Porous Catalyst Particle Problem

The starting point of a number of theoretical studies of packed catalytic reactors, where an exothermic reaction is carried out, is an analysis of heat and mass transfer in a single porous catalyst since such system is obviously more conductive to reasonable, analytical or numerical treatment. As can be expected the mutual interaction of transport effects and chemical kinetics may give rise to multiple steady states and oscillatory behavior as well. Research on multiplicity in catalysis has been strongly influenced by the classic paper by Weisz and Hicks (5) predicting occurrence of multiple steady states caused by intrapellet heat and mass intrusions alone. The literature abounds with theoretical analysis of various aspects of this phenomenon however, there is a dearth of reported experiments in this area. Later the possiblity of oscillatory activity has been reported (6). 

The significance of the key parameters which govern concentration and temperature distribution in a single porous catalyst may be illustrated most effectively by presentation and manipulation of the differential equations which are supposed to describe the system. 

1 For those who are interested in recent developments of the theory the excellent book by Aris (/) and papers by Aris (2), Schmitz (J), and Ray (4) are recommended. 

The present article is not designed to review the work devoted to theoretical treatment of multiple steady states and oscillatory activity predicted from these equations, for those readers who seek a profound review the texts by Aris (7), Schmitz (5), and others (7,9) are recommended. 

Below we outline the quantitative results that have been gained by the theoreticians without recourse to rigorous analytical treatment of governing equations. The conditions under which a porous catalyst particle may exhibit multiple steady states have been conceived by Aris (2), Luss (5), and others (9)  

---

Studies with porous catalyst particles conducted during the late 1930s established that, for very rapid reactions, the activity of a catalyst per unit volume declined with increasing particle size. Mathematical analysis of this problem revealed the cause to be insufficient intraparticle mass transfer. The engineering implications of the interaction between diffusional mass transport and reaction rate were pointed out concurrently by Damkohler [4], Zeldovich [5], and Thiele [6]. Thiele, in particular, demonstrated that the fractional reduction in catalyst particle activity due to intraparticle mass transfer, r, is a function of a dimensionless parameter, 0, now known as the Thiele parameter. 

In heterogeneous catalytic reactions, the problem is even more subtle what exactly is the reaction volume If the catalyst has a porous structure accessible to the reactants, is it the bulk volume of the porous catalyst particles Or perhaps the volume of the pores themselves What if the catalyst is non-porous and reaction takes place only on the external surfaces of the non-porous particles Should we then use the void volume or the catalyst volume as the reaction volume Or perhaps all catalytic reaction rates should be expressed per unit of time per unit of available surface area. 

This review paper is concentrated on problems in scaling-up multiphase catalytic fixed bed reactors such as trickle-bed or packed bubble column reactors, in which two fluid phases (gas and liquid) pass concurrently through a bed of solid (usually porous) catalyst particles. These types of reactors are widely used in chemical and petrochemical industry as well as in biotechnology and waste water treatment. Typical processes are the hydrodesulphurization of petroleum fractions, the butinediol syntheses in the Reppe process for synthetic rubber, the anthrachinon/hydrochinon process for H202 production, biochemical processes with fixed enzymes or the oxidative treatment of waste water under pressure. 

As long as product is deposited within the micropores of the catalyst by capillary condensation only, there should be no problem, as the particle will behave as a dry one. Incipient wetness corresponds to a situation where hydrocarbon product starts to condense on the outer surface of the porous catalyst particle. This situation, which is characterized by the hydrocarbon dew point, marks the onset of particle agglomeration and defluidization. 

In addition to these kinetic steps, there are also physical processes of heat and mass transfer to be considered. The external transport problem is one of heat and species exchange through the boundary layer between the surrounding bulk fluid and the catalyst surface (Figure 5). Concentration and temperature gradients are necessarily present in this case and would have to be accounted for in the modeling equations. Also, there is often an internal transport problem of heat conduction through the catalytic material -- and in the case of porous catalyst particles, an internal diffusion problem as well. Internal transport problems are beyond the scope of this paper. It must be noted, however, that any model intended to describe real-life systems will have to account for these effects. 

The GDE for hydrochloric acid electrolysis is characterised by micro-scale hydraulic problems connected with the competition between the gas phase (oxygen), which has to diffuse towards the catalyst, and the liquid phase (water), which must be released. This competition is managed basically by a flow-through structure provided with hydrophobic channels of relatively large diameter. These are formed from PTFE (the binder of the structure) and catalyst particles and account for regulating the gas phase. Hydrophilic channels with smaller diameters (one order of magnitude smaller), which are located in the micro-porous carbon particles of the catalyst support (e.g. Vulcan XC-72), act as water absorbers. A consequence of the electrolysis process is that the catalyst itself is partially covered by liquid. This reduces its effectiveness and accounts for extra voltage. 

It helps to distribute the reactant gases or liquids evenly from fhe FF channels of the bipolar plates to the CL so that most of fhe active zones (and catalyst particles) are used effectively. Thus, the DL has to be porous enough for all the gases or liquids (e.g., liquid fuel cells) to flow without major problems. 

Several length-scales have to be considered in a number of applications. For example, in a typical monolith reactor used as automobile exhaust catalytic converter the reactor length and diameter are on the order of decimeters, the monolith channel dimension is on the order of 1 mm, the thickness of the catalytic washcoat layer is on the order of tens of micrometers, the dimension of the pores in the washcoat is on the order of 1 pm, the diameter of active noble metal catalyst particles can be on the order of nanometers, and the reacting molecules are on the order of angstroms cf. Fig. 1. The modeling of such reactors is a typical multiscale problem (Hoebink and Marin, 1998). Electron microscopy accompanied by other techniques can provide information on particle size, shape, and chemical composition. Local composition and particle size of dispersed nanoparticles in the porous structure of the catalyst affect catalytic activity and selectivity (Bell, 2003). 

Considerations based on the known physical phenomena can guide the choice of catalyst porosity and porous structure, catalyst size and shape and reactor type and size. These considerations apply both to the laboratory as well as to large-scale operations. Many comprehensive reviews and good books on the problem of reactor design are available in the literature. The basic theory for porous catalysts is summarized in this book and simple rules are set forth to aid in making optimum choices to obtain fully effective catalyst particles, which give the best performance from an economic point of view. 

First, let us examine the criteria applicable to diffusion effects in the gas phase, i.e., the spaces and channels over or between catalyst particles. When the catalyst solids are not porous but have all their active surfaces located in their geometric contours, diffusion in the outside gas space will be the only existing diffusion problem. However, even when the catalyst particles are subject to internal diffusion effects, the external gas space conditions need still be examined separately. The criteria will be examined assuming the reaction to be of first order, keeping in mind that deviation from exact first-order kinetics does not alter the diffusion picture by considerable magnitudes, as was seen above. 

Chapter 5 is dedicated to the single particle problem, the main building block of the overall reactor model. Both porous and non-porous catalyst pellets are considered. The modelling of diffusion and chemical reaction in porous catalyst pellets is treated using two degrees of model sophistication, namely the approximate Fickian type description of the diffusion process and the more rigorous formulation based on the Stefan-Maxwell equations for diffusion in multicomponent systems. 

When a catalyst is used in a dissolved or suspended form, separation of the catalyst from the mixture may offer a special problem. Suspended fine catalyst particles may be kept in the reactor by using porous filter elements in the reactor exit. Periodically the exit flow is reversed to clean the filters. Alternatively the separation may take place outside the reactor, and the catalyst has to be recycled. Dissolved catalysts such as organic metal salts will have to be precipitated and filtered. Extraction or adsorption may also be feasible. In all cases traces of a dissolved catalyst may still be present in the process stream, which requires additional purification steps. Eventually traces of the catalyst may contaminate the main product or the waste streams. Especially the removal of traces of metal ions from waste water streams can require considerable effort. To this end, ion exchange resins or complexing resins are often used. These have to be regenerated from time to time this operation creates again waste water. Obviously... 

The problems of monomer recovery, reaction medium viscosity, and control of reaction heat are effectively dealt with by the process design of Montedison Fibre (53). This process produces polymer of exceptionally high density, so although the polymer is stiU swollen with monomer, the medium viscosity remains low because the amount of monomer absorbed in the porous areas of the polymer particles is greatly reduced. The process is carried out in a CSTR with a residence time, such that the product k jd x. Q is greater than or equal to 1. is the initiator decomposition rate constant. This condition controls the autocatalytic nature of the reaction because the catalyst and residence time combination assures that the catalyst is almost totally expended in the reactor. 

---