CO Oxidation Activities of Ni and Pd-TiO₂@SiO₂ Core-Shell Nanostructures

Abstract

We prepared Ni and Pd-modified $TiO_2@SiO_2$ core-shell nanostructures and then analyzed them by scanning electron microscopy, optical microscopy, X-ray diffraction crystallography, FT-IR and UV-Visible absorption spectroscopy. In addition, their CO oxidation performance was tested by temperature-programmed mass spectrometry. The CO oxidation activity showed an order of Ni-$TiO_2@SiO_2$ ($900^{\circ}C$) < Ni-$TiO_2@SiO_2$ ($90^{\circ}C$) < Ni-$TiO_2@SiO_2$ ($450^{\circ}C$) in the first CO oxidation run, and greatly improved activity in the same order in the second run. The $T_{10%}$ (the temperature at 10% CO conversion) corresponds to the CO oxidation rate of $2.8{\times}10^{-5}$ molCO $g{_{cat}}^{-1}s^{-1}$. For Ni-$TiO_2@SiO_2$ ($450^{\circ}C$), the $T_{10%}$ was observed at $365^{\circ}C$ in the first run and at $335^{\circ}C$ in the second run. For the Pd-$TiO_2@SiO_2$ ($450^{\circ}C$), the $T_{10%}$ was observed at a much lower temperature of $263^{\circ}C$ in the first CO oxidation run, and at $247^{\circ}C$ in the second run. The CO oxidation activities of transition metal modified $TiO_2@SiO_2$ core-shell nanostructures presented herein provide new insights that will be useful in developing catalysts for various environments.

Introduction

Removing toxic gases is very important for clean environments and respiratory biological systems.1-3 Carbon monoxide (CO) is a toxic gas commonly produced from incomplete combustion of fossil fuels. Catalytic converters have played a major role in eliminating CO emission via the oxidation of CO to nontoxic CO2 using a catalyst.3 The CO oxidation reaction over transition metal oxide catalysts has been studied extensively in efforts to improve catalytic performance. 1 Among the many factors involved in oxidation performance, oxide supports and thermal pretreatments have been shown to be important key factors.4-6 The use of silica (SiO2) spheres as an oxide support in the development of catalysts has attracted a great deal of attention in recent years.7-18 For example, Yan et al. prepared Co3O4 nanoparticles in SiO2 nanocapsules and found a 100% CO conversion rate at 150 ℃,7 which was much better than that observed for other Co3O4 nanostructures such as nanowires with exposed reactive {111} planes. Ni@SiO2, Co@SiO2 and Fe@SiO2 yolk-shell structures were prepared by Park et al. and shown to be good for steam reforming of methane with high temperature stability and recyclability.11 Kim et al. coated SiO2 with TiO2 and ZrO2 using a sol-gel method and embedded Pt. They found that Pt-support interaction became much stronger, and that CO oxidation performance was superior to that of uncoated Pt/SiO2 and Pt/Al2O3.8 Dye degradation under UV irradiation was tested for a transition metal catalyst with and without TiO2/SiO2 hybrid structures. The activity showed the order of Co-TiO2/SiO2 >> Co-TiO2 > TiO2/SiO2 > Co-SiO2 > TiO2 > SiO2,12 reflecting the fact that the TiO2/SiO2 hybrid structure shows a higher synergic effect due to increased surface defects through the formation of Si-O-Ti bonds. AuCuOx hybrid nanocrystals were supported on silica, which resulted in the catalyst becoming more active and selective for acetaldehyde formation from ethanol.5 SiO2 has also been used to support Au and Ag nanoparticles.13-18 In tests of the use of Ag/SiO2 and Au/SiO2 catalysts for CO oxidation, size and thermal pretreatment were shown to be important factors.

In the present study, we employed nonreducible SiO2 as a base oxide support, and then coated the surface with reducible TiO2. In this process, TiO2 acts as a good interfacial binder for transition metals and surface oxygen defect sites. Additionally, Ni and Pd were used as transition metals for comparison and different annealing temperatures were applied before testing the CO oxidation activity. We chose the two different metals because they are positioned in the same group 10 of the periodic table, but have different outer electron configurations of [Ar] 4s2 3d8 (or [Ar] 4s1 3d9) and [Kr] 4d,10 respectively. However, the oxidation state of 2+ for NiO and PdO show the same d8 electron configuration.

 

Experimental

SiO2 sphere oxide supports were synthesized by the Stöber method.19 To coat the SiO2 surface with TiO2, we dispersed 1.0 g of SiO2 powders in 50 mL ethanol, added 1.0 mL of titanium tetraisopropoxide, and then ultra-sonicated the mixed solution for 3 h, which yielded white TiO2@SiO2 catalysts. After the reaction, the powder products were first fully washed with ethanol only, then redispersed in 25.0 mL ethanol. Next, 10.0 mL of 0.1 M Ni(II) or 0.01 M Pd(II) solution was added to deposit Ni or Pd on the TiO2@SiO2 core-shell structures, after which the mixed solution was ultra-sonicated for 3 h. The final powder products were fully washed with water and ethanol repeatedly. The powder samples were then dried at 90 ℃ in an oven, which yielded green and gray Ni-TiO2@SiO2 and Pd-TiO2@SiO2, respectively. The dried Ni-TiO2@SiO2 samples were subsequently annealed at 450 ℃ and 900 ℃, while the Pd-TiO2@SiO2 sample was annealed at 450 ℃. All samples were annealed in a conventional furnace. The surface morphologies and the colors of the as-prepared and annealed samples were examined by scanning electron microscopy (SEM, Hitachi S- 4800) and optical microscopy, respectively. The X-ray diffraction (XRD) patterns were generated using a PANalytical X’Pert Pro MPD diffractometer with Cu Kα radiation (40 kV and 30 mA). UV-Vis absorption spectra for the powder samples were obtained using a Varian Cary 5000 UV-Vis- NIR spectrophotometer. FT-IR spectra were measured using a Thermo Scientific Nicolet iS10 spectrometer in attenuated total reflection mode. The CO oxidation performance was tested using a SRS RGA200 quadrupole mass spectrometer at a heating rate of 20 K/min under a mixture of CO (1%) and O2 (2.5%) in N2, with a flow rate of 40 mL/min. For the CO oxidation in a flow reactor, we used 10 mg of the sample in a quartz U-tube (with an inner diameter of 4 mm).

 

Results and Discussion

Figure 1 shows the SEM images of bare SiO2, TiO2-coated SiO2 (TiO2@SiO2 core-shell), Ni and Pd-modified TiO2@SiO2 core-shells annealed at various temperatures. The SEM images of bare SiO2 show a smooth surface and fairly uniform sphere shape with a diameter of about 200-300 nm. For the TiO2-coated SiO2, the surface became rougher and larger. The bare SiO2 and TiO2@SiO2 were white. Upon modification with Ni (Ni-TiO2@SiO2, 90 ℃-dried), the surface and the color of TiO2@SiO2 particles were dramatically changed, and the SiO2 spheres became aggregated. The color of the sample was pale green. Upon annealing at 450 ℃, the over-coated Ni complex decomposed to form islands on the TiO2@SiO2 surface. The islands became bigger upon annealing at a higher temperature of 900 ℃. The colors of the 450 ℃ and 900 ℃ annealed samples (Ni-TiO2@SiO2, 450 ℃ and 900 ℃) showed gray-light green and yellowgreen colors, respectively. Upon Pd modification (Pd- TiO2@SiO2, 90 ℃-dried), the color changed to gray. Upon annealing at 450 ℃ (Pd-TiO2@SiO2, 450 ℃), the color changed to light-chocolate, and islands on the TiO2@SiO2 surface were observed.

Figure 1.SEM images of bare SiO2, TiO2@SiO2 core-shell, 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni-TiO2@SiO2 core-shell catalysts, and 90 ℃-dried, 450 ℃-preannealed Pd-TiO2@SiO2 coreshell catalysts. The optical microscope images show the colors of the corresponding samples.

Figure 2.Powder X-ray diffraction patterns of bare SiO2 spheres, TiO2@SiO2 core-shell structures, 90 ℃-dried, 450 ℃ and 900 ℃ preannealed Ni-TiO2@SiO2 catalysts, and 90 ℃-dried and 450 ℃ preannealed Pd-TiO2@SiO2 catalysts.

Figure 2 displays the XRD patterns of bare SiO2 sphere particles, TiO2@SiO2 core-shell structures, 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni-TiO2@SiO2 catalysts, and 90 ℃-dried and 450 ℃-preannealed Pd-TiO2@SiO2 catalysts. For bare SiO2, a very broad peak was found at around 22°, which is consistent with previous studies.19 Upon coating with TiO2 followed by annealing at 450 ℃ (TiO2@SiO2 core-shell structures), sharper peaks were found at 2θ = 25.3°, 37.9°, 48.2°, 54.6°, and 62.7°. These findings are in good agreement with those of the tetragonal anatase TiO2 structure (JCPDS 1-084-1286), and were assigned to (011), (004), (020), (121) and (024) planes, respectively. Upon modifying the surface of TiO2@SiO2 catalysts with Ni and drying at 90 ℃, various sharp XRD peaks were observed, possibly due to Ni complex formation on the surface. Upon annealing at 450 ℃, three strong peaks appeared at 2θ = 37.3°, 43.3°, and 62.9°. These peaks were found to correspond to the (101), (012), and (110)/(104) planes of rhombohedral (R-3m) NiO (JCPDS 44-1159, a = b =2.96 Å, c = 7.23 Å).20 Further annealing at 950 °C resulted in new XRD patterns in addition to those of NiO. The XRD peaks matched those of rhombohedral (R-3) NiTiO3 (JCPDS 1-085- 0451, a = b = 5.45 Å) well.21 These findings indicate that nickel titanium oxide was formed at the interface between NiO and TiO2@SiO2. The XRD peak of TiO2 was not observed due to the reaction between TiO2 and thick Ni forming NiTiO3.22

Figure 3 displays the UV-visible reflectance absorption spectra of bare SiO2 sphere particles, 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni-TiO2@SiO2 catalysts, and 90 ℃dried and 450 ℃-preannealed Pd-TiO2@SiO2 catalysts. The optical microscope images (inset) show the colors of the corresponding samples. The Kubelka-Munk method was used to convert the diffuse reflectance to the absorbance (YFigureaxis). For bare white SiO2, no discernible absorption peak was observed in the UV-Visible region. For the 90 ℃-dried Ni-TiO2@SiO2 catalyst, two broad absorption regions were found at around 250 nm and 650 nm. The broad band at around 650 nm was related to the absorption of Ni(II) ions. The band below 400 nm was likely due to a band gap absorption of the samples.24 For the 450 ℃-preannealed Ni- TiO2@SiO2 sample, the absorption band of NiO was very broad, and covered the entire visible range.24 As a result, the sample appeared gray. The UV-vis absorption spectrum of the 900 ℃-preannealed Ni-TiO2@SiO2 sample was similar to that of the 90 ℃-dried Ni-TiO2@SiO2 sample, but shifted to a longer wavelength. The band at 450 nm could be attributed to the Ni(II)Ti(IV) charge-transfer band arising from NiTiO3.25 The 90 ℃-dried Pd-TiO2@SiO2 catalyst was gray and its absorption covered the entire visible range. For the 450 ℃-annealed Pd-TiO2@SiO2 sample, the broad absorption at around 470 nm was attributed to PdO.26

Figure 3.UV-Visible diffuse reflectance absorption spectra of bare SiO2 spheres, 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni- TiO2@SiO2 catalysts, and 90 ℃-dried and 450 ℃-preannealed Pd-TiO2@SiO2 catalysts.

Figure 4 shows the FT-IR spectra of bare SiO2, TiO2@SiO2 core-shell, 90 ℃-dried, 450 ℃-preannealed, and 900 ℃preannealed Ni-TiO2@SiO2 catalysts, and 90 ℃-dried, 450 ℃-preannealed Pd-TiO2@SiO2 catalysts. In the FT-IR spectra of bare SiO2, two major peaks were observed at 1070 and 800 cm−1 in the region of 650-1700 cm−1, which were attributed to Si–O–Si and Si-O bending modes, respectively.27 The two peaks were dominant for all samples, indicating that the coating thickness is thin enough to detect the IR absorption signal from the core SiO2. The weak broad peak at about 3200 cm−1 is attributed to adsorbed water and -OH. The broad peak at 3200 cm−1 was strong for 90 ℃-dried Ni- TiO2@SiO2 and Pd-TiO2@SiO2 samples. In addition, some extra peaks were observed due to vibrational modes of Pd or Ni complexes.28 Upon annealing at 450 ℃, the broad peak intensity was dramatically diminished, as expected. Upon annealing of the Ni-TiO2@SiO2 sample at 900 ℃, no peak was observed at around 3000 cm−1 due to absence of water and OH groups.

Figure 4.FT-IR spectra of bare SiO2, TiO2@SiO2 core-shell, 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni-TiO2@SiO2 catalysts, and 90 ℃-dried, 450 ℃-preannealed Pd-TiO2@SiO2 catalysts.

Figure 5.CO oxidation conversion (%) profiles for the first and second CO oxidation runs of 90 ℃-dried, 450 ℃ and 900 ℃preannealed Ni-TiO2@SiO2 catalysts. Inset optical microscope images show the color of the samples after the CO oxidation reaction.

We tested the CO oxidation performance of all of the aforementioned samples. Since the bare SiO2 and TiO2@SiO2 samples showed no CO oxidation activity between room temperature and 700 ℃ these samples are not discussed further. Figure 5 displays the CO oxidation conversion (%) profiles versus reaction temperature (℃) for the first and second CO oxidation runs of 90 ℃-dried, 450 ℃and 900 ℃-preannealed Ni-TiO2@SiO2 catalysts. The corresponding Arrhenius plots (ln rate versus 1/T) between the CO conversion of 10% and 15% are shown in Figure 6. The reaction rate (v) was calculated by the equation v = (molar CO flow rate, mol·s-1) × (CO conversion fraction)/(weight of the catalyst, gcat). In the first run of 90 ℃-dried Ni- TiO2@SiO2 catalyst (Figure 5), the T10% (the temperature at 10% CO2 conversion) was observed at 395 ℃, corresponding to an oxidation rate of 28 mmolCO s-1 gcat -1. In the second run of CO oxidation, the T10% decreased to 368 ℃, indicating that the activity was increased. For the 450 ºC-preannealed Ni-TiO2@SiO2 catalyst, the CO oxidation activity was greatly improved, and the T10% was observed at 368 ℃ and 335 ℃in the first and second runs of CO oxidation, respectively. However, for the Ni-TiO2@SiO2 catalysts preannealed at 900 ℃, the CO oxidation activity was degraded significantly, and the T10% was observed at 471 ℃and 402 ℃in the first and second runs, respectively. Each sample showed different CO oxidation activities and colors after the reaction.

Figure 6.Arrhenius plots of the first and second CO oxidation runs of 90 ℃-dried, 450 ℃ and 900 ºC-preannealed Ni-TiO2@SiO2 catalysts.

Table 1.aThe unit is ×10−4 molCO gcat −1 s−1 for the CO oxidation rate. bThe activation energy was measured in the CO conversion range of 10-15%.

Table 1 summarizes the calculated kinetic parameters based on the CO conversion profiles (Figure 5) and the Arrhenius plots (Figure 6). In the first run, the CO oxidation rates at 400 ℃ were measured to be 0.32 × 10−4, 0.54 × 10−4 and 0.012 × 10−4 mol gcat −1 s−1 for the 90 ℃-dried, 450 ℃ and 900 ℃-preannealed Ni-TiO2@SiO2 catalysts, respectively. The rates were increased to 0.51 × 10−4, 0.85 × 10−4 and 0.26 × 10−4 mol gcat −1 s−1 in the second run, respectively. The activation energies (Ea, kJ/mol) in the first run were measured to be 96.3, 81.9, and 54.9 kJ/mol, respectively, while these values were greatly reduced to 68.9, 75.7, and 85.4 kJ/ mol, respectively, in the second run.

It is worth comparing our results with the CO oxidation rate and the thermal treatment effects of pure NiO samples. Wang et al. synthesized NiO nanorings and found CO oxidation rates of 0.38−2 × 10−4 molCO gcat −1 s−1, which is in good agreement with our results.6 They also observed that samples preannealed at 550 ℃ showed the best performance, and that catalytic activity was found to decrease above the annealing temperature. This observation is in good agreement with our temperature dependent result.

Figure 7.CO oxidation conversion (%) profiles for the first and second CO oxidation runs of 90 ℃-dried and 450 ℃-preannealed Pd-TiO2@SiO2 catalysts. Inset optical microscope images show the color of the samples after CO oxidation reaction.

Figure 7 shows the CO oxidation conversion (%) profiles for the first and second runs of 90 ℃-dried and 450 ℃preannealed Pd-TiO2@SiO2 catalysts. The corresponding Arrhenius plots (ln rate versus 1/T) are displayed in Figure 8. In the first run of 90 ℃-dried Pd-TiO2@SiO2 catalyst (Figure 7), the T10% was observed at 495 ℃. In the second run, the T10% was significantly decreased (by 224 ℃) to 271 ℃. The significant decrease in temperature indicates that the catalyst became more active at a lower temperature. Based on the XRD and FT-IR data, Pd complex changes to PdO after high temperature thermal annealing, indicating that until the Pd complex changes to active PdO, CO oxidation may not occur. As a result, the onset CO oxidation temperature in the first run was much higher than in the second run. For the 450 ℃-preannealed Pd-TiO2@SiO2 catalyst, the CO oxidation activity was greatly improved, even in the first run. The T10% was observed at 263 ℃ and 247 ℃ and in the first and second runs of CO oxidation, respectively.

Figure 8.Arrhenius plots for the first and second CO oxidation runs of 90 ℃-dried, 450 ℃-preannealed Pd-TiO2@SiO2 catalysts.

Table 2.aThe unit is × 10−4 molCO gcat −1 s−1 for the CO oxidation rate. bThe activation energy was measured in the CO conversion range of 10-15%.

The calculated kinetic parameters are summarized in Table 2 for the Pd-TiO2@SiO2 catalysts. In the first run, the CO oxidation rates at 500 ℃ were 0.30 × 10−4 and 2.0 × 10−4 mol gcat −1 s−1 for the 90 ℃-dried and 450 ℃-preannealed Pd-TiO2@SiO2 catalysts, respectively. For the 90 ℃-dried sample, the rate was dramatically increased to 1.95 × 10−4 mol gcat −1 s−1 in the second run, possibly due to formation of PdO with higher catalytic activity.23 For the 450 ℃-preannealed sample, the rate was slightly increased in the second run when compared to the first run. Enhanced catalytic performance in the second run was also observed for other catalytic systems such as the metal-organic frameworksupported Pd nanoparticles.29 The activation energies (Ea) in the first run were measured to be 88.6 and 74.0 kJ/mol, respectively. In the second run, the values were measured to be 73.0 and 68.5 kJ/mol, respectively.

Based on the results for the two different catalyst systems, the Pd catalyst showed the much lower CO conversion temperature (or more energetically efficient catalytic performance) than the Ni catalyst. We further tested the stability29 of the two selected catalysts (Ni-TiO2@SiO2 and Pd- TiO2@SiO2 450 ℃). Figure 9 displays T10% (℃) up to fourth runs, and CO conversion with reaction time at a fixed temperature (at 550 ℃ for Ni-TiO2@SiO2, and at 450 ℃ for Pd-TiO2@SiO2 450 ℃ catalysts). The T10% was commonly decreased in the 2nd run, compared with the 1st run, as discussed above. The T10% in the 3rd and 4th runs was not greatly different from that of the 2nd run within 10 ℃. The stability of the two catalysts was tested by measuring CO conversion with reaction time at a selected high temperature. In Figure 9 (right), the Pd-TiO2@SiO2 450 ℃ catalyst showed no decrease CO conversion while the Ni-TiO2@SiO2, catalyst shows 8% decrease in conversion yield after 2 h. This indicates that Pd-TiO2@SiO2 catalyst shows much better stability than Ni-TiO2@SiO2 catalyst.

Figure 9.T10% (left) for the 1st, 2nd, 3rd and 4th CO oxidation runs, and CO conversion with time at a selected temperature (right) for Ni-TiO2@SiO2 and Pd-TiO2@SiO2 450 ℃ catalysts.

 

Conclusion

We tested CO oxidation performance of Ni-TiO2@SiO2 and Pd-TiO2@SiO2 core-shell nanostructures and obtained CO conversion (%) profiles and Arrhenius plots versus reaction temperature. The CO oxidation rate (mol gcat −1 s−1) occurred in the order of Ni-TiO2@SiO2 (900 ℃) < Ni- TiO2@SiO2 (90 ℃) < Ni-TiO2@SiO2 (450 ℃) in the first and second CO oxidation runs. For the Pd-TiO2@SiO2 catalysts, the oxidation rate occurred in the order of Pd- TiO2@SiO2 (90 ℃) < Pd-TiO2@SiO2 (450 ℃). The CO oxidation rate commonly increased in the second run when compared with the first run for the Ni-TiO2@SiO2 and Pd- TiO2@SiO2 core-shell catalysts. For the Ni-TiO2@SiO2 (450 ℃) catalyst, the T10% (corresponding to a reaction rate of 2.8 × 10−5 molCO gcat −1 s−1) occurred at 368 ℃ and 335 ℃ in the first and second runs of CO oxidation with activation energies of 81.9 and 75.7 kJ/mol, respectively. For Pd-TiO2@SiO2 (450 ℃) catalyst, the T10% occurred at much lower temperatures of 263 ℃ and 247 ℃ in the first and second runs with activation energies of 74.0 and 68.5 kJ/mol, respectively. The results of the present study confirm that the overlayer metal and thermal pretreatment are important factors for improving CO oxidation performance, and further highlight the importance of designing more efficient catalysts for environmental and chemical industrial applications.