Page 95 - The engineering of chemical reactions

Reaction-Rate Data 79

1. Real processes almost always involve multiple reactions. This is the subject of
Chapter 4, where we will see that selectivity to form a desired product is a crucial issue and
consequently there are many ks and orders that we need to know.
2. Real processes frequently involve catalysts or enzymes, which usually give rate
expressions much more complex than these.
3. Catalysts and enzymes also can vary significantly between batches and exhibit
activation and deactivation, so that reaction rates may be expected to vary with time. Thus
it is not unusual to find that a reaction activation energy increases with the time that a process
has been onstream, which one might need to fit by assuming that E is a function of time.
As you might expect, problems such as these require careful consideration and caution. We
will consider catalytic reactions and their kinetics in Chapter 7.
4. It is frequently difficult to maintain reactors strictly isothermal because all reactions
liberate or absorb considerable heat. These effects can be minimized by diluting reactants
and using low temperatures, thus making reaction rates sufficiently slow that the system can
be thermostatted accurately. However, kinetics under these conditions are not those desired
in a reactor, and one must be careful of the necessary extrapolation to operating conditions.
We will discuss heat effects in detail in Chapters 5 and 6.
5. Most reaction systems involve many simultaneous reactions, and the analysis of
these systems can be very difficult or impossible. For a large set of reactions with unknown
kinetics, it is common simply to assume reactions and rates (for example, with all reactions
first order) and make empiricalJits of experimental data to these rate expressions. Such rate
expressions have little fundamental significance, but empirical reaction rates allow one to
formulate models of the reactor behavior. Special caution must be used in extrapolating
these empirical rate expressions beyond the range of data from which they were obtained.
Many industrial reaction processes such as in the petroleum refinery are analyzed through
such empirical models. These are developed over long periods of time to fit plant behavior,
and each large producer of a given process has carefully guarded empirical models with
which to optimize plant performance for given feedstocks and customer product demands.
6. Kinetic data are frequently acquired in continuous reactors rather than batch
reactors. These data permit one to determine whether a process has come to steady state
and to examine activation and deactivation processes. These data are analyzed in a similar
fashion to that discussed previously for the batch reactor, but now the process variables
such as reactant flow rate (mean reactor residence time) are varied, and the composition
will not be a function of time after the reactor has come to steady state. Steady-state reactors
can be used to obtain rates in a differential mode by maintaining conversions small. In this
configuration it is particularly straightforward to vary parameters individually to find rates.
r One must of course wait until the reactor has come to steady state after any changes in feed
br or process conditions.

Small steady-state reactors are frequently the next stage of scaleup of a process
from batch scale to full commercial scale. Consequently, it is common to follow batch
experiments in the laboratory with a laboratory-scale continuous-reactor process. This
permits one both to improve on batch kinetic data and simultaneously to examine more
rile I properties of the reaction system that are involved in scaling it up to commercial size.
ese ! Continuous processes almost by definition use much more reactants because they run
continuously. One quickly goes from small bottles of reactants to barrels in switching to