Richard K. Herz

Surface chemistry models of carbon monoxide oxidation on supported platinum catalysts

Rates of CO oxidation on noble metal catalysts are often represented by a Langmuir-Hinshelwood rate expression that assumes the competitive equilibrium adsorption of CO and O2 on the active metal surface. Surface chemistry studies reported in the literature have shown, however, that this assumption cannot be justified under all conditions. As a result, the usual Langmuir-Hinshelwood rate expression is unable to explain the data of several reported studies of CO oxidation on Pt catalysts. In this report we develop two reaction models which include separate adsorption, desorption, and surface reaction steps, and which do not assume adsorption equilibrium. The ability of each model to fit CO oxidation rate data taken with an alumina-supported Pt catalyst is compared with that of the usual Langmuir-Hinshelwood rate expression. Unlike the usual rate expression, the surface chemistry models successfully simulate the abrupt transition in steady-state rate that occurs between the CO inhibition regime and the first-order regime. The parameter values used to fit the supported Pt data are similar to those determined with Pt crystals. However, they indicate that CO may be adsorbed less strongly on the supported Pt and that most of the surface Pt atoms in the supported catalyst were deactivated by an oxidizing pretreatment.

Dynamic behavior of automotive catalysts: III. Transient enhancement of water-gas shift over rhodium

The participation of the water-gas shift reaction in the transient response of Rh-containing catalysts in exhaust was demonstrated by comparing the responses of a Pt/Rh/Ce/Al2O3 catalyst, a Rh-free Pt/Ce/Al2O3catalyst, and a Ce-free Pt/Rh/Al2O3 catalyst. Each of the bead-type catalysts was mounted in a converter in the exhaust system of a gasoline engine. Infrared diode laser spectroscopy was used to measure CO concentrations in exhaust simultaneously at the converter inlet and outlet with high time-resolution. In the first series of experiments, a catalyst was stabilized under rich exhaust and then given a 1-s exposure to lean exhaust. Following the subsequent lean-to-rich transition, a transient reduction in CO emission was observed over all three catalysts. The reduction in CO emission over the PtCe catalyst could be explained by reaction of CO with oxygen “stored” in the catalyst and CO adsorption and accumulation in the catalyst. The reductions in CO emissions over the two Rh-containing catalysts were greater than the amounts that could be explained by CO and oxygen accumulation and reaction but were consistent with a transient enhancement of the water-gas shift reaction over Rh. Lean-to-rich step response experiments provided further evidence of the participation of water-gas shift in the dynamic behavior of three-way automotive catalysts. During air-fuel ratio cycling and the warmed-up portion of the Federal Test Procedure, time-averaged CO conversions decreased in the order Pt/Rh/Ce>PtCe>PtRh. This ranking suggests that oxygen accumulation and reaction over Ce contributes somewhat more to CO conversion during driving than water-gas shift over Rh.

Adsorption effects during temperature-programmed desorption of carbon monoxide from supported platinum

Experimental data are presented for the temperature-programmed desorption (TPD) of CO from porous PtAl2O3 into a vacuum. Most of the preadsorbed CO desorbs in a peak between 380 and 550 K. Other workers have measured desorption at substantially higher temperatures during TPD of CO into a carrier gas rather than a vacuum. A comparison of the experimental conditions suggests that the competition of CO adsorption with CO desorption may contribute to the differences between the TPD results. To investigate the effects of CO adsorption, we develop a mathematical model and use it to compute desorption spectra for the TPD of CO from Pt dispersed over a porous support into (a) an inert carrier gas and (b) a vacuum. Over the realistic parameter range considered, our model predicts that adsorption effects, caused by high concentrations of gaseous CO in the system, are always an important feature, broadening the desorption peaks and shifting them to higher temperatures. Indeed, we find that adsorption competes with desorption to the extent that adsorption equilibrium is always approached closely within the porous supported Pt samples. For desorption into a carrier gas, the adsorption effects result from limitations to the flow of CO from the sample cell, whereas for desorption into a vacuum, the adsorption effects result from limitations to the diffusion of CO from the porous sample. Our results suggest that significant adsorption effects will also be present during the TPD of CO from other Group VIII precious metals dispersed over porous supports.

Transient oxidation and reduction of alumina-supported platinum

Abstract The oxidation state of alumina-supported Pt following treatment with O2 and the transient reaction between CO and preadsorbed oxygen are reported. Pt/Al2O3 (65%-dispersed) was exposed to O2 (1.23 kPa) at 523 K. The maximum oxygen uptake was obtained within 120 s. Following flushing with inert, a flow of CO (1.23 kPa) was switched over the sample. During the transient reaction, the production of gaseous CO2 was monitored with a mass spectrometer and the absorption of infrared radiation by adsorbed CO was recorded at a fixed wavenumber. The experiments were repeated to obtain the infrared data over a range of wavenumbers. From the complete set of infrared data, spectra were obtained which show how the population of five different CO adsorption states varied with time. The total amount of CO2 produced during the transient reaction corresponds to removal of two O atoms per Pt atom in the sample and, thus, indicates complete oxidation of the Pt. Two peaks in the CO2 production rate are observed: an initial sharp peak lasting for 2.5 s in which about 20% of the preadsorbed oxygen is removed followed by a broad peak lasting for about 120 s in which all of the remaining oxygen is removed. With 30%-dispersed Pt/Al2O3 the relative heights and shapes of the two peaks are slightly different but the same initial O atom to total Pt atom ration is indicated. The first CO2 peak is ascribed to removal of oxygen atoms in the surfaces of supported Pt oxide particles. The second CO2 peak is ascribed to removal of oxygen atoms bound within supported Pt oxide particles. Similar transient reduction behavior was also observed with Rh/Al2O3.

Identification of a weak adsorption state of carbon monoxide on highly dispersed, alumina-supported platinum

Abstract Temperature-programmed desorption of CO from 0.1 wt% Pt Al 2 O 3 was measured as a thin wafer of the material was heated in vacuum. At a Pt dispersion of 100% ( CO Pt = 0.92 ), a relatively tall, narrow CO desorption peak was obtained at low temperatures (355 K at 0.93 K s −1 ) followed by a low, broad desorption band extending to 750 K. As the Pt dispersion decreased to 80% ( CO Pt = 0.70 ) and then 70% ( CO Pt = 0.64 ) as a result of pretreatment in O 2 and H 2 , the fraction of CO that desorbed in the low-temperature peak decreased and the fraction of CO that desorbed in the broad desorption band increased. At Pt dispersions below 40%, only a broad desorption band was obtained, a result which is in agreement with data reported by others. The low-temperature peak is interpreted as desorption of CO from very small Pt particles ( 2 O 3 . The broad desorption band is interpreted as CO desorption from larger Pt particles. Results from experiments performed at different heating rates indicate that the enthalpy of CO adsorption in the state that corresponds to the low-temperature peak is 74 ± 4 kJ mol −1 , an adsorption energy which is lower than energies obtained from measurements of CO desorption from Pt single crystal surfaces.

Dynamic Measurement of Carbon Monoxide Concentrations in Automotive Exhaust Using Infrared Diode Laser Spectroscopy

Infrared diode laser spectroscopy was used to make the first measurements of carbon monoxide (CO) concentration in automotive exhaust with a time response fast enough to be useful in the analysis of engine and emission control system dynamics. Carbon monoxide concentrations were measured before and after a three-way catalytic converter. The response time of the laser system (for a change in amplitude of 10%-90% of fullscale) was 25 ms, more than adequate to resolve all CO transients of interest. The instrument was capable of resolving concentration changes on the order of 0.1 vol % before the catalyst and 0.02 vol % after it. A minicomputer was used to simultaneously collect data on CO concentrations before and after the catalyst, and also the output of an oxygen sensor located in the exhaust line.