i8o 



NATURE 



[Dec. 22, 1887 



Fluoride of arsenic, AsFs, the next fluoride experi- 

 mented upon, was first prepared by M. Dumas, who was 

 severely injured in the experiment. It is a liquid which 

 boils at 63" C, and may be easily maintained in a gaseous 

 condition, by use of a steam jacket, and submitted to the 

 action of the spark. It is, however, a most disagreeable 

 substance to work with, as it produces most terrible sores 

 when by any mischance it comes in contact with the 

 operator's skin. On passing sparks through it for an 

 hour, as in case of the pentafluoride of phosphorus, the 

 platinum wires became covered with a black incrustation 

 of arsenic, while the walls of the tube were strongly cor- 

 roded. On testing the gas, it was found to contain a 

 large quantity of silicon tetrafluoride mixed with a smaller 

 quantity of free fluorine, which displaced sufficient iodine 

 from a solution of potassium iodide to give a good colora- 

 tion to several cubic centimetres of chloroform. Clearly, 

 progress was being slowly made, though still far from the 

 isolation of fluorine. 



And now a remarkable experiment of a new type was 

 performed. It had been noticed that, on passing an 

 electric current through a platinum wire in an atmosphere 

 of phosphorus trifluoride, the platinum fused owing to the 

 formation of a fusible phosphide of platinum ; at the same 

 time the glass of the containing vessel was etched and the 

 mercury attacked. So the experiment was repeated on a 

 grander scale. A quantity of spongy platinum, previously 

 washed with hydrofluoric acid and calcined, was placed 

 in a platinum tube 80 cm. long, and of v$ cm. diameter ; 

 that portion of the platinum tube which required to be 

 heated was incased in a second outer tube of glazed 

 porcelain, so that between the two a current of nitrogen 

 could be kept circulating, and so prevent access of furnace 

 gases. The tube was then heated in a furnace, and pure 

 hydrogen passed through it for some time to remove all 

 other gases ; afterwards pure nitrogen was substituted, 

 and finally phosphorus trifluoride. After passing a short 

 time, the current of fluoride was suddenly stopped with a 

 most singular result : a partial vacuum was caused, owing 

 to absorption by the platinum. 



When, however, the current of trifluoride was passed 

 more rapidly, a small quantity of pentafluoride was 

 formed ; the fluorine liberated, when the absorption of 

 phosphorus by the platinum occurred, having combined 

 with the trifluoride just as in the spark experiment. But, 

 on examining the gas which passed out of the tube under 

 these conditions, it was found to liberate iodine from 

 potassium iodide, attack mercury, and etch glass. In 

 fact, it was proved that free fluorine was liberated, and 

 mostly absorbed by the platinum, causing the diminution 

 of pressure on stopping the current, but being more or less 

 carried away when the current was more rapid. The 

 fluophosphide of platinum formed was found to contain 

 only 70 to 80 per cent, of platinum, and the formation of 

 this substance was so rapidly effected that every experi- 

 ment required a new tube. The action of pentafluoride 

 of phosphorus upon platinum was next tried, and with 

 still more encouraging results. On sweeping the tube, 

 heated in a coke blast-furnace, with a rapid stream of the 

 pentafluoride for some minutes, then moderating the 

 rapidity, and five minutes later again increasing the 

 speed, the issuing gas was found to "blacken solid potas- 

 sium iodide by liberating free iodine, inflame phosphorus, 

 and attack crystalline silicon, glass, and mercury. It was, 

 in fact, free fluorine drowned in excess of trifluoride of 

 phosphorus. This was a decided advance, and the out- 

 look was becoming considerably more hopeful. 



The next experiments were made with liquid fluoride of 

 arsenic, AsFg, a quantity of which was placed in a platinum 

 crucible, which served as the negative electrode. A pla- 

 tinum wire, dipping into the liquid in the crucible, and 

 reaching to within 5 millimetres of the base, served as 

 the positive electrode. The current from three Grenet 

 cells was then passed through the liquid, causing a de- 



position of arsenic upon the interior surface of the crucible, 

 but no gas could be perceived at the positive pole. How- 

 ever, on dipping the platinum wire into a solution of 

 starch paste and potassium iodide, blue striae were at 

 once formed in the solution, showing the presence of a 

 condensed gas sheath of fluorine around the platinum 

 wire. Following up this indication, the current from 

 twenty-five Bunsen cells arranged in series was next 

 employed, and immediately the deposition of arsenic 

 commenced upon the walls of the crucible, while bubbles 

 of gas were evolved around the platinum wire. Un- 

 fortunately the action soon ceased, owing to the bad 

 conductivity of the liquid and of the thick deposit of 

 arsenic. The wire, however, was strongly attacked. So 

 attempts were next made to increase the conductivity of 

 the fluoride by the addition of metallic fluorides, and it 

 was soon discovered that the best results were obtained 

 by use of the double fluoride of hydrogen and potassium, 

 HF. KF. It was probably this discovery which led to 

 the grand success with which these efforts have been 

 finally crowned, for, as has been previously mentioned, it 

 was by the electrolysis of this double fluoride that M. 

 Moissan eventually succeeded in preparing free fluorine. 



Before leaving the experiments upon arsenic fluoride, it 

 may be mentioned that it was eventually electrolyzed in a 

 continuous manner by use of seventy to ninety Bunsen 

 cells, the arsenic liberated remaining in suspension in the 

 liquid, instead of adhering to the tube, but the bubbles 

 were rapidly seen to diminish in size in passing through 

 the liquid, and scarcely a trace of gas escaped ; instead of 

 permitting its isolation, the fluorine preferred to form a 

 new fluorid •, the pentafluoride of arsenic, thus once more 

 baffling the ingenious experimenter. 



But success was not now far away. The wonderful 

 manner in which the double fluoride of potassium and 

 hydrogen increased the conductivity of arsenic fluoride 

 determined M. Moissan in employing it for the same pur- 

 pose in an attempt to electrolyze pure anhydrous hydro- 

 fluoric acid. Faraday long ago showed that the electric 

 current will not pass through the anhydrous acid, and 

 Gore more recently came to the same conclusion. The 

 current from fifty Bunsen cells was found by M. Moissan 

 to be absolutely powerless to penetrate the acid used in 

 these later experiments. But, on dissolving a few frag- 

 ments of the double fluoride HF.KF in the acid, the 

 current at once passed freely, and the experiment thus 

 became possible. The apparatus used in the first attempts 

 with this mixture consisted of a platinum U-tube, of which 

 each branch was closed by a paraffined cork, through 

 which the rods of platinum forming the poles were passed. 

 Upon each branch, just above the level of the liquid and 

 beneath the cork, was soldered a little platinum delivery- 

 tube to lead off the gases evolved. As hydrofluoric acid 

 boils at I9°'4 C, the apparatus was immersed in a bath of 

 methyl chloride, which boils at -23°, but which could be 

 reduced in temperature to —50' by driving through it a 

 current of dry air. Hence the electrolysis could be con- 

 ducted without fear of the gaseous products being drowned 

 in excess of vapour of hydrofluoric acid, and the activity 

 of the liberated fluorine was at the same time moderated. 

 On passing the current, a gas was at once produced at 

 each electrode, a regular evolution of hydrogen at the 

 negative pole, and a continuous disengagement of gas at 

 the positive pole. But still affairs were not satisfactory : 

 crystalline silicon did not take fire when held in the gas 

 coming off from the positive pole ; so the apparatus was 

 taken to pieces an hour later, in order, if possible, to find 

 a clue to the source of failure. The paraffined cork at 

 the negative branch was intact, but, behold the mischief, 

 the other was carbonized to the depth of a centimetre ; so 

 the liberated fluorine had extracted hydrogen out of the 

 cork, and passed on as hydrofluoric acid. The positive 

 platinum rod was also much corroded. Closely-fitting 

 stoppers of fluor-spar were next tried, coated with melted 



