only weakly adsorbed on the Fe2O3 nanorods and is easily desorbed below 250 °C (Figure 4 c). However, the presence of molecular oxygen greatly promotes the adsorption of NO. The co-adsorption of NO/O2 yields a strong desorption peak of NO2 at 372 °C (Figure 4 d). This is in accord with a comparative test of NO oxidation on the Fe2O3 nanorods where the conversion of NO to NO2 reaches 26 % at about 370 °C (Supporting Information, Figure S9), suggesting that the γ-Fe2O3 nanorods enable to catalyze NO oxidation effectively as well. TPD profiles of a) NH3, b) NH3/O2, c) NO, and d) NO/O2 on the γ-Fe2O3 nanorods. Catalytically, the outstanding SCR performance of the γ-Fe2O3 nanorods is intimately associated with the exposed {110} and {001} planes. Ammonia is well adsorbed on the Lewis acidic Fe3+ sites, whereas NO is only weakly adsorbed.28–30 The inferior performance of the α-Fe2O3 nanorods in a comparative test confirms that the exposed {210} and {001} planes are much less active (Supporting Information, Figure S4 d,e). This is because these Fe-terminated surfaces in the α phase (space group Rc (no. 167)) only provide ferric sites for the adsorptions of NO and NH3 but lack neighboring oxygen anions for their activations. In contrast, the {110} and {001} surfaces on the γ-Fe2O3 nanorods contain iron and oxygen ions simultaneously. The ferric site and its neighboring basic oxygen site jointly accomplishes the reactive adsorptions of NO and NH3, and their subsequent reaction yields nitrogen and water, probably through surface nitrates.27 The apparent activation energies are 49 and 56 kJ mol−1 at 120–160 °C and 340–360 °C, respectively. The turnover frequencies of the Fe3+ site are 1.39×10−3 s−1 at 150 °C and 6.01×10−3 s−1 at 350 °C (Table S2 in the Supporting Information). In summary, we have successfully fabricated crystal-phase and morphology controlled Fe2O3 nanomaterials. In particular, γ-Fe2O3 nanorods that are enclosed by the reactive {110} and {100} facets are highly active and distinctively stable for the SCR of NO with NH3. This provides a possibility for obtaining highly efficient SCR catalysts using less costly and non-toxic Fe2O3 nanomaterials, and potentially offers a new more general strategy to promote the catalytic property of metal oxides by tuning crystal phase and shape at the nanometer level. The β-FeOOH nanorods were prepared by an aqueous precipitation method. An aqueous solution containing FeCl3⋅6 H2O (5.38 g), NaCl (11.60 g), and PEG (10 mL) in water (190 mL) was gradually heated to 120 °C. A Na2CO3 aqueous solution (200 mL, 0.2 M) was then added through a syringe pump at a rate of 5.5 mL min−1. The mixture was then aged at 120 °C for 1 h. The precipitate was washed with water and ethanol, and finally dried at 50 °C for 6 h under vacuum. The γ-Fe2O3 nanorods were prepared by heating the β-FeOOH precursor in PEG to reflux. A slurry mixture containing β-FeOOH nanorods (5.0 g) and PEG (500 mL) was gradually heated to 200 °C and heated to reflux for 24 h under a stream of nitrogen. The resulting solid was washed with water and ethanol, followed by drying at 50 °C for 12 h under vacuum. TEM images were taken on a Philips FEI Tecnai G2 microscopy operating at 120 kV. HRTEM images were recorded on a FEI Cs-corrected Titan 80–300 microscope operated at 300 kV with a Gatan filter. EELS analyses were performed with an energy dispersion of 0.1 eV. TPD of NO and NH3 on the γ-Fe2O3 nanorods were conducted with a quartz-tube reactor that was connected to a mass spectrometer. 100 mg samples were pre-treated with a 3.0 vol % O2/He mixture at 400 °C for 0.5 h, followed by NH3 or NO adsorption at room temperature. SCR of NO with NH3 was conducted with a continuous-flow quartz-tube reactor under atmospheric pressure. Before the test, γ-Fe2O3 nanorods (100 mg, 40–60 mesh) were treated with a 3.0 vol % O2/He mixture (60 mL min−1) at 400 °C for 0.5 h. Typically, the feed gas contained δ=1000 ppm NO, δ=1000 ppm NH3, and 3 vol % O2 balanced with He (120 mL min−1). When necessary, 10.0 vol % H2O and δ=100 ppm SO2 were added to the gas stream. The concentrations of nitrogen oxides and nitrogen in the inlet and outlet streams were continuously detected by a NO/NOx analyzer and a mass spectrometer. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.