One-Dimensional Oxygen-Deficient Metal Oxides

1

One-Dimensional Oxygen-Deficient Metal Oxides

Wei-Qiang Han†

†Center for Functional Nanomaterials, Brookhaven National Laboratory, Upton, NY 11973

Abstract Metal oxides are usually non-stoichiometric. Non-stoichiometry affects their physical properties and chemical reactivity and can lead to novel devices and new applications in renewable energies. This chapter introduces two types of oxygen-deficient metal oxides. The first is non-stoichiometric oxygen-deficient one-dimensional nano-ceria (CeO2-x) and non-stoichiometric oxygen-deficient TiO2-x nanowires with nanocavities; the second is sub-stoichiometric Magnéli phasesTinO2n-1 (4≤ n ≤ 10) nanowires and sub-stoichiometric Cr2O2.4 nanobelts with modulation structures. The application of 1D-nano-ceria for water-gas shifting reaction also is detailed.

11.1 Introduction

Metal oxides contain at least one metal cation and the oxygen anion (O2-). They usually are semiconductors and non-stoichiometric under typical experimental processing conditions. Oxides can be classified into two main categories, As listed in Table 11.1, viz. p-type and n-type, and into four subcategories, metal deficient, metal excess, oxygen deficient, and oxygen excess. [[1]] For example, cerium oxide (CeO2) falls into the oxygen deficient category. For this oxide, O2- vacancies are the cause of the metal excess; the oxide will have the real formula CeO2-x to keep the crystal neutral because two electrons are needed to be incorporated for each oxygen ion that is removed. A good site for these electrons is the Ce4+cation. If one electron is associated with one Ce4+cation, it will change from a Ce4+ ion to a Ce3+ ion. Interestingly, titanium oxide falls into both the metal-excess and oxygen-deficient categories.

Table 11.1 Semiconducting properties of binary metal oxides of non-stoichiometric composition 1

P-type semiconductor / n-type semiconductor
Deficit of metal / CoO, NiO, FeO, MnO, Cu2O / -
Excess of oxygen / UO2 / -
Excess of metal / - / TiO2, ZnO, CdO
Deficit of oxygen / - / TiO2, CeO2, ZrO2, SnO2, Nb2O5, Ta2O5, WO3, PuO2, Bi2O3, PbO2

Nanostructured objects have attracted wide attention in recent years because of their (1) new physics phenomena that affect physical properties; (2) unusual quantum effects and structural properties; and, (3) promising applications in optics, electronics, thermoelectric, magnetic storage, and renewable energies. One-dimensional (1D) systems are realized by creating nanostructures that are quantum confined in one direction. When the dimensionality of the material is lowered, the new variable of length scale becomes available to control of materials properties. Then, as the system’s size declines and approaches nanometer length-scales, it is possible to elicit dramatic differences in the density of electronic states, opening new opportunities to alter physical- and chemical-properties.[[2],[3]]1D nanostructures, including nanotubes (NTs) and nanowires (NWs), are used as elements for miniaturized electrical-, nanofluidic- and optical-nanodevices and have played important roles in renewable energies. [[4], [5]]

This chapter will introduce and detail the non-stoichiometric oxygen-deficient 1D-nano-ceria (CeO2-x) and its applications in water-gas shifting (WGS) reaction. It also will describe non-stoichiometric oxygen-deficient TiO2-x NWs with nanocavities, sub-stoichiometric Magnéli-phasesTinO2n-1 NWsand sub-stoichiometric Cr2O2.4 nanobelts (NBs) with modulation-structures.

11.2 Oxygen-Deficient 1D-Nano-CeO2-x and its Applications in the WGS Reaction

11. 2. 1. Crystal Structure of Cubic-Ceria

Cerium with a 4f25d06s2 electron configuration can exhibit both +3 and 4+ oxidation states, and intermediate oxides whose composition is in the range Ce2O3–CeO2 can be formed. The dioxide CeO2 crystallizes in the fluorite structure. It has a face-centered cubic cell (f. c. c) with space group Fm3m, (a=5.41134 Å, JCPDS 43-1002). In this structure, each cerium cation is coordinated by eight equivalent nearest-neighbor oxygen anions at the corner of a cube and each anion is coordinated tetrahedrally by four cations. The structure, illustrated in Figure 1, is considered as a cubic close-packed array of cerium ions with oxygen ions occupying all the tetrahedral holes.

Fig. 1The crystal structure of CeO2 in the fluorite structure

The cerium oxides, ranging from Ce2O3-CeO2, earlier were treated using the classic point-defect model of non-stoichiometry, in which oxygen-vacant sites were considered to occur randomly in the lattice, in conformity with the law of statistical thermodynamics. Later experiments indicated that non-stoichiometric phases originating from the fluorite lattice were formed at low-temperature by the removal of oxygen ions and ordering of the vacancies formed. [[6]]

Reduced ceria results from the removal of O2- ions from the CeO2 lattice, so generating a vacant anion site according to the following equation:

4Ce4+ + O2- → 4Ce4+ + 2e-/□ +0.5O2 → 2Ce4+ + 2Ce3+ + □ +0.5O2

where □ represents an empty position (anion-vacant site) originating from the removal of O2- from the lattice, here represented as an oxygen tetrahedral site (Ce4O). Electrostatic balance is maintained by the reduction of two cerium cations from +4 and 3+.

11. 2. 2 Background of the WGS Reaction

The WGS reaction, typically used to generate H2 through the reaction of a gas mixture of CO and H2O (CO + H2O → H2+CO2, Hº298 = -41.1 kJ/mol), is a well-known catalytic process of industrial importance. [[7]] Based on thermodynamic and kinetic considerations, a high conversion of CO is obtained with a two-bed operation at low (180-250 ºC) and high-temperatures (350-500 ºC). In continuously operating industrial applications, the classical catalysts employed are Fe2O3-Cr2O3 for the first stage (high-temperature shift (HTS)), and Cu/ZnO/Al2O3 for subsequent stages (low-temperature shift (LTS)), to obtain a good performance under steady state conditions.

For proton-exchange membrane (PEM) fuel cells, the anode catalyst usually is Pt/C, chosen as it is more sensitive to CO that those mentioned above, because the PEM fuel cell operates at lower temperatures at which CO can de-active the Pt. Usually, the CO in the fuel must be deeply reduced to < 10 ppm. The WGS reaction is a critical step in fuel processors for preliminary clean up of CO, and the additional generation of hydrogen before the preferential oxidation of CO, or the methanation step. WGS units sited downstream of the fuel reformer to further lower the CO content, and improve the H2 yield. To obtain this equilibrium outlet concentration of CO from the reformate fuel, the WGS catalyst must be active at low temperatures, 200–280°C, depending on the inlet concentrations of CO in it. The reaction is moderately exothermic, with low temperatures resulting in low CO levels; however, the kinetic of the reaction is favorable, even at higher temperatures. [[8]]

Employing Fe–Cr and Cu–ZnO catalysts in fuel processors poses a series of disadvantages: The low activity of Fe–Cr as HTS catalyst and its thermodynamic limitations at high temperatures; the sensitivity of the Cu-ZnO catalyst to air or temperature excursions; the lengthy pre-conditioning of such catalysts for intermittent operation (pre-reduction/passivation); and, the large reactor volume dictated by the slow WGS kinetics of the Cu–ZnO catalyst at low temperatures. Therefore, Fe-Cr and Cu-ZnO catalysts are unsuitable for automotive applications, where the need for fast start-ups dictate the need for using a low volume of a non-pyrophoric catalyst (fewer stages). Thus, it is critical to develop efficient safe catalysts for the WGS reaction in the fuel-cell process.

Ceria recently attracted great interest, particularly for reducing the emissions of CO, NOx and hydrocarbons from automobile exhaust, to abate soot formation in diesel fuels, and to minimize CO content in fuel-cells. [[9]] The key to these applications is that CeO2-x easily produces oxygen vacancies in an oxygen-deficient environment, shifting some Ce4+ to Ce3+ ions in the stable fluorite structure. [[10]] Oxygen vacancies also are crucial for binding catalytically active species to ceria. Thus, oxygen vacancies in ceria are considered to play an essential role in catalytic reactions.[[11]]Ceria-supported noble-metal co-catalysts, such as Pt-, Au-, Pd-loaded ceria, exhibit very interesting properties for the WGS reaction with fuel cells.[[12]]Under some conditions related to H2 production, the WGS reaction rates were higher on noble-metals loaded ceria than on commercial catalysts. [[13]]

This section discusses the use of pure- and Pd-loaded 1D-nano-ceria, a mixture of NTs and NWs, as catalysts for the WGS reaction at low-temperature.

11.2.3 Synthesis of 1D- Ceria

Various methods are used to prepare special ceria morphologies with enhanced reducibility. Zhou et al. generated ceria nanoparticles (NPs) by adding an aqueous ammonium hydroxide precipitant into a solution of cerium nitrate at room-temperature and then introduced oxygen into the reactor to oxidize Ce3+ to Ce4+. [[14]] Chen, et al. obtained ceria NWs via adding this same precipitant into cerium nitrate at 70 °C, and subsequently allowed it to age at 0 °C for one day. [[15]] Yu et al. prepared ceria nanocrystals (NCs) in spherical-, wire- and tadpole-shapes from a nonhydrolytic sol-gel reaction of cerium (III) nitrate and diphenyl in the presence of surfactants. [[16]] Natile et al. synthesized ceria NPs by two different synthetic routes: Precipitation from a basic solution (sizes around 8-15 nm) and microwave-assisted heating hydrolysis (size around (3.3-4.0 nm). [[17]] They found that the NPs made by the latter method were more reduced than those from the former. Methanol oxidation is also favored on the ceria NPs prepared by the latter method because of their high specific area and the presence of greater amount of active sites of Ce3+ cations. Zr4+ and La3+ doped porous ceria NPs with a high BET surface area of 160 m2/g exhibited a photovoltaic response, directly derived from the NPs’ size; normal ceria does not show this response. [[18]] All these results suggest that ceria with high surface area can increase the Ce3+ ratio that leads to high reducibility.

The 1D-nano-CeO2-x used for WGS reaction, described in the next section, was synthesized by two successive stages: Precipitation and aging. At the precipitation stage, 1.5 grams of cerium nitrate (Ce(NO3)3.6H2O) was added to 15 ml de-ionized water and heated at 100 °C. Once a large amount of vapor formed, 10 ml 5% ammonia hydroxide solution was added. Very fine yellowish precipitates formed immediately and the mix started boiling. After 3 minutes, the solution was transferred quickly to a 0 °C refrigerator. [[19]] Figure 2 displays the powder profile refinement of the as-produced material using GSAS/EXPGUI code.

Fig. 2Powder profile refinement of fresh 1D-ceria

Fig. 3Pure 1D-ceria sample (a) typical morphology with three kinds of nanostructures: NPs, NWs and NTs; (b) a high-magnification TEM image of a ceria NT, with a wall of about 5.5 nm thick; and, (c) a high-magnification TEM image of a ceria NR

A careful inspection reveals that there are two kinds of 1D nanostructures of CeO2-x. One is a NW with consistent cross-wise lattice; while the other is the NT with weak contrast in the middle (Fig 13.3(a)). These characteristics can be seen more clearly in high resolution images of a CeO2-x NT and a NW, respectively (Fig. 3(b-c)). The TEM image of the pure 1D-nano-CeO2-x shows polycrystalline ceria NWs and NTs (~80 %), together with ceria NPs (~20%) with a diameter similar to that of the 1D-nano-CeO2-x (Figure 3(a)). Most 1D-nano-cerias are NWs whose diameters typically range from 6-25 nms, and lengths up to a couple of microns. In Figure 3b, the selected area diffraction patterns (obtained by fast Fourier transform (FFT) techniques) in the upper right corner correspond to cubic ceria. The direction of the incident electron-beam is along <110>, i. e., the exposed crystal plane is (110). The axis of the CeO2-x NT is along the <110> direction. Two kinds of lattice-fringe directions attributed to (111) and (200) are observed that, respectively, have an interplanar spacing of 3.1 Å and 2.7 Å. For most NTs, the thickness of the wall is almost uniform over the tube, though thickness differs from tube to tube. Fig. 3c shows the CeO2-x NW has the same crystalline features as the NT. For the 1D-ceria, the preferred exposed crystal planes for both NWs and NTs are {110} and {100. Based on electron diffraction analyses and high resolution imaging, the CeO2-x NPs, NWs and NTs were shown to have the same crystal structure, a cubic fluorite structure, consistent with the x-ray measurements. The lattice parameters of the CeO2-x NTs vary from 0.54 nm to 0.56 nm depending on their diameters. In general, the lattice parameter increases with decreasing diameter of the NTs. Cerium-nitrate solution reacts with ammonium hydroxide to form Ce(OH)3 as an intermediate product with a 1D nanostructure that is retained if the pH of the reaction is higher than 8.[[20]] Excess ammonium hydroxide was used in the present experiment so the intermediate Ce3+ oxidized to Ce4+. Quickly cooling the samples to 0 °C retained the 1D nanostructure. The precipitates were dehydrated further, and re-crystallized during the aging time. Prolonging aging time leads to more 1D-like hollow structure, i.e. NTs.

Fig. 4 EELS spectra showing different M5 peak intensity for CeO2-x NTs with (a) d=14.6 nm, (b) d=17.3 nm, and, (c) 25.5 nm. The thicknesses of the wall of the NTs are 5.5, 6.0, and 10.8 nm for (a), (b) and (c), respectively. The spectra are normalized for the M4 peak

The increase in the lattice parameter of the CeO2-x NTs implies that the oxidation state of the CeO2-x NTs may differ from that of bulk CeO2. EELS (electron energy-loss spectrometry) can analyze the chemical composition of TEM specimens with a lateral resolution down to about one nanometer. The valence of the cerium ions is determined from the relative intensity of the white lines (M4 and M5) of the cerium in the EELS spectra.The NPs are almost completely reduced to CeO1.5 when the diameter < 3 nm. This reduced CeO2-x has a fluorite structure, the same as that of bulk CeO2. Also, EELS spectra taken from the edge and center of the NP indicated that for large NPs the valence reduction of cerium ions occurs mainly at the surface, forming a Ce1.5 layer and leaving the core essentially as CeO2.The fraction of Ce3+ ions in the NPs rapidly increased with declining NP’s size.[[21]] Fig.4 shows the M4 and M5 edges of the EELS spectra from three NTs with diameter, d = 14.6, 17.3 and 25.5 nm. It qualitatively illustrates the systematic change in the EELS spectra is correlated with the diameters of the NTs, that is the intensity of the M5 edge rises with the decrease in the diameter of the NTs. To determine the relative amounts of cerium ions Ce3+ and Ce4+, the second derivative method is used to measure the M5/M4 ratio, since it is insensitive to variations in thickness. The M5/M4 ratio these three NTs, d=14.6, 17.3 and 25.5 nm , respectively, are 1.27, 1.22 and 1.05; based on M5/M4 being 1.31 for Ce3+ and 0.91 for Ce4+, the fraction of Ce3+ (Ce3+/[Ce3++Ce4+]) therefore is estimated correspondingly as 0.90, 0.78 and 0.35. Compared with the CeO2-x NPs of the same diameter, the fraction of Ce3+ in the CeO2-x NTs is significantly larger. The main reason is that NTs have two surfaces: The outer surface and the inner one. Actually, the total surface area depends on the thickness of the wall of the NTs. If the cerium ions in the CeO2-x NTs follow the same distribution as that of the CeO2-x NPs, that is, Ce3+ exists on the surface, while Ce4+ inside, the fraction of Ce3+ mainly would be determined by the thickness of the wall. In fact, the thicknesses of the wall of the NTs for Fig. 4(a-c) are about 5.5, 6.0 and 10.8 nm, respectively. Oxygen vacancies in ceria NT combined with their inner and outer surfaces could offer more functional, effective features and play an essential role in applications, such as catalytic reactions. Techniques to make high-yield ceria NTs with sustainable stability during the WGS reaction still is challenging and worthy of further effort. [19]

11.2.4 Testing 1D-Ceria for the WGS Reaction

The in-situ time-resolved XRD experiments were performed at beam line X7B (λ = 0.922 Å) of the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory. The in-situ Ce LIII-edge x-ray absorption near-edge spectra (XANES) and the in-situ Pd K-edge data were collected there, at beam line X19A and X18B, respectively.[[22]] The products from time-resolved XRD and XAFS experiments were measured with a 0-100 amu quadruple mass spectrometer (QMS, Stanford Research Systems). The portion of the exit gas flow that passed through a leak valve and into the QMS vacuum chamber provided the relative pressure of the products. [[23]]

Fig. 5 Pure 1D-ceria sample (a) a 3D plot of in-situ time-resolved XRD patterns collected during the hydrogen reduction process. (b) H2 and CO2 relative pressure during the WGS reaction; (d) TEM image of the sample after the WGS reaction; and, (d) The lattice parameter of the ceria determined from the in situ diffraction during WGS and H2 reduction conditions as a function of temperature, which show relative cell expansion of H2 versus WGS

Samples of 1-2 mg were loaded into a 1-mm sapphire capillary tube attached to a flow system. The 1D-nano-ceria was exposed in pure H2 up to 400 ºC for activation before the WGS reaction. [[24], [25]] A similar set up to that used for the WGS reaction was employed for the temperature-programmed reduction and oxidation, for which pure H2 and 5% O2 in He were used, respectively. The temperature ramp rate was ~ 2 ºC/min. The in-situ time-resolved XRD patterns (Fig. 5a) showed the retention of the cubic-fluorite structure, and peak widths that were nearly constant during the reduction process, although there were significant changes in the lattice parameter from thermal expansion and the partial reduction of the cerium oxide, viz. from 5.43 Å at 25 oC to 5.47 Å at 400 oC. The reduction of pure 1D-nano-ceria in H2 started at 150 oC, a much lower temperature than those previously reported for bulk or 3D ceria NPs (i.e. NPs with no preferred growth in any direction).