The Behaviour of Glasses when Slowly Heated with Special Reference to the Thermal Expansion
JSGT 1933 V17 T324-T347
Measurements of the thermal expansion from room temperature to the softening point in each case have been made on seven glasses differing widely in composition and in viscosity and other properties. The glasses include fused boric oxide, three soda–silica glasses of soda content ranging between 19·2 and 39·29%, two soda–lime–silica glasses corresponding approximately to types used in commerce, one actually a commercial window glass, the other a pure synthetic glass, and finally an extra light flint optical glass of comparatively complex composition. Determinations of the linear thermal expansion were made when the specimens, of standard length and approximately the same diameter, were heated at the rate of 1° per minute throughout the range and also when heated at the rate of 1/6° per minute. Prior to the determination at 1° per minute, the specimen was annealed according to a special schedule described, involving cooling from the softening temperature at a rate of not more than 1° per minute. Prior to the determination at the very slow rate of heating, the annealing schedule involved cooling from the softening point at a rate not exceeding 1/6° per minute. The results show that with but special exceptions the extreme difference in linear thermal expansion coefficient of the glasses heated at the two rates did not exceed 4%. Of the seven glasses, with five the transformation point was reduced at the slow rate of heating by an amount which did not exceed 8°. Under similar conditions the softening point was also lowered, but in no case could the softening point be reduced to coincide with the transformation point. Transition points at temperatures below 500° were even more noticeable at the slow rate of heating than at the quicker rate. In several glasses also heat absorption effects caused retrogression in the curve just below the transformation point. Four successive slow beatings of a pure soda–lime–silica glass produced thermal expansion coefficient results differing by not more than 4% over the range 0–400°. The conclusion is drawn that whilst the determinations at very slow rates of heating are of special interest in studying the constitution of glass, there is no need, for practical purposes, for a specimen of 5 mm thickness to be heated at a rate slower than 1° per minute.
Eric Seddon & W. E. S. Turner
