Selective CO Oxidation in Excess of H PrOx over the Nanostructured Cux CexOy Catalyst

To lower the cost and improve the selectivity of the catalyst, a novel nonstoichio-metric nanostructured CuxCe1-xO2-y catalyst for the selective low-temperature oxi dation of CO in excess of H2 was synthesized by coprecipitation and sol-gel methods; these methods have been patented [18, 34]. The sol-gel method of catalyst preparation is particularly convenient for deposition on diverse geometries of support (i.e., honeycomb supports) and on reactors that can be used in PrOx processes. This type of catalyst is also capable of converting methanol directly into hydrogen and CO2 by steam reforming through a water-gas-shift reaction [45, 46]. By use of this catalyst, the four previously mentioned reactors (reformer, two-stage WGS reactors [LT and HT], and PrOx reactor) could be incorporated into a single unit.

Here we present a summary of recent results obtained in a study on the kinetics of the selective CO oxidation in excess of hydrogen over the nanostructured CuxCe1-xO2-y catalyst as obtained in the fixed-bed reactor operating in a differential mode. The inlet gaseous mixture composition simulates the real composition at the outlet of the low-temperature water-gas-shift reactor concerning the concentrations of CO, H2, and O2, except that no CO2, H2O, and unconverted CH3OH were present.

The experimental kinetic results obtained are presented in Figure 8.17 for three samples of CuxCe1-xO2-y catalysts with different content of copper prepared by coprecipitation. It was found that the reaction kinetics could be best represented by the redox mechanism [36]. Such a redox reaction can be described by the following two-step reaction:

The first step in this reaction mechanism is the catalyst reduction. Cat-O represents oxidized catalyst, which is attacked by a reductant (Red). The catalyst itself undergoes reduction, while the reductant is oxidized. The second step represents reoxidation of the catalyst by the oxidant (Ox-O), which donates an oxygen atom to the catalyst while it reduces itself.

The kinetics of selective CO oxidation over the CuxCe1-xO2-y nanostructured catalysts can be well described by employing a Mars and van Krevelen type of kinetic equation derived from a redox mechanism:

'm kCOkO2 PCOPO2

The parameters kCO and kO2 are taken to be the reaction rate constants for the reduction of surface by CO and reoxidation of it by O2. The parameters kCO, kO2, and n at one temperature were obtained by fitting experimental values of PCO, PO2, o u

F? - 100cc/m ??, W - 50m g -■-10% C?/CeO2 5% C?/CeO2 -:"~15% C?/CeO2

F? - 100cc/m ??, W - 50m g -■-10% C?/CeO2 5% C?/CeO2 -:"~15% C?/CeO2

FIGURE 8.17 CO conversion and selectivity over nanostructured CuxCe1-xO2-y catalysts with x = 0.05, 0.10, and 0.15, respectively.

FIGURE 8.17 CO conversion and selectivity over nanostructured CuxCe1-xO2-y catalysts with x = 0.05, 0.10, and 0.15, respectively.

and reaction rate with the rate equation. The parity plot for calculated vs. experimental values of reaction rate is presented in Figure 8.18. The agreement between experimental and calculated values is very good over three orders of magnitude of reaction rate.

We have performed the dynamic oxidation of carbon monoxide over Cua1Ce0.9O2_y nanostructured catalyst using a step change in CO concentration over the preoxidized catalyst. Figure 8.19 represents the CO and CO2 responses after a step change from He to 1 vol% CO/He over the fully oxidized Cu01Ce09O2-y nanostructured catalyst. At low temperatures, CO breakthrough is delayed for a few seconds, as can be seen from Figure 8.19a. At the temperature of 250°C, however, 20 sec is needed for the first traces of CO to exit the reactor. On the other hand, the evolution of CO2 in the reactor effluent stream has no delay, as seen in Figure 8.19b. However, the nature of the CO2 peak as a function of temperature changes significantly. At temperatures lower than 100°C, only one peak in CO2 response is visible.

FIGURE 8.18 Calculated vs. experimental values of reaction rates for the selective CO oxidation in excess hydrogen.

At 100°C, the CO2 peak broadens, and at 125°C, two separate peaks are clearly visible. The first peak is narrow, followed by a second broader peak. When the temperature is increased further, the first peak in the CO2 response becomes invisible because it is covered by the second peak. Only the origin of first peak is signified by fast evolution of CO2 in the reactor effluent stream. It is also important to notice how the maximum of the second peak shifts to the right when the temperature is increased. At 250°C, the catalyst surface responds almost instantly to a CO step change by producing CO2. The concentration of CO2 in the reactor effluent gas after 3 sec is 0.65 vol%, as shown in Figure 8.19b. However, the CO2 concentration in the reactor effluent gas rises further and reaches 0.80 vol% after 25 sec. This is followed by a sharp decrease in the CO2 concentration, which stabilizes after 100 sec at 0.2 vol%. Afterward, the concentration in CO2 decreases very slowly and falls to zero after 13 min.

A detailed elementary-step model of the CO oxidation over Cu01Ce09O2_y nano-structured catalyst under dynamic conditions was developed. The model discriminates between adsorption of carbon monoxide on catalyst-inert sites as well as on oxidized and reduced catalyst-active sites. Apart from that, the model also considered the diffusion of subsurface species in the catalyst and the reoxidation of reduced catalyst sites by subsurface lattice oxygen species. The model allows us to calculate activation energies of all elementary steps considered as well as the bulk diffusion coefficient of oxygen species in the Cu01Ce0.9O2-y nanostructured catalyst. The diffusion coefficient obtained by the mathematical modeling of step experiments is

50 100

FIGURE 8.19 a) CO and (b) CO2 concentrations in the reactor effluent stream as a function of temperature. The markers represent experimental points. The solid lines represent the model predictions obtained by the integration of the rate equations given in Table 8.11 with initial and boundary conditions given in Table 8.12. The kinetic parameter values are given in Table 8.13. Conditions: mcat = 200 mg, Ov = 200 ml min-1.

50 100

FIGURE 8.19 a) CO and (b) CO2 concentrations in the reactor effluent stream as a function of temperature. The markers represent experimental points. The solid lines represent the model predictions obtained by the integration of the rate equations given in Table 8.11 with initial and boundary conditions given in Table 8.12. The kinetic parameter values are given in Table 8.13. Conditions: mcat = 200 mg, Ov = 200 ml min-1.

shown to be within the range of bulk diffusion coefficients measured over other oxide catalysts. The elementary reaction steps, the mass balance equations, the initial and boundary conditions, and the estimated kinetic parameters are given in Table 8.10, Table 8.11, Table 8.12, and Table 8.13.

In our studies, we have demonstrated that the redox mechanism that was used to model dynamic behavior of CO oxidation is consistent with a kinetic model of the selective CO oxidation obtained under a steady-state mode of operation [37]. We propose the following tentative scheme (Figure 8.20) for the selective CO oxidation over the Cu01Ce0.9O2-y nanostructured catalyst: CO and H2 adsorb on the copper/ceria interfacial region of the catalyst, the most reactive places for both CO

TABLE 8.10

Elementary Reaction Steps Considered in the Kinetic Modeling of the CO Concentration-Step-Change Experiments for the Oxidation of CO over Completely Oxidized Cu01Ce0.9O2-Y Nanostructured Catalyst without the Presence of Oxygen in the Reactor Feed

Step No.

Elementary Reaction Step

CO ■■■ Cu2+Os,s ——— CO—Cu+q CO + Cu+q ——— CO-Cu+q Ce4+Ob,b ——- Ce3+q +Os,b CO ■ ■ ■ Cu+q +Osb ——— CO ■ Cu2+Os,b CO-Cu+q ——— CO2 + Cu+q

Note: Oxygen vacancy is represented by □. The meaning of subscripts accompanying oxygen species and oxygen vacancies is explained in text.

and H2 oxidation reactions. It is further proposed that CO (and H2) use mostly copper cations as the adsorption sites, while cerium oxide must also be present in the close vicinity. Copper oxide might also form a solid solution with cerium oxide, at least in the form of small intergrowths at the interface, which are x-ray diffraction (XRD )-invisible. In this concerted mechanism of copper and cerium oxide, copper cation has the following role: it is the adsorption site for the CO (and H2). When either of the two reactants is adsorbed on the copper cation, it extracts oxygen from the surface, and copper is reduced from Cu2+ to Cu+. Cerium cation, which lies next to copper cation, can supply an additional oxygen atom from the catalyst lattice while it reduces itself simultaneously from the Ce4+ into the Ce3+ form. Cerium oxide acts as an oxygen supplier (buffer) when it is needed at the place of reaction. Only one single copper ion is enough to convert one molecule of CO (or H2) into CO2 (or H2O), respectively. When the product molecule is desorbed, the site becomes available for the next reactant molecule, either CO or H2. Upon extraction of surface oxygen from the catalyst lattice, the oxygen vacancy can be refilled directly from the gas phase or by oxygen diffusion through the bulk of the catalyst. The latter mechanism is observed at higher temperatures.

The CuxCe1-xO2-y nanostructured catalyst prepared by the sol-gel method is a very efficient selective CO oxidation catalyst even under the highly reducing conditions that are present in a PrOx reactor. The catalyst is energy efficient toward the PEM fuel cell technology because it oxidizes CO with 100% selectivity close to the PEM fuel cell working temperature. These performances are obtained with a catalyst that contains cheap copper and cerium oxides rather than costly noble metals.

Was this article helpful?

0 0

Post a comment