An Improved Joule Radiometer and its Applications

Abstract

The first part of the Paper relates to improvements which have been made in order to convert the original Joule convection apparatus into an instrument for the exact measurement of small steady rates of evolution or absorption of heat. These improvements consisted in (1) replacing the badly conducting glass enclosure and cardboard partition by others made of brass and copper respectively; (2) replacing the uncertain and variable magnetic control of the movement of the vane in Joule's apparatus by the elastic control of a quartz fibre; (3) shaping the channels, in which the vanes moved, so that the angular deflection of the vanes was proportional to the rate of evolution of heat; (4) reducing the size, so that a more uniform temperature of its various parts could be easily maintained by (5) placing the radiometer within a concentric brass tube to exclude all extraneous heat excepting that which might be directed through apertures in its side towards the radiometer. The sensibility of the instrument was measured by passing a current through a resistance loop in one of the compartments of the partitioned tube, and found to be equal to 0.52 mm. per microwatt, as measured on a scale at a distance of 1 metre from the mirror. Thus the instrument may be used for the measurement of feeble oscillating currents, it being about as rapid as a Duddell milliammeter. To convert the apparatus into an instrument for the measurement of radiant heat it is suggested that the radiant heat be directed through a small rock salt or fluorite window in the side of a compartment on to a thin blackened metal disc supported centrally by a badly conducting fibre within the compartment. Its use for the quick measurement of the heat given out by radium is also suggested. It is suggested that small steady rates of evolution or absorption of heat might be measured by the compensation methods of Callendar or Angstrom. The second part of the Paper relates to a suggested method of measuring the Thomson effect with this radiometer. The method hinges on an experiment described by the author in Nature, May 18, 1911, p. 380. In that apparatus the halves of a thin wire on either side of the partition are heated by the passage of an alternating current through thicker leads of the same metal. The Joule effects are compensated very nearly by an electro-deposit of the same metal, by scraping the thin wire, or by an auxiliary heating coil. The substitution of a direct current for the alternating current causes a slight heating in one and a cooling in the other half of the wire. The heating or cooling due to the Thomson effect in one half of the wire is compensated by passing a small measured current in the proper direction through the other half of the wire. This small current is passed through a thin lead of the same metal attached to the centre of the thin wire, and may be adjusted in four different ways. The temperature difference between the centre of the wire and the thick leads is measured with suitably attached thermojunctions. The Thomson coefficient is expressible in terms of measurable quantities, and is equal to the product of the mean compensating current and the mean resistance of the halves of the wire divided by the temperature difference between the centre of the thin wire and its thick leads.