During this operation the platinum U tube, dried with the greatest care, has been fixed with a cork in a cylindrical glass vessel surrounded with chloride of methyl. Up to the moment of introducing the hydrofluoric acid, the leading tubes are attached to drying tubes containing fused caustic potash. To introduce the hydrochloric acid into the apparatus, it may be absorbed through one of the lateral tubes in the receiver in which it is condensed.
In some experiments we have directly condensed the hydrofluoric acid in the U tube surrounded with chloride of methyl; but in this case care must be taken that the tubes are not clogged up by small quantities of fluoride carried over, which would infallibly lead to an explosion and projections, which are always dangerous with so corrosive a liquid.
When we have introduced in advance in the small platinum apparatus a determined amount of hydrofluoric acid cooled with chloride of methyl, in tranquil ebullition at a temperature of -23°, the current of 20 cells of Bunsen large size, arranged in series, is passed through by means of the electrodes. An amperemeter in the circuit admits of the intensity of the current being observed.
If the hydrofluoric acid contains a small quantity of water, either by accident or design, there is always disengaged at the positive pole ozone, which has no action on crystallized silicium. In proportion as the water contained in the acid is thus decomposed, it is seen by the amperemeter that the conductivity of the liquid rapidly decreases. With absolutely anhydrous hydrofluoric acid the current will no longer pass. In many of our experiments we have succeeded in obtaining an acid so anhydrous that a current of 25 amperes was entirely arrested.
To render the liquid conducting, we have added before each experiment a small quantity of dried and fused fluorhydrate of fluoride of potassium. In this case, decomposition proceeds in a continuous manner; we obtain at the negative pole hydrogen, and at the positive pole a regular disengagement of a colorless gas in which crystallized silicium in the cold burns with great brilliancy, becoming fluoride of silicium. This latter gas has been collected over mercury, and accurately characterized.
Deville's adamantine boron burns in the same manner, but with more difficulty, becoming fluoride or boron. The small quantity of carbon and aluminum which it contains impedes the combination. Arsenic and antimony in powder combine with this gaseous body with incandescence. Sulphur takes fire in it, and iodine combines with a pale flame, losing its color. We have already remarked that it decomposes cold water, producing ozone and hydrofluoric acid.
The metals are attacked with much less energy. This is due, we think, to the small quantity of metallic fluoride formed preventing the action being very deep. Iron and manganese in powder, slightly heated, burn with sparks. Organic bodies are violently attacked. A piece of cork placed near the end of the platinum tube, where the gas is evolved, immediately carbonizes and inflames. Alcohol, ether, benzol, spirit of turpentine, and petroleum take fire on contact.
The gas evolved at the negative pole is hydrogen, burning with a pale flame, and producing none of these reactions.
When the experiment has lasted several hours, and there is not enough hydrofluoric acid left at the bottom of the tube to separate the two gases, they recombine in the apparatus in the cold, with violent detonation.
We have satisfied ourselves, by direct experiment, that a mixture of ozone and hydrofluoric acid produces none of the reactions described above.