DISTILLATION 261 second might be called " destructive distillations," if it were not customary to reserve this term for the particular case in which the substance operated on consists of vegetable or animal matter which is being decomposed by the application of heat alone, i.e., without the aid of re-agents. The general object of simple distillation is the separation of substances of different degrees of volatility. The apparatus used varies very much according to the nature of the substance operated on and of the product extracted, and according to the scale on which the operation is carried out. Of the various contrivances used in chemical laboratories, the simplest is a glass retort, the descending neck of which is inserted into, and goes to near the bottom of, a slanting globular flask. The retort serves for the reception of the substance to be distilled, and is heated by means of char coal or gas fire ; the vapours pass into the flask, which is kept cool by a continuous current of cold water running over it, or, in the case of more volatile substances, by being immersed in ice or some freezing mixture. This somewhat primitive arrangement works satisfactorily only when the vapours are easily condensible, and when the product is meant to be collected as a whole. In the majority of cases, however, the distillate has to be " fractionated," i.e., collected in a number of separate, consecutive portions; and it is then desirable that the apparatus should be so constructed as to enable one at any moment to examine the distillate as it is coming over. For this purpose it is necessary to condense the vapours on their way to, and not within, the receiver, so that the latter can, at any time, be removed and replaced by another. The condenser most generally used in chemical laboratories is that known as Liebig s condenser. It consists of a straight glass or metal tube, 1 to 3 feet long and ^ to 1 inch wide, fitted co-axially, by means of corks or india-rubber tubes, into a wider tube (made of glass or iron) which communicates at the lower end with a water tap, and at the upper with a sink, so that a stream of cold water can be made to run against the current of the vapour. The condenser tube is fixed in a slanting position, and the vapours made to enter at the upper end. The dimensions of the condenser and rate of water-flow depend on the speed at which the vapour is driven over, and on the temperature of that vapour, and, last not least, on the latent heat of the vapour and specific heat of the distillate. To show the importance of the last- named point, let us compare the quantities of heat to be withdrawn from 1 Ib of steam and 1 Ib of bromine vapour respectively, to reduce them to liquids at C. We have in the case of water and bromine Water. Bromine. For the temperature of the vapour. .. . 100 63 For the latent heat 536 45 0> 6 For the specific heat of the liquids 1 106 For the total heats of the vapours 636 52 3 The withdrawal of 52 3 units of heat from 1 Ib of bromine vapour reduces it to liquid bromine at C. By the with drawal of (- x 52-3 = J 83 units from the steam, as an easy calculation shows, only O lGBb of liquid water, of even 100, could be produced hence more than O84 fib of steam remains uncondensed (at a temperature of about 96 C., assuming the steam to remain saturated, and to have the temperature of the condensed water). But obviously a condenser under all circumstances is the more efficacious the greater its surface and the thinner its body. It is also obvious, cceteris paribus, that the most suitable material for a condenser tube is that which conducts heat best. Hence a metal tube will generally condense more rapidly than one of glass, and for metal tubes copper is better than tin. and silver better than either. In chemical laboratories glass is the 01) ] v material which is quite generally appli cable. In chemical works, on the other hand, glass, on account of its fragility, is rarely used ; condensers there, wherever possible, are made of metal, usually fashioned into spirals (" worms ") and set in tub-shaped refrigerators Where acids have to be condensed, stoneware worms are generally employed. In the distillation of acetic acid pla tinum worms, notwithstanding their high price, have been found to work best, and in the long run to be cheapest. The theory and successful execution of the process assume their greatest simplicity when the substances to be separated differ so greatly in their volatility that, without appreciable error, one can be assumed to be non-volatile at the boiling point of the other. A. good illustration of this special case is afforded by the customary process used for the purification of water. A natural sweet water may in general be assumed to consist of three parts 1st, water proper, which always forms something like 98 per cent, or more of the whole ; 2d, non-volatile salts ; 3d, gases. To obtain pure water from such material, we need only boil it in a distillation apparatus, so as to raise from it dry steam, which steam when condensed yields water con taminated only with the gases. To expel these all that is necessary is to again boil it for a short time ; the gases go off with the first portions of steam, so that the residue, when allowed to cool in absence of air, constitutes pure water. To pass to a less simple case, let us assume that the substance to be distilled is a solution of ether in water, and the object is the separation of these two bodies. Ether boils at 35 C., water at 100 C. The elastic force of saturated steam at 35 is 42 mm., = /<,% = -^th of an atmosphere. Assuming now the mixture to be distilled from a flask, what will go on ? Neglecting for the sake of simplicity the small tension of the steam at 35, we should expect that at first the ether would simply boil away, so to speak, from a bath of warm water at 35 C. ; that the vapour would be pure ether, and maintain that composition until all the ether had boiled off ; then there would be a break the tempera ture of the liquid would gradually rise to 100, and the water then distil over in its turn. And so it is approxi mately, but not exactly. Our theory obviously neglects some important points. Water at 35 has a tension of Y^th atmosphere, ether of one atmosphere ; hence the two saturated vapours together should press with a force of l^th atmosphere in other words, the mixture should com mence to boil at less than 35. This, however (as in the majority of analogous cases), is not confirmed by experi ment. The mixture commences to boil at a little above 35, and the boiling point rises steadily as the proportion of ether in the liquid decreases. Now, a priori, we should presume that at every given moment the volumes of ether and water in the vapour should be, approximately at least, proportional to the respective vapour tensions at the temperature at which the mixture happens to boil. Thus, for instance, assuming at the first that the liquid boils at 40 C., when the two tensions are equal to 910 and 55 910 mm. respectively, the vapour will contain-^ -? = 094 yiO ~r Ou of its volume of ether vapour, and 06 of its volume of steam, supposing both substances to have the same chances of forming saturated vapour, which, of course, holds only so long as they both are present in appreciable quantities. We easily see that, as the distillation progresses, the ether vapour must get more and more largely charged with vapour of water, until at last what goes off is steam, con taminated with less and . less of ether vapour. A thermometer placed near the entrance end of the condenser will, of course, record lower than one plunged into the boiling liquid, because the vapour in rising undergoes partial condensation, and the thermometer being bedewed
with the condensed vapour will approximately indicate the