Terminal Hydroxyl Groups on Al2O3 Supports Influence the Valence State and Dispersity of Ag Nanoparticles: Implications for Ozone Decomposition.

Ozone is a poisonous gas, so it is necessary to remove excessive ozone in the environment. Catalytic decomposition is an effective way to remove ozone at room temperature. In this work, 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 catalysts were synthesized by the impregnation method. The 10%Ag/nano-Al2O3 catalyst showed 89% ozone conversion for 40 ppm O3 for 6 h under a space velocity of 840 000 h-1 and a relative humidity of 65%, which is superior to 10%Ag/AlOOH-900 (45% conversion). The characterization results showed Ag nanoparticles to be the active sites for ozone decomposition, which were more highly dispersed on nano-Al2O3 as a result of the greater density of terminal hydroxyl groups. The understanding of the dispersion and valence of silver species gained in this study will be beneficial to the design of more efficient supported silver catalysts for ozone decomposition in the future.

Introduction

Due to the deterioration of the environment, ozone has increasingly drawn our attention. Outdoors, the photochemical reaction of VOCs and NO in the atmosphere produces ozone, which leads to excessive ozone concentrations in the near-surface troposphere.[1,2] Indoors, air cleaners, printers, and copiers can release ozone.[3] In 2005, the World Health Organization (WHO) stipulated that the 8 h average ozone concentration in the working environment shall not exceed 100 μg/m3. Excessive ozone seriously threatens people’s health[4] and reduces crop yields.[5] Thus, the disposal of excess ozone is a pressing problem.

There are a variety of methods to remove ozone, such as active carbon absorption, liquid absorption, thermal decomposition, and catalytic decomposition. Among these, catalytic decomposition is considered to be the most efficient, safe, and economical method. Catalysts for ozone decomposition mainly include transition metal oxides (e.g., MnO,[6−9] CoO,[10] Fe2O3,[11] NiO,[12] and ZnO[13]) and noble metals (e.g., Au,[14,15] Ag,[16−18] Pt,[19] Pd,[20,21] and Rh[22]). However, most transition metal oxides have the disadvantages of poor moisture resistance and eventual deactivation. Noble metals can effectively improve the water resistance and increase the lifetime of the catalysts. Ag-based catalysts are much cheaper than other noble metal catalysts and are widely used. Ag/perlite, Ag/SiO2, Ag/α-Al2O3, and Ag/clinoptilolite have been investigated. The results indicate that Ag species are the active sites for ozone decomposition.[23−26] Li et al.[16] confirmed that metallic silver particles (Ag0) on Ag–Mn catalysts showed much better ozone removal performance than Ag1.8Mn8O16 and Agδ+ species. Kumar et al.[27,28] prepared Ag/MCM-41-20, Ag/H-MCM-41-50, and Ag/H-beta-11 and confirmed that the acidic properties and structure of catalyst supports had an effect on the state of Ag on these ozone decomposition catalysts. However, it was not clear how the support could influence the state of silver and then have an impact on the activity of catalysts.

γ-Al2O3 has often been used as a nonreactive catalyst support because it is easy to determine the active sites and the reaction mechanism in catalytic studies. Ag has typically been used as the active component in catalysts because of its low price and good performance. Thus, Ag/γ-Al2O3 has been widely studied in a variety of catalytic reactions due to its excellent performance and easy preparation. Examples include the catalytic oxidation of HCHO[29,30] and selective catalytic reduction and selective catalytic oxidation of NO.[31−34] Wang et al.[35] compared the state of Ag on 1%Ag/nano-Al2O3 and 1%Ag/micro-Al2O3 and found that the use of nano-Al2O3 as a support can lead to single-atom silver dispersion, while Ag on micro-Al2O3 existed in the form of Ag clusters. They confirmed that Ag was anchored by terminal hydroxyl groups and that abundant terminal hydroxyl groups on a support would result in high dispersion and utilization of Ag.

In this work, we successfully synthesized 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 catalysts by the impregnation method. The characterization results showed that silver in the metallic state was a superior active site for ozone decomposition and that the abundant terminal hydroxyl groups on nano-Al2O3 led to high dispersion of Ag.

Results and Discussion

Ozone Decomposition Performance

The ozone decomposition performance of AlOOH-900, nano-Al2O3, 10%Ag/AlOOH-900, and 10%Ag/nano-Al2O3 is displayed in Figure . The AlOOH-900 and nano-Al2O3 supports had no ozone removal ability at room temperature under the conditions of 40 ppm ozone concentration and RH = 65%. The results were consistent with the study of Tao et al.,[21] who found that the ozone conversion of Al2O3/Al-fiber was nearly zero. When silver at 10 wt % was supported on AlOOH-900 and nano-Al2O3, the ozone conversion was improved dramatically. After 6 h, the activity of 10%Ag/nano-Al2O3 was up to 90% and remained stable. However, 10%Ag/AlOOH-900 only exhibited 45% ozone conversion, which was about half that of 10%Ag/nano-Al2O3. The above phenomenon was possibly related to the state, dispersity, or size of silver species on the support.

Figure 1

Ozone conversion of nano-Al2O3, AlOOH-900, 10%Ag/nano-Al2O3, and 10%Ag/AlOOH-900; Conditions: ozone inlet concentration, 40 ppm; temperature, 30 °C; relative humidity, 65%; and space velocity, 840 000 h–1.

Specific Surface Area and the Crystal Structure

Figure shows the N2 adsorption–desorption isotherms, X-ray diffraction (XRD) patterns, and morphology of the samples. As shown in Figure a, the isotherms of all samples were type IV with a hysteresis loop of type H3 (according to IUPAC). There was no change in the type of pore structure after silver was impregnated on the samples. As shown in Table , the specific surface areas of nano-Al2O3 and AlOOH-900 were similar, 176.7 and 199.8 m2/g, respectively. The specific surface area declined after the addition of Ag for both supports, but the area for 10%Ag/AlOOH-900 was still much larger, indicating that the specific surface area was not the main factor in the ozone decomposition activity. As shown in Figure b, nano-Al2O3 and AlOOH-900 had the same crystalline structure as γ-Al2O3. Figure c,d shows that nano-Al2O3 had a sheet-like morphology with a particle size of about 10 nm, and AlOOH-900 had a morphology consisting of sheets and rods longer than 20 nm.

Figure 2

(a) N2 adsorption–desorption isotherms and pore-size distribution curves (inset) of nano-Al2O3, AlOOH-900, 10%Ag/nano-Al2O3, and 10%Ag/AlOOH-900; (b) XRD patterns of nano-Al2O3 and AlOOH-900; (c) high-resolution transmission electron microscopy (HRTEM) of nano-Al2O3; and (d) HRTEM of AlOOH-900.

Table 1

Brunauer–Emmett–Teller (BET) Surface Area, Pore Size, and Pore Volume of the Four Samples

sample	SBET [m2/g]	pore diameter (d) [nm]	pore volume (V) [cm3/g]
nano-Al2O3	176.7	13.5	0.7
AlOOH-900	199.8	13.5	0.8
10%Ag/nano-Al2O3	160.5	13.1	0.6
10%Ag/AlOOH-900	177.0	12.5	0.6

Chemical States and Dispersion of Silver

Figure shows the XRD patterns of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900. The diffraction peaks at 38.1, 44.3, 64.4, and 77.5° correspond to the (111), (200), (220), and (311) lattice planes of the Ag metal (JCPDS 87-0717),[34] respectively. Also, it could be observed that the peak intensity of Ag on 10%Ag/AlOOH-900 was stronger than that on 10%Ag/nano-Al2O3. The results may be due to the coalescence of some Ag nanoparticles or a better degree of crystallinity of Ag on AlOOH-900. In addition, the diffraction peaks at 19.3, 32.3, 33.8, and 37° assigned to AgO species appeared on 10%Ag/AlOOH-900, indicating that part of Ag was oxidized and may form AgO particles. Perhaps this was the reason why the ozone conversion of 10%Ag/AlOOH-900 was worse than that of 10%Ag/nano-Al2O3.

Figure 3

XRD patterns of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900.

Figure shows the HRTEM images of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900. As shown in Figure a,c, many silver nanoparticles with a size of 4–5 nm were anchored on nano-Al2O3, displaying an unclear lattice fringe, but the silver nanoparticles began to coalesce to form larger particles with a size of 10–15 nm with clearer lattice fringes on 10%Ag/AlOOH-900, which may have adverse effects on ozone decomposition. The results further confirmed the above XRD results. The lattice spacings of Ag nanoparticles on 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 were measured and are shown in Figure b,d. Fringe spacings of 0.236 and 0.228 nm on sliver particles may represent the (111) plane of Ag0 and the (2̅02) plane of AgO, respectively.[36]

Figure 4

(a, b) HRTEM image of 10%Ag/nano-Al2O3 and (c, d) HRTEM image of 10%Ag/AlOOH-900.

X-ray photoelectron spectroscopy (XPS) spectra were then measured to confirm the state of silver. Figure a shows the Ag 3d spectra of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900, which showed that the binding energies of Ag 3d5/2 for 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 were 368.3 and 368.2 eV. The peak of Ag 3d on 10%Ag/nano-Al2O3 was located at a higher binding energy, indicating that there was more Ag0 in the Ag nanoparticles. According to the inductively coupled plasma optical emission spectroscopy (ICP-OES) data in Table , the content of Ag for 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 in the bulk was 8.6 and 8.4%, respectively. The Ag content on the surface of the two catalysts was measured by XPS as 10.97 and 8.94%, respectively, as shown in Table . The above results indicated that the silver particles were more dispersed on the surface of nano-Al2O3, which was consistent with the results of HRTEM and XRD. To determine the percentage of Ag0 in the silver particles, the peaks of Ag 3d5/2 were deconvoluted into two peaks at 368.7 and 367.9 eV, which were deemed as arising from silver in the metallic and oxidized states[37−39] displayed in Figure b,c. The percentages of Ag0 in 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 were 53 and 39%, respectively (Table ), indicating that 10%Ag/nano-Al2O3 had more Ag in the metallic state, which was beneficial to ozone removal.[16]

Figure 5

(a) Ag 3d XPS spectra of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900; (b) Ag 3d5/2 XPS spectra of 10%Ag/nano-Al2O3; and (c) Ag 3d5/2 XPS spectra of 10%Ag/AlOOH-900.

Table 2

ICP-OES and XPS Results of Ag Species of 10%Ag/Nano-Al2O3 and 10%Ag/AlOOH-900

	ICP-OES	XPS
sample	Ag (wt %)	Ag (wt %)	Agn0 (%)	Agnδ+ (%)
10%Ag/nano-Al2O3	8.6	10.97	53	47
10%Ag/AlOOH-900	8.4	8.94	39	61

UV–vis DRS spectra of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 were obtained by subtracting the spectrum of bare BaSO4. As can be seen in Figure , all samples have a broad absorption band from 200 to 700 nm. The bands at 209 and 232 nm can be attributed to the Ag+ species.[40] The band at 289 nm is attributed to Agδ+.[41] The broad bands at 364 and 491 nm belong to Ag0 and AgNPs, respectively.[42] It is obvious that 10%Ag/nano-Al2O3 has a stronger Ag0 absorption band and a weaker AgNPs absorption band than those of 10%Ag/AlOOH-900. These results indicate that 10%Ag/nano-Al2O3 has more small Ag0 particles in the metallic state, but the Ag species on 10%Ag/AlOOH-900 coalesce to form larger particles.

Figure 6

UV–vis profiles of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900.

H2-Temperature-programmed reduction (H2-TPR) (Figure ) was used to characterize the relative content of Ag in the oxidized state. The peaks at 117 and 131 °C corresponded to the reduction of large AgO clusters; the peaks at 199 and 212 °C were due to the reduction of small AgO clusters.[43] Compared to the 10%Ag/AlOOH-900 catalyst, 10%Ag/nano-Al2O3 had less oxidized Ag and a lower reduction temperature. The AgO species on 10%Ag/nano-Al2O3 were not detected by XRD or HRTEM, possibly due to their highly dispersed or ultrathin amorphous outer layers over the surface of silver nanoparticles, and they did not form AgO particles like that on 10%Ag/AlOOH-900.[34,44]

Figure 7

H2-TPR image of 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900.

Hydroxyl Group Content

The anchoring mechanism of noble metals on supports has been investigated. The effect of electronic interactions between reducible oxides and noble metals was considered to be an important factor affecting metal anchoring on the supports.[45] However, for nonreducible supports, there may be other reasons for the anchoring of metal species on the supports. Kwak et al.[46] reported that unsaturated pentacoordinate Al3+ (Al3+ penta) on the surface of γ-Al2O3 was the anchoring site of Pt. Wang et al.[35] found that terminal hydroxyl groups on γ-Al2O3 were the anchoring points of Ag species. To study the anchoring mechanism of Ag on Al2O3 further, in situ diffuse reflectance infrared spectroscopy (DRIFTS) with NH3 adsorption was performed to measure the content of OH groups.

Figure shows the spectra of in situ DRIFTS with NH3 adsorption (Figure a) and the peak fitting (Figure b,c) results for the wavenumbers from 3850 to 3590 cm–1. As shown in Figure a, the peak at about 3771 cm–1 was due to terminal hydroxyl groups (type I), the peak at about 3729 cm–1 was derived from doubly bridging hydroxyls (type II), and the peak at about 3675 cm–1 was due to triply bridging hydroxyl groups (type III).[35,47,48] Compared with AlOOH-900, nano-Al2O3 had more terminal hydroxyl groups. After Ag was impregnated on the supports, the peaks for terminal hydroxyls (type I) of nano-Al2O3 and AlOOH-900 decreased sharply and nearly disappeared, indicating that Ag consumed almost all of the terminal hydroxyl groups on the supports. The result further indicated that terminal hydroxyl groups were the main anchoring sites of Ag. Besides, the peak intensity due to type II hydroxyl groups on 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900 became stronger, which may be because of the generation of a new type of hydroxyl group after Ag was added to the surface. There was no change in the peak intensity of type III hydroxyl groups on nano-Al2O3 and AlOOH-900, indicating that they were not the main anchoring sites for Ag nanoparticles. Besides the three kinds of hydroxyl groups (types I, II, and III) on nano-Al2O3 and AlOOH-900, a peak for a new type of doubly bridging hydroxyl group generated at about 3748 cm–1, marked as type II′, appeared on 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900, as shown in Figure b,c. To see the proportion of each hydroxyl group a bit more intuitively, the percentage composition of each type of hydroxyl group was calculated, as shown in Table . The proportion of terminal hydroxyl groups (type I) on nano-Al2O3 was 38%, higher than that on AlOOH-900 (15%). As for 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900, type I contents were 10 and 7%, respectively. As indicated by the results above, nano-Al2O3 had more terminal hydroxyl groups available to anchor Ag particles, which were the active sites of ozone decomposition. Therefore, the ozone conversion of 10%Ag/nano-Al2O3 was higher than that of 10%Ag/AlOOH-900.

Figure 8

(a) In situ DRIFTS spectra of NH3 adsorption on the four samples; (b) peak fitting of OH consumption peaks after in situ DRIFTS of NH3 adsorption over nano-Al2O3 and 10%Ag/nano-Al2O3; and (c) peak fitting of OH consumption peaks after in situ DRIFTS of NH3 adsorption over AlOOH-900 and 10%Ag/AlOOH-900.

Table 3

Proportion of OH Consumption for Four Types of Hydroxyl Groups According to the Peak Fitting Results of In Situ DRIFTS of NH3 Adsorption

sample	type I (%)	type II (%)	type III (%)	type II′ (%)
nano-Al2O3	38	25	37
AlOOH-900	15	57	28
10%Ag/nano-Al2O3	10	33	39	18
10%Ag/AlOOH-900	7	53	35	5

Conclusions

This study compared two kinds of Ag-containing catalysts, 10%Ag/nano-Al2O3 and 10%Ag/AlOOH-900, which were prepared by the impregnation method. The ozone conversion of 10%Ag/nano-Al2O3 was 89%, higher than that of 10%Ag/AlOOH-900 (45%). The characterization results showed that nano-Al2O3 had nearly four times the number of terminal hydroxyl groups as AlOOH-900, leading to better dispersion of Ag particles with the metallic state (Ag0), which were superior active sites for ozone decomposition. Thus, 10%Ag/nano-Al2O3 had better ozone removal performance. The understanding of the dispersion and valence of silver species detailed in this paper will be beneficial to the design of more efficient supported silver catalysts for ozone decomposition in the future.

Materials and Methods

Preparation of Catalysts

The AlOOH-900 support was prepared by calcining commercial boehmite at 900 °C for 3 h, and the nano-Al2O3 support was purchased (Aladdin). Catalysts with a silver loading of 10% were prepared by impregnating nano-Al2O3 or AlOOH-900 in silver nitrate solution. After stirring for 2 h on a magnetic stirring apparatus, the sample was evacuated on a rotary evaporator at 333 K for 1 h until the water evaporated. Then, the catalysts were dried at 100 °C for about 12 h and calcined at 500 °C for 3 h in air (Scheme ).

Scheme 1

Catalyst Characterization

The BET surface areas and pore properties of the samples were obtained by a Quantachrome Quadrasorb SI-MP analyzer. The Brunauer–Emmett–Teller (BET) method was used to calculate the specific surface area of the samples. The Barrett–Joyner–Halenda (BJH) method was used to determine the diameter and volume of pores. After outgassing at 300 °C for 4 h, N2 was adsorbed on the samples at −196 °C to obtain the adsorption and desorption curves.

The crystalline structure of the catalysts was determined by X-ray powder diffraction (XRD) (D8-Advance, Bruker, Germany) with Cu Kα radiation (λ = 0.15406 nm) operated at 40 kV and 40 mA. The scan range was set from 5 to 90° with a step size of 0.02°.

High-resolution transmission electron microscopy (HRTEM) images were obtained on a JEOL JEM 2010 TEM with an acceleration voltage of 200 kV.

The silver content of the catalysts was determined by an inductively coupled plasma emission spectrometer (ICP-OES) (720, Varian). The elemental composition, content, and valence state of species on the surface of the samples were analyzed by X-ray photoelectron spectra (XPS) (Axis UItra, Kratos Analytical Ltd., U.K.) using Al Kα radiation.

The valence state of silver was analyzed by diffuse reflectance UV–vis spectra. The spectra were measured at room temperature in air with BaSO4 as reference (U-3100 UV–vis spectrophotometer, Hitachi Co., Japan) and were collected in the range of 200–700 nm with a resolution of 1 nm.

The hydroxyl content was determined by an in situ DRIFTS system (Nexus 670, Thermo Nicolet) equipped with an MCT/A detector. Prior to NH3 absorption (500 ppm), the sample was pretreated at 350 °C for 1 h in 10%O2/N2 (500 mL/min) and cooled to room temperature.

The relative content of silver oxide of the catalysts was obtained by H2 temperature-programmed reduction (H2-TPR) experiments performed on a Micromeritics Autochem II 2920 equipped with a thermal conductivity detector (TCD). The samples (100 mg) were pretreated at 300 °C for 30 min in Ar (30 mL/min) and then were cooled to 30 °C. Next, after the baseline had stabilized, the catalysts were heated in 10%H2/Ar (30 mL/min) from 30 to 400 °C (10 °C/min).

Catalysts’ Ozone Decomposition Activity

Catalysts of 40–60 mesh with a mass of 100 mg were placed into a fixed-bed continuous flow quartz reactor (i.d., 4 mm). Then, all of the samples were tested at 30 °C with a gas flow of 1.4 L/min and RH = 65%. The concentration of ozone was 40 ± 2 ppm and the space velocity was about 840 000 h–1. The ozone was generated by low-pressure ultraviolet lamps, and the inlet and outlet ozone concentrations were detected by an ozone monitor (model 202, 2B Technologies). A humidity probe (HMP110, Vaisala OYJ, Finland) was used to control the humidity at 65%. The formula used to compute ozone conversion was as followswhere Cin and Cout are the inlet and outlet concentrations of ozone, respectively.