Harvey G. Stenger

Oxidation of sulfur dioxide over supported vanadia catalysts: molecular structure – reactivity relationships and reaction kinetics

The oxidation of sulfur dioxide to sulfur trioxide over supported vanadium oxide catalysts occurs as both a primary and secondary reaction in many industrial processes, e.g., the manufacture of sulfuric acid, the selective catalytic reduction of NOx with ammonia and the regeneration of petroleum refinery cracking catalysts. This paper discusses the fundamental information currently available concerning the molecular structure and sulfur dioxide oxidation reactivity of surface vanadia species on oxide supports. Comparison of the molecular structure and reactivity information provides new fundamental insights on the following topics related to the catalytic properties of surface vanadia species during the sulfur dioxide oxidation reaction: 1. role of terminal VO, bridging V–O–V and bridging V–O-support bonds, 2. number of surface vanadia sites required to perform SO2 oxidation, 3. influence of metal oxide additives, 4. generation and influence of the surface sulfate overlayer, 5. effect of surface acidity on the reaction turnover frequency, 6. competitive adsorption between SO2 and SO3 and 7. reaction kinetics.

The identification of kinetic expressions and the evolutionary optimization of specialty chemical batch reactors using tendency models

Abstract This paper presents a method for the identification of kinetic models for a wide variety of batch/semibatch processes related to the production of fine and specialty chemicals. The developed methodology is particulary suitable for those industrial processes where a detailed understanding of process fundamentals (reaction mechanisms, process kinetics) is not available. It is also readily applicable to those processes where it is difficult to obtain time dependent concentration data. The modeling effort is aimed at developing a low-order, nonlinear “Tendency Model” which is descriptive of the qualitative and approximate quantitative behavior of the overall process. the physical/chemical insight gained from a set of two-level multifactor experiments in primary operating variables (e.g. temperature, catalyst concentration, initial concentrations of key reactants, promotors, etc) is used to select an approximate functional form (viz. power-law kinetics, Langmiur-Hinshelwood, etc.) for various rate equations in the kinetic model. A priori process knowledge and process understanding gained from statistical analysis of the data collected is used to aid in the model identification. The final values of reaction orders and the remaining kinetic parameters are determined by minimizing a prespecified model-fitting function. A factorial analysis of the models prediction for process responses is used to determine the accuracy of the model. An inaccurate model is modified to account for any observed discrepencies. The developed “Tendency Model” is used for the optimization of the process in an evolutionary manner. As the process operates and additional data becomes available, the model parameters are updated using the information on the sensitivity of the plant-model mismatch with respect to various reaction orders. The developed methodology is illustrated by means of experimental data for an example process related to the production of fatty acid epoxides which are used as stabilizer/plasticizers for PVC resins.

Analysis of Alloy Nanoparticles

Abstract. Quantitative analysis of particles less than 10 nm in diameter requires a focused electron beam to isolate individual particles for X-ray emission spectrometry. Effects such as phase separation among particles and surface segregation within particles can only be determined by this technique. This analysis can be made quantitative with minimal use of the usual correction factors provided the small particles are supported on ceramic materials about the same thickness as the particles themselves. Two alloy nanoparticle systems are examined here: Pt-Rh and Pt-Re. In each case the catalytic properties resulting from various processing procedures have been correlated with the microstructure within and among individual particles.

Effects of drying on the preparation of HF co-impregnated rhodium/Al2O3 catalysts

Abstract The effects of drying rate on the internal distribution of rhodium within HF co-impregnated rhodium catalysts is reported in this work. Quantitative electron probe microanalysis (EPMA) revealed catalysts with uniform rhodium distributions, inner bands (egg-whites) of Rh and inner cores (egg-yolks) of Rh. The type of internal rhodium distribution produced was found to depend strongly upon the rate of catalyst drying. Measured drying curves showed two governing regimes of drying. At temperatures below 500°C evaporation appeared to take place primarily at the external surface of the support. Also for drying temperatures less than 500°C, EPMA distributions displayed a significant degree of rhodium segregation to the external surface of the support. This segregation occurred when characteristic times for solute back diffusion were greater than the time to reach the critical moisture content. At a temperature of 500°C, evaporation appeared to occur solely from a receding front within the support (i.e. shrinking core drying). Regardless of the regime in which drying takes place, it was shown that for the egg-white catalyst character, which exists immediately after impregnation to be preserved, the total drying time must be less than the time required for diffusion to the center of the support.

Distributions of HF Co-impregnated rhodium, platinum and palladium in alumina honeycomb supports

Abstract The co-impregnation of Rh, Pt, and Pd with hydrofluoric acid (HF) into 1.2 mm thick γ-alumina monolith walls has been investigated. Quantitative electron probe microanalysis (EPMA) results are presented which reveal the effect of co-impregnation with HF on the resulting internal catalyst distribution. Co-impregnation of RhCl 3 (H 2 O) x with HF produced samples with inner shells (egg-whites) of Rh. Co-impregnation of PtCl 4 with HF produced samples with both inner shells (egg-whites) and inner cores (egg-yolks), of platinum. The co-impregnation of Pd(NO 3 ) 2 (H 2 O) x with HF yielded catalysts with uniform outer shells, (egg-shells), of palladium. Also, two approaches for the correction of atomic number, absorption, and fluorescence effects in the X-ray emission spectroscopy (XES) data taken in EPMA were compared for these catalyst systems. The atomic number correction, Z , was found to be the dominant term and to depend significantly upon its method of calculation. Total metal loadings obtained by integrating quantitative EPMA profiles were in close agreement with bulk chemical analyses.

Use of tendency models and their uncertainty in the design of state estimators for batch reactors

Tendency models have been successful in the modeling and optimization of batch reactor processes where a detailed understanding based on fundamental principles and detailed kinetic studies is not available. The evolutionary nature of the Tendency modeling algorithm has proven useful in updating the process model between batches, as new process data or insight become available. But optimization is not the only task that can be undertaken with a Tendency model. In this work, the use of Tendency models in the design of state estimators to estimate reactor concentrations is investigated. The primary goal is to use the knowledge of the uncertainty in the Tendency model (which, by its nature, is an approximate model) to tune an extended Kalman filter. Two examples are presented to illustrate that even though Tendency models can feature a significant amount of uncertainty, they can be used successfully in state estimators.

Effect of process-model mismatch on the optimization of the catalytic epoxidation of oleic acid using Tendency models

Abstract The optimization of batch chemical reactors is strongly affected by the accuracy of the Tendency model used. In this work, the effect of this process-model mismatch on the optimization of the epoxidation of oleic acid using Tendency modeling is studied. The primary objective is to quantify the effect of the models uncertainty on process optimization so as to select the less uncertain between two Tendency models resulting for the experimental data. By placing confidence limits on the optimal operating policy and the expected performance index predicted by each model, it is possible to predict the poor performance that results when using a power-law kinetic model to optimize this process. The improved performance that results from using an alternate kinetic model, featuring inhibition kinetics, can also be predicted through this type of uncertainty analysis.

Effect of Model Uncertainty on the Tendency Modeling, Optimization and Control of Batch Reactors

Abstract The optimization of chemical batch reactor processes is strongly affected by the accuracy of the Tendency model used. In this work, the effect of this process-model mismatch on the convergence of the Tendency modeling, optimization and control algorithm is studied. The primary objective is to determine whether an uncertain Tendency model can still be successfully used for process optimization. In the two examples examined, if the magnitude of the uncertainty is limited, it is shown that by operating the next batch under suboptimal but less uncertain conditions, the optimization of the process can be achieved by Tendency models that have some structural mismatch with the process.