Paper to the American Chemical Society Symposium Series 109 (Advances in Chemistry) 610 in 1972

S F Bush

Introduction

The experiments and mathematical analysis abstracted here were prompted by the observation, on the industrial scale, of a sudden instability in certain vapor-phase chlorine-hydrocarbon reactions. Reactions of this sort are often carried out in reactors where the velocities have been deduced from calculation and measurement on the laboratory scale. Under normal conditions the reaction of chlorine to completion occurs in the region below the injection point, but under some conditions the reaction ceases abruptly, and unreacted chlorine leaves this region to react eventually elsewhere in a dangerous and uncontrolled manner.

The steady reaction states observed cannot be regarded as unstable to infinitesimal disturbances since reaction is typically maintained for periods which are long compared with the time for such disturbances to pass through the system. The sudden cessation of reaction can be explained, however, as a result of a sufficiently large disturbance being applied to a system which is locally stable.

The stability of the systems studied is characterized as the time in seconds for which a given percentage reduction in the flow of the leaner feed component (usually chlorine) will just permit the system to recover. Experiments of this sort have been carried out in laboratory reactors with methyl chloride and methane as the hydrocarbon feedstocks.

Group II Technical Note, ICI Central Instrument Research Lab.

S F Bush with P A Sinclair

Summary

An analysis is presented of the mixing of chlorine and methyl chloride in the dip-pipes of chlorinators of Rocksavage type. Conclusions are drawn about the length of pipe needed for adequate mixing.

See also the section on Applications to existing products and processes.

Group II paper, ICI Central Instruments Lab (CIRL), for CIRL/Mond Division Symposium, Runcorn Heath, paper 5.

S F Bush

Summary

This paper contains a review of a number of aspects of gas-phase reactor design which have been encountered in the course of determining, and subsequently employing, the kinetics of methane and methyl chloride chlorination. Kinetic studies based on reaction in a paddle-stirred reactor are summarised together with results obtained from a mathematical model of the reactor. This model has led to the design of a jet-stirred reactor and preliminary results obtained from this device are also given. Finally, the problems surrounding the design of tubular chlorinators of the Rocksavage type are outlined and partial solutions described.

Preliminary report from Group II, ICI Central Instrument Research Lab.

S F Bush

Summary

The possible used of stirred reactors at Rocksavage depends on their performance relative to that of the existing plant and proposed extensions. The conclusions of this report are therefore based on estimates made by the author and by Mond Division of the performance of a train of chlorinators and on estimates made by the author of the performance of stirred reactor trains. These estimates are made with respect to conversion and capacity, and suggest a simple model of the chlorinators which will be refined and presented, together with more general information, in a later report. It must be emphasised that the estimates made here depend on somewhat scanty information about the plant performance and on kinetics derived from a small stainless steel stirred vessel. However, whenever possible the least favourable, but still realistic, data have been used in estimating. Thus it is thought, and has been assumed here, that the steel walls of the stirred reactor would yield a lower overall rate of reaction than the brick lined walls of the chlorinators, and that kinetics so based would thus yield a pessimistic estimate of plant capacity.

Group II Report, ICI Central Instrument Research Lab.

S F Bush

Summary

1. A stirred reactor system chlorinating methane would occupy a volume of about one-eighth that occupied by a system of uncooled static chlorinators, for the same output.
2. For the same conversion of methane to methylene chloride and chloroform, the production of methyl chloride is lower in a stirred reactor than in one of the existing chlorinators. The production of carbon tetrachloride is only slightly higher.
3. A stirred reactor controls the secondary C₂ and soot forming reactions better than the existing chlorinators can. In a chlorinator system with intermediate injection of chlorine the tendency to these unwanted reactions will be somewhat enchanced in the methane chlorination compared with the methyl chloride chlorination.
4. A stirred reactor, approximately one-sixth full-size is proposed, for erection at rocksavage, as a pilot plant for the chlorination of methane/methyl chloride mixtures. An outline design is appended.
5. The product distribution in the stirred reactor rules out its use in place of the present system for chlorinating methyl chloride alone. It is estimated, however, that by redesign of the dip pipe and selective use of water cooling, the chlorinators could be modified to carry the envisaged increase in chloroform production, avoiding the erection of an additional chlorinator per train.
6. In order to find the best design of static chlorinators for the chlorination of methyl chloride, a programme of experiments on a model chlorinator is proposed, the experiments to be carried out in about five months at CIRL.