Autoclaves or pressure vessels are always used in cases where it is necessary to raise the temperature of a substance above its boiling point, or where gases are evolved on heating, which are necessary for the reaction. In this connection the most important substances dealt with in colour technology are aqueous solutions and mixtures, and after that alcohol and alkyl chlorides. The limits of pressure and temperature are about 60 atmos. and 300o C. Generally speaking the ordinary type of apparatus is not calculated to withstand greater stresses than those indicated.

Both vertical and horizontal autoclaves are made use of, either with or without stirring-gear. If the mixture is homogeneous, stirring is unnecessary; but if several layers are formed, or it is necessary to bring together solid and liquid substances, then continuous stirring is essential. As an example of a reaction carried out under pressure in which stirring is unnecessary, the ordinary preparation of dimethylaniline may be given; whilst all alkali-fusions afford examples where continuous stirring is requisite.

The autoclaves used are hollow, cylindrical vessels of 100-10,000 litres capacity, provided with a flange to which the cover is affixed by means of nuts and bolts; for the sake of greater strength, the bottom is usually made hemispherical. They are almost invariably made of iron, either cast iron or cast steel being used, the latter offering the greater measure of safety for working at high pressures

(see also steel and iron as structural materials). Besides cast iron and steel, tin is also used; autoclaves made of tin are manufactured, having walls up to 40 mms. thick; they are either riveted or auto-genously welded. The objection taken to welded autoclaves by many colour factories is quite unjustified, and is explained simply by the fact that at first the welding process had not been fully perfected.

The screw-bolts form the weak point in every autoclave, and they must therefore be made from the best hand-forged wrought iron. The cover has flange-pieces cast on, as indicated in Fig. 32. These serve for affixing the armatures and the stuffing-box of the agitator. The bracket which bears the agitator should be high enough to make it easy for the packing of the stuffing-box to be looked after and renewed. The stuffing-box itself, through which the axle of the stirrer goes, should be as simple as possible, and cooled with water. Hollow stuffing-boxes are also cast, through which water may be circulated. Cooling by means of circulating oil, such as is done with success in the case of steam turbines is here unnecessary, as the conditions are quite different. The cover is provided with two manometers and two thermometers, together with two safety valves. Of recent years it has become the custom in certain cases, with the approval of experts, to do away with the safety valves, as these never work properly and are a constant source of annoyance. By using two thermometers and manometers it is easily possible to follow the course of a reaction exactly. With large autoclaves the cover is provided with a special opening or man-hole, this alone being opened from time to time as occasion requires. The joint between the cover and the body of the vessel is made tight by means of a special packing ring let into the flange of the autoclave itself, suitable materials for the ring being copper, lead, lead-covered iron, or asbestos board. The rings must be accurately turned, and must have a width of 20-50 mms. and a thickness of 1-6 mms. Lead is somewhat easily squeezed out by the pressure of the screws, but withstands ammonia very well; copper is the ideal packing, but is attacked by ammonia. Asbestos can be used for low pressures, but has the disadvantage that it nearly always tears when the autoclave is opened. The cover is screwed on by first tightening lightly diametrically opposite bolts, and then working round in a circle, always screwing up tighter and tighter, the final tightening being effected by hammering the long wrenches employed for the purpose; there is no danger of the bolts being broken off in the process.

The walls of an autoclave should never be brought into contact directly with the substances under reaction, as every melt attacks the walls, so that after a time the apparatus becomes too weak and must be taken down. For this reason a lining is nearly always inserted in the actual pressure-vessel, and is kept in position by pouring in solder. It is not permissible simply to place the liner in the autoclave as the heat transmission is inadequate and the walls of the autoclave may become red-hot. When placing the liner in position the latter is fixed immovably in the autoclave by means of a strong girder, and then the solder is poured in through a sheet-iron funnel. Enamel and lead may be protected by covering the inside of the liner with wet cloths, or it may be filled with water which must, however, be heated up and afterwards cooled by means of a worm. If the heating be omitted it is quite possible for the solder never to reach the bottom, but to solidify half-way down; if, on the other hand, the inner vessel be cooled by water without using a coil, then the water will boil away altogether. An inadequate heat transmission may be caused, however, not only by the presence of air in the intervening space, but still more by the formation of crusts of salts on the inside of the vessel. In cases where salt separates out, stirring must be effected continuously, and the stirrer must approach the walls as closely as possible, in order to keep these scraped clean. If, however, large quantities of salt have separated out no amount of stirring will be of any use. A case is known to the writer where, owing to the formation of a crust of salt only 4 cms. thick, an autoclave became heated up to redness, and with an internal temperature of 2400 and a pressure of 48 atmos., blew out like a balloon, after which the bottom split open. The issuing stream of vapours cooled the steel sufficiently to prevent any further danger. It is quite certain that if cast iron had been used it would have exploded.1