Some time ago I exhibited at a soirée of the Royal Society a few experiments with the Crookes radiometer, the object being to show that when helium is the residuary gas filling the instrument, an attached charcoal condenser, even when placed in liquid hydrogen, is unable to diminish the pressure by absorption to such an extent that the radiometer will not rotate (when subjected to the concentrated beam of an electric arc lamp focussed upon the black surface of the mica vanes); while, on the other hand, if the gas instead of being helium is hydrogen, all radiometer motion is suspended. Even when the charcoal condenser of the helium radiometer was cooled in solid hydrogen under exhaustion so that a temperature of 15° absolute was reached, the rotation of the instrument was still very marked. If the radiometer is repeatedly washed out with the mixed oxygen and nitrogen got from the evaporation of liquid air, the charcoal and the whole of the glass being thoroughly heated and the apparatus finally exhausted to a fraction of a millimetre and sealed off, then, on placing the charcoal tube in liquid air, generally after an hour or two the vacuum is so high that no motion is induced by the beam of the electric arc. But if instead of liquid air the cooling agent is liquid hydrogen, then two minutes’ immersion is sufficient to effect the same result, provided the radiometer is small and the gases get down a quill tube direct into the charcoal. Instead of the gases from liquid air being used to clean out the radiometer as described, it is for some purposes better to seal on a side tube containing perchlorate of potassium, which, when heated, gives pure oxygen. Further, in many experiments it is advantageous to exhaust the radiometer with its little charcoal condenser by means of a larger quantity of charcoal placed in liquid air for a night and then to seal the latter off before cooling the special charcoal bulb attached to the radiometer. When a McLeod gauge was sealed on to the end of the bulb containing the charcoal condenser which is cooled in liquid air (no stop-cocks of any kind being used), all the mercury vapour was eliminated from the radiometer, and the pressure of the permanent gas was found to be 0·00001 mm., or one seventy-six millionth of an atmosphere. In this condition the radiometer moved when the image of the poles was focussed on the black vanes and after some 15 minutes’ heating the pressure was found to be only one twenty-five millionth of an atmosphere and the pressure remained at this after 10 hours’ cooling of the charcoal condenser in liquid air. The gas produced was no doubt hydrogen, got from the lamp black of the mica vanes, this being the first time the instrument was used. As a rule the radiometers require to be refilled, exhausted and tested more than once in order to get the motion reduced to a minimum. The importance of the removal of traces of gases like helium, hydrogen or neon is shown from the fact that a radio meter which has the charcoal removed from the attached bulb, and the latter cooled in liquid hydrogen (the instrument having been previously filled with dry air and exhausted to a fraction of a millimetre of mercury), will not reach such a vacuum as to stop the radiometer motion. Now, as the pressure of nitrogen at the boiling point of hydrogen must be of the order of a millionth of a millionth of an atmosphere, the action must come either from uncon-densable gases, or the persistent adhesion of gas molecules to the glass and vanes of the radiometer or to some solid matter volatile under the conditions of the experiment. The lowest pressure reached in a charcoal vacuum after 10 minutes’ cooling in solid hydrogen was still one hundred millionth of an atmosphere. The pressure observed is thus far too high and it may he that some of this is due to hydrogen coming from the charcoal. To get really high vacua by the charcoal method, even when liquid hydrogen is the cooling agent, it seems necessary to allow the absorption to go on for an hour or more, when the space to be exhausted is relatively large, and where narrow tubes or orifices constitute part of the apparatus, as in the McLeod gauge. Further, the presence of any organic matter on the vanes is fatal. No amount of cooling of the charcoal in liquid hydrogen of a radiometer filled as usual and tested in the ordinary manner, in which the vanes were made of pith, makes a vacuum sufficient to stop the radiometer motion. The concentrated beam, each time it was applied, was generating gas. In all the experiments the arc used was expending 10 ampères and the focus was adjusted to about 3 feet from the lamp. The radiometers had a volume of from 150 to 20 c. c. Finding the McLeod gauge very difficult to use, a new method of defining the maximum limit of the working pressure (under the defined, circumstances) depending upon the vapour-pressure of mercury was devised. For this purpose a side tube was sealed on to the top of the radio-meter and this, after being bent twice at right angles, ended in a little bulb containing a globule of mercury. After the radiometer and charcoal were heated and exhausted and repeatedly washed out with the gas from liquid air, the charcoal was cooled in liquid air and the mercury allowed to distil for an hour or two. After this treatment, on cooling the mercury with liquid air, the radiometer in a short time became inactive. In this condition the mercury was placed in an alcohol bath at - 80° C. and the temperature allowed to rise slowly.
The Measurement of the Intensity of the Raman Spectrum. II-III. III. The Measurements of the Depolarization Degree and the Absolute Intensity Ratio of Raman Lines.