Article pubs.acs.org/est
Manganese Oxides with Rod-, Wire-, Tube-, and Flower-Like Morphologies: Highly Effective Catalysts for the Removal of Toluene Fang Wang, Hongxing Dai,* Jiguang Deng, Guangmei Bai, Kemeng Ji, and Yuxi Liu Laboratory of Catalysis Chemistry and Nanoscience, Department of Chemistry and Chemical Engineering, College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, China S Supporting Information *
ABSTRACT: Nanosized rod-like, wire-like, and tubular α-MnO2 and flower-like spherical Mn2O3 have been prepared via the hydrothermal method and the CCl4 solution method, respectively. The physicochemical properties of the materials were characterized using numerous analytical techniques. The catalytic activities of the catalysts were evaluated for toluene oxidation. It is shown that αMnO2 nanorods, nanowires, and nanotubes with a surface area of 45−83 m2/g were tetragonal in crystal structure, whereas flowerlike spherical Mn2O3 with a surface area of 162 m2/g was of cubic crystal structure. There were the presence of surface Mn ions in multiple oxidation states (e.g., Mn3+, Mn4+, or even Mn2+) and the formation of surface oxygen vacancies. The oxygen adspecies concentration and low-temperature reducibility decreased in the order of rod-like α-MnO2 > tube-like α-MnO2 > flower-like Mn2O3 > wire-like α-MnO2, in good agreement with the sequence of the catalytic performance of these samples. The best-performing rod-like α-MnO2 catalyst could effectively catalyze the total oxidation of toluene at lower temperatures (T50% = 210 °C and T90% = 225 °C at space velocity = 20 000 mL/(g h)). It is concluded that the excellent catalytic performance of α-MnO2 nanorods might be associated with the high oxygen adspecies concentration and good low-temperature reducibility. We are sure that such one-dimensional well-defined morphological manganese oxides are promising materials for the catalytic elimination of air pollutants.
■
dimensional (1D) α-, β-, γ-, and δ-MnO2 nanorods using a hydrothermal method and observed good catalytic activities for CO oxidation.16 Gao et al. obtained 1D α-MnO2 nanowires by hydrothermally treating the mixture of KMnO4 and NH4Cl at 140 °C for 24 h.17 Zheng et al. generated single-crystalline 1D β-MnO2 nanotubes (diameter 200−500 nm and length several micrometers) via a poly(vinyl pyrrolidone)-assisted hydrothermal route with MnSO4 with NaClO3 as precursor.18 Without the use of a template but with CCl4 and water as medium, Yuan et al. synthesized flower-like α- and γ-MnO2 (surface area 239 m2/g), which showed a good electrochemical capacitive behavior.19 It was reported that manganese oxides were catalytically active for the complete oxidation of VOCs, such as propane, n-hexane, benzene, and toluene.20−23 For instance, Finocchio and Busca investigated the surface and redox properties of Mn3O4, Mn2O3, and MnO2, and claimed that the bulk oxygen diffusion rate had an effect on the catalytic oxidation rate in the oxidation of propane.20 Delmon and coworkers observed that the γ-MnO2 catalyst outperformed the 0.3 wt % Pt/TiO2 catalyst in the oxidation of n-hexane.23 After studying the oxidation of benzene over manganese oxide
INTRODUCTION Most volatile organic compounds (VOCs), such as formaldehyde, methanol, benzene, and toluene, are harmful to the atmosphere and human health. It is highly desired to control the emissions of VOCs. Up to now, a number of methods (e.g., adsorption) have been developed for the removal of hazardous materials.1−6 Among the strategies for VOCs elimination, catalytic oxidation is believed to be one of the most effective pathways because it can operate at low temperatures and no secondary pollution products are generated.7−11 The key issue of such a technology is the availability of high-performance catalysts. Although supported precious metal catalysts show excellent activities for the total oxidation of toluene at low temperatures,12−14 the high cost and some involved problems (e.g., sintering and volatility) prohibit their wide applications. Cheap transition-metal oxides, such as manganese oxides, cobalt oxides, and chromia, are active at high temperatures,7,8,15 but they are inferior to precious metals in catalyzing the combustion of toluene at low temperatures. Hence, it is of significance to develop a catalyst that is cheap and effective for the removal of toluene at low temperatures. In the past years, a large number of works have been focused on the controlled preparation of manganese oxides with various morphologies. Up to now, manganese oxides with rod-like, wire-like, tubular, and spherical shapes have been generated.16−19 For example, Zhu and co-workers prepared one© 2012 American Chemical Society
Received: Revised: Accepted: Published: 4034
November 11, 2011 January 20, 2012 March 11, 2012 March 13, 2012 dx.doi.org/10.1021/es204038j | Environ. Sci. Technol. 2012, 46, 4034−4041
Environmental Science & Technology
Article
basis of the toluene consumption and CO2 production, the carbon balance and the conversion of toluene were calculated. The relative errors for the gas concentration measurements were less than ±1.5%. The balance of carbon throughout the investigation was estimated to be ca. 99.5%.
octahedral molecular sieve (OMS-2) catalyst, Luo et al. believed that the excellent activity and stability of OMS-2 at low temperatures were due to the hydrophobic property and the facile evolution of lattice oxygen.21 Aguero et al. observed good catalytic performance over Al2O3-supproted MnOx catalyst for the combustion of ethanol and toluene, which was attributed to the high capacity for adsorbing oxygen, the existence of surface defects, and the good reducibility of the catalyst.22 It has been generally accepted that catalytic activity is related to the surface area, defective structure, reducibility, and morphology of a catalyst. The particle morphology has an important impact on catalytic oxidation performance of transition metal oxides (e.g., Co3O4).24 However, up to now, rarely has work been done on the comparative investigation of manganese oxides with various well-defined morphologies. Previously, our group prepared a series of three-dimensionally (3D) ordered or wormhole-like mesoporous transitionmetal oxides (e.g., chromia,25,26 iron oxide,27 manganese oxide,28 and cobalt oxide28,29) by using the 3D ordered mesoporous silica KIT-6- or SBA-16-nanocasting method, and investigated their physicochemical properties. We found that these 3D mesoporous transition-metal oxides performed well in catalyzing the combustion of formaldehyde, acetone, methanol, and toluene. Recently, we adopted the hydrothermal method to generate a number of transition-metal oxides with well-defined morphologies. In this paper, we report the controlled preparation and catalytic properties of rod-like, wire-like, and tubular MnO2 as well as flower-like Mn2O3 for the combustion of toluene.
■
RESULTS AND DISCUSSION Crystal Phase Composition. Figure 1 shows the XRD patterns of the as-prepared manganese oxide samples. By
Figure 1. Wide-angle XRD patterns of (a) rod-like MnO2, (b) wirelike MnO2, (c) tube-like MnO2, and (d) flower-like Mn2O3. Δ: Impurity Mn3O4 phase; ○: impurity MnO2 phase.
■
EXPERIMENTAL SECTION Catalyst Preparation. The manganese oxide catalysts were prepared according to the hydrothermal16,30 or solution method.19 The detailed procedures are described in the Supporting Information. The as-prepared samples are referred to as rod-like MnO2, wire-like MnO2, tube-like MnO2, and flower-like Mn2O3. Catalyst Characterization. All of the as-prepared samples were characterized by techniques such as X-ray diffraction (XRD), N2 adsorption−desorption (BET), scanning electron microscopy (SEM), transmission electron microscopy (TEM), selected-area electron diffraction (SAED), X-ray photoelectron spectroscopy (XPS), and H2 temperature-programmed reduction (H2-TPR). The detailed methods are stated in the Supporting Information. Catalytic Evaluation. Catalytic activity of the samples was evaluated in a continuous-flow fixed-bed quartz microreactor (i.d. 4 mm). To minimize the effect of hot spots, the catalyst (0.1 g, 40−60 mesh) was diluted with 0.3 g of quartz sands (40−60 mesh). The reactant feed (flow rate 33.3 mL/min) was 1000 ppm toluene + O2 + N2 (balance), with the toluene/O2 molar ratio and space velocity (SV) being 1/400 and 20 000 mL/(g h), respectively. For the change of SV, we altered the total flow rate of the reactant feed by the mass flow controller (D07 19CM, Beijing Sevenstar Electronics Co.). The outlet gases were analyzed online by a gas chromatograph (Shimadzu GC-2010) equipped with a flame ionization detector (FID) and a thermal conductivity detector (TCD), using a 1/8-in. Chromosorb 101 column (3 m long) for toluene separation and a 1/8-in. Carboxen 1000 column (3 m long) for permanent gas separation. The outlet gases were also monitored online by a mass spectrometer (HPR20, Hiden). We found that no other products were detected in addition to CO2 and H2O. On the
comparing to the XRD patterns of the standard α-MnO2 (JCPDS PDF 72-1982), Mn2O3 (JCPDS PDF 41-1442), and Mn3O4 (JCPDS PDF 24-0734) samples, one can realize that the hydrothermally derived rod- and tube-like manganese oxide samples were single-phase α-MnO2 and of tetragonal crystal structure; in addition to the main phase of tetragonal α-MnO2, there was a trace amount of tetragonal Mn3O4 phase in the wire-like manganese oxide sample. In the CCl4-solution derived flower-like manganese oxide sample, however, there were a cubic Mn2O3 phase in majority and a tetragonal α-MnO2 phase in minority. All of the diffraction peaks could be well indexed, as indicated in Figure 1c and d. From Figure 1, one can also observe no significant difference in XRD signal intensity of the four samples, indicating that they possessed similar crystallinity, a result due to the same subsequent thermal treatments. The XRD results demonstrate that the preparation conditions had an important influence on crystal structure of the manganese oxide sample. Morphology, Surface Area, Surface Element Composition, and Oxygen Species. Figure 2 shows the SEM images of the as-prepared manganese oxide samples. It is observed that the manganese oxide particles derived hydrothermally at 140 °C for 12 h (Figure 2a and b), 240 °C for 24 h (Figure 2c and d), and 120 °C for 12 h (Figure 2e and f) were, respectively, rod-, wire-, and tube-like in morphology, whereas those obtained with CCl4 solution displayed a flower-like spherical shape with sharp edges. It should be noted that the wire-like morphology can be differentiated from the rod-like morphology in terms of the bending or straight shape. The diameter and length of the rods in the rod-like α-MnO2 sample were ca. 43 nm and 2−4 μm, those of the wires in the wire-like α-MnO2 sample were ca. 40 nm and 1−10 μm, and those of the tubes in 4035
dx.doi.org/10.1021/es204038j | Environ. Sci. Technol. 2012, 46, 4034−4041
Environmental Science & Technology
Article
Figure 2. SEM images of (a, b) rod-like MnO2, (c, d) wire-like MnO2, (e, f) tube-like MnO2, and (g, h) flower-like Mn2O3.
Figure 3. TEM and high-resolution TEM images as well as SAED patterns (insets) of (a, b) rod-like MnO2, (c, d) wire-like MnO2, (e, f) tube-like MnO2, and (g, h) flower-like Mn2O3.
the tubular α-MnO2 sample were ca. 65 nm and 1−3 μm, respectively. For the Mn2O3 sample, the size of the flower-like spheres was in the range of 800−1000 nm. Shown in Figure 3 are the TEM and high-resolution TEM images as well as the SAED patterns of the manganese oxide samples. Well-grown nanorods (Figure 3a), nanowires (Figure 3c), and nanotubes (Figure 3e) of α-MnO2 could be clearly observed. The TEM images (Figure 3g and h) were recorded on the edge of a broken flower-like Mn2O3 nanoentity. From the high-resolution TEM images (Figure 3b, d, and f), one can
see well-resolved lattice fringes. The lattice spacings (d values) of the (121) crystal plane of the rod-, wire-, and tube-like αMnO2 samples were ca. 0.239, 0.238, and 0.239 nm, respectively, rather close to that (0.2388 nm) of the standard α-MnO2 sample (JCPDS PDF 72-1982). The d value (0.271 nm) of the (222) crystal plane of the flower-like spherical Mn2O3 sample estimated from the high-resolution TEM image (Figure 3h) was also not far away from that (0.2716 nm) of the referenced Mn2O3 sample (JCPDS PDF 41-1442). Furthermore, the recording of linearly aligned bright electron 4036
dx.doi.org/10.1021/es204038j | Environ. Sci. Technol. 2012, 46, 4034−4041
Environmental Science & Technology
Article
Table 1. Preparation Conditions, BET Surface Areas, and Surface Element Compositions of the Manganese Oxide Samples surface element molar ratio catalyst
preparation method
Mn source
rod-like MnO2 wire-like MnO2 tube-like MnO2 flower-like Mn2O3
hydrothermal treatment at 140 °C for 12 h hydrothermal treatment at 240 °C for 24 h hydrothermal treatment at 120 °C for 12 h In CCl4 solution at RT
KMnO4 KMnO4 and MnSO4 KMnO4 KMnO4 and MnCl2
a
calcination condition
BET surface area (m2/ g)
Mn3+/Mn4+
Oads/ Olatt
°C, °C, °C, °C,
83.0 83.2 44.8 162.3
0.58 0.31 0.49 4.17 (0.38)a
1.50 0.78 1.15 0.92
500 500 500 500
3 3 3 3
h h h h
The datum in parentheses is the surface Mn2+/Mn3+ molar ratio.
Figure 4. (A) Mn 2p3/2 and (B) O 1s XPS spectra of (a) rod-like MnO2, (b) wire-like MnO2, (c) tube-like MnO2, and (d) flower-like Mn2O3.
the product. Among the hydrothermally prepared manganese oxide samples, the rod-like α-MnO2 sample showed the highest surface Mn3+/Mn4+ molar ratio (0.58), whereas the lowest surface Mn3+/Mn4+ molar ratio (0.31) was achieved on the wire-like α-MnO2 sample. It is noted that there was also the copresence of Mn2+, Mn3+, and Mn4+ on the surface of the flower-like Mn2O3 sample due to the formation of tetragonal Mn2O3 and α-MnO2 phases, with the surface Mn3+/Mn4+ and Mn2+/Mn3+ molar ratios being 4.17 and 0.38, respectively. Based on the principle of electroneutrality, we deduce that the surface oxygen vacancy density was the highest on the rod-like α-MnO2 surface, while the lowest was on the wire-like α-MnO2 surface. Usually, oxygen molecules are adsorbed at the oxygen vacancies of an oxide material. Therefore, we believe that the oxygen adspecies locate at the surface oxygen vacancies of αMnO2 or Mn2O3. This result is in good agreement with the result of O 1s XPS investigations. The formation of surface oxygen vacancies on the α-MnO2 or Mn2O3 sample was beneficial for the oxidation of VOCs, which provides a good interpretation for the higher catalytic activity of rod-like αMnO2 at low temperatures (shown in Section 3.4). As can be seen from Figure 4B, the asymmetrical O 1s signal could be deconvoluted to two components: one at BE = 529.0 eV and the other at BE = 531.7 eV; the former was assigned to the surface lattice oxygen (Olatt) species, whereas the latter was assigned to the surface adsorbed oxygen (Oads) species.25−27,29 It is found from Table 1 that the estimated surface Oads/Olatt molar ratios of the samples were dependent upon the preparation method. The surface Oads/Olatt molar ratio decreased in the order of rod-like α-MnO2 (1.50) > tube-like α-MnO2 (1.15) > flower-like Mn2O3 (0.92) > wire-like αMnO2 (0.78). The formation of oxygen adspecies was due to
diffraction spots in the SAED patterns (insets of Figure 3b, d, and f) means that the tetragonal α-MnO2 samples with rod-like, wire-like, and tubular morphologies were single crystalline. For the flower-like spherical Mn2O3 sample, however, the SAED pattern (inset of Figure 3h) showed multiple bright electron diffraction rings, suggesting that this cubic Mn2O3 sample was mainly polycrystalline. As can be seen from Table 1, the BET surface areas (ca. 83 m2/g) of the rod- and wire-like α-MnO2 samples were similar, and much higher than that (ca. 45 m2/g) of the tubular α-MnO2 sample. However, the flower-like spherical Mn2O3 sample possessed a surface area of ca. 162 m2/ g, significantly higher than those of the hydrothermally derived α-MnO2 samples. The difference in preparation method led to a big difference in surface area of the MnOx catalysts with similar crystallinity, similar phenomena also took place in the preparation of α-Fe2O3 samples.31,32 XPS is a good tool to investigate the surface element composition, element oxidation state, and adsorbed species of a material. Figure 4 illustrates the Mn 2p3/2 and O 1s XPS spectra of the manganese oxide samples. As shown in Figure 4A, there was one asymmetrical signal at BE = ca. 642 eV for the three αMnO2 samples and at BE = ca. 641 eV for the Mn2O3 sample, in which the former could be decomposed to two components at BE = 641.6 and 642.8 eV, whereas the latter could be decomposed to three components at BE = 640.6, 641.6, and 642.8 eV. The components at BE = 640.6, 641.6, and 642.8 eV were attributable to the surface Mn2+, Mn3+, and Mn4+ species,7,33 respectively. A quantitative analysis on the Mn 2p3/2 XPS spectra of the samples gives rise to the surface Mn3+/ Mn4+ as well as Mn2+/Mn3+ molar ratios, as summarized in Table 1. Apparently, the preparation method had an important impact on the surface Mn3+/Mn4+ or Mn2+/Mn3+ molar ratio of 4037
dx.doi.org/10.1021/es204038j | Environ. Sci. Technol. 2012, 46, 4034−4041
Environmental Science & Technology
Article
Figure 5. (A) H2-TPR profiles and (B) initial H2 consumption rates of (a) rod-like MnO2, (b) wire-like MnO2, (c) tube-like MnO2, and (d) flowerlike Mn2O3.
the presence of surface oxygen vacancies on α-MnO2 or Mn2O3, which implies that there might be the coexistence of Mn3+ and Mn4+ ions in/on the α-MnO2 samples or Mn2+ and Mn3+ ions on/in the Mn2O3 sample.27,29 Such a deduction was supported by the Mn 2p3/2 XPS results of these samples. Reducibility. Figure 5A illustrates the H2-TPR profiles of the as-prepared manganese oxide samples. For the rod-like αMnO2 sample, there was a main reduction band at 265 °C with a shoulder at 285 °C, the total H2 consumption was 11.28 mmol/g (Table 2). However, only one strong reduction band
H2 consumptions (10.00−11.28 mmol/g) of the hydrothermally derived α-MnO2 samples and that (5.93 mmol/g) of the flower-like Mn2O3 sample were quite close to their theoretical H2 consumptions (11.50 and 6.30 mmol/g, respectively). This result indicates that a substantial fraction of Mn4+ in α-MnO2 or Mn3+ and Mn4+ in Mn2O3 had been reduced to Mn2+ below 400 °C. To better compare the lowtemperature reducibility of these samples, we calculated the initial H2 consumption rate of the first reduction band of each sample before the occurrence of phase transformation (where the initial H2 consumption of the first reduction band of the catalyst is less than 25%25−27,35,36), and the results are shown in Figure 5B. Obviously, the initial H2 consumption rate decreased in the sequence of rod-like α-MnO2 > tube-like α-MnO2 > flower-like Mn2O3 > wire-like α-MnO2. That is to say, the lowtemperature reducibility of these manganese oxide samples followed the above order. Catalytic Performance. In the blank experiment (only quartz sands were loaded), no conversion of toluene was detected below 400 °C, indicating that under the adopted reaction conditions there was no occurrence of homogeneous reactions. Figure 6 shows the catalytic performance of the bulk α-MnO2 sample (surface area = ca. 10 m2/g, Beijing Chemical Reagent Company, A.R., 99.9%) and the as-prepared α-MnO2 and Mn2O3 samples for the combustion of toluene. Under the conditions of toluene concentration =1000 ppm, toluene/O2 molar ratio = 1/400, and SV = 20 000 mL/(g h), toluene conversion increased with the rise in reaction temperature, and the α-MnO2 and Mn2O3 catalysts with various morphologies performed much better than the bulk α-MnO2 catalyst. It is worth pointing out that toluene was completely oxidized to CO2 and H2O over the as-prepared α-MnO2 and Mn2O3 catalysts, and there was no detection of products of incomplete oxidation, as confirmed by the good carbon balance of ca. 99.5% in each run. It is convenient to compare the catalytic activities of these samples by using the reaction temperatures T10%, T50%, and T90% (corresponding to the toluene conversion = 10, 50, and 90%), as summarized in Table 2. It is clearly seen that rod-like α-MnO2 was inferior to wire- and tube-like αMnO2 and flower-like Mn2O3 in catalytic performance at lower
Table 2. Reduction Temperatures, H2 Consumptions, and Catalytic Activities of the Manganese Oxide Samples reduction temperature (°C) catalyst bulk MnO2 rod-like MnO2 wire-like MnO2 tube-like MnO2 flower-like Mn2O3
catalytic activity (°C) H2 consumption (mmol/g)
T10%
T50%
T90%
11.28
225 176
292 210
322 225
275
10.00
143
225
245
268
10.95
157
222
233
5.93
145
226
238
band 1 band 2 265
260
285
332
centered at 275 °C for the wire-like α-MnO2 sample and at 268 °C for the tube-like α-MnO2 sample was recorded, with the total H2 consumption being 10.00 and 10.95 mmol/g (Table 2), respectively. In the case of the flower-like spherical Mn2O3 sample, there were two weaker reduction bands at 260 and 332 °C, corresponding to a total H2 consumption of 5.93 mmol/g (Table 2). According to the results reported previously,28,34 the reduction process could be reasonably divided into two steps: (i) Mn4+ → Mn3+ and (ii) Mn3+ → Mn2+. Theoretically, the H2 consumptions for the reduction of MnO2 to Mn3O4 and of Mn3O4 to MnO are 7.67 and 3.83 mmol/g, respectively; while a H2 consumption of 6.30 mmol/g is needed if the Mn2O3 is completely reduced to MnO. In the present studies, the total 4038
dx.doi.org/10.1021/es204038j | Environ. Sci. Technol. 2012, 46, 4034−4041
Environmental Science & Technology
Article
temperature, a larger amount of O− species might be available through the conversion of O2− and O22− to O− species on the rod-like MnO2 catalyst,37,39 giving rise to a great enhancement in catalytic activity. The T50% and T90% values for the rod-like αMnO2 catalyst were 82 and 97 °C lower than those for the bulk α-MnO2 catalyst, respectively. Therefore, it is concluded that in terms of T50% and T90% values, the catalytic performance decreased in the order of rod-like α-MnO2 > tube-like α-MnO2 > flower-like Mn2O3 > wire-like α-MnO2, coinciding with the sequences of oxygen adspecies concentration obtained in the XPS studies and of low-temperature reducibility revealed by the H2-TPR investigations.25−27,29 Figure 7A and B shows the effects of SV and toluene/O2 molar ratio on the catalytic activity of the rod-like α-MnO2 sample, respectively. It is observed that the catalytic activity of rod-like α-MnO2 decreased with the rise in SV value (Figure 7A) or toluene/O2 molar ratio (Figure 7B). Obviously, the rise in O2 concentration of the reactant feed favored the enhancement of toluene conversion, suggesting that the oxygen adspecies might play an important role in the total oxidation of toluene. That is to say, the oxygen nonstoichiometry relevant to structural defects might be a critical factor in determining the catalytic activity of manganese oxide.25−30 To examine the catalytic stability of the rod-like α-MnO2 sample, we carried out the on-stream reaction experiment at 225 °C and the result is shown in Figure S1 of the Supporting Information. It is found that there was no significant decline in catalytic activity within 60 h of on-stream reaction. Hence, we believe that the rod-like α-MnO2 sample was catalytically durable. In the past years, a number of materials have been used as catalysts for the oxidative removal of toluene. It was reported that under similar conditions for the combustion of toluene, the T50% and T90% values were 245−340 and 265−375 °C over the commercial Mn3O4, Mn2O3, or MnO2 catalyst at SV = 15 000 mL/(g h),15 140−200 and 234−240 °C over the mesoporous CrOx or MnO2 catalyst at SV = 20 000 h−1,26,28 254−279 and 295−306 °C over the LaMnO3 or LaCoO3 catalyst at SV = 178 h−1,40,41 270 and 300 °C over the 5 wt % Au/CeO2 catalyst at SV = 186 h−1,42 and 180 and 250 °C over the 0.5 wt % Pd/
Figure 6. Toluene conversion as a function of reaction temperature over the catalysts under the conditions of toluene concentration = 1000 ppm, toluene/O2 molar ratio = 1/400, and SV = 20 000 mL/(g h).
temperatures (
