Flash powders should never be stored in glass-stoppered bottles, for the friction between the stopper and a small quantity of the powder may be quite sufficient to cause the explosion of the bottle and its contents.
300. Constituents of Flash Powders. Flash powders should contain no poisonous substances, nor any substances which give off poisonous fumes, such as chromates.
There should be no deliquescent substance present, which will absorb the moisture from the air and in this manner prevent or delay the combustion. All substances which attack mag nesium, or which are liable to cause spontaneous combustion of the mixture (this sometimes occurs when cadmium nitrate is used), should be excluded. Products which make the mixture susceptible to shock, as an example of which potassium chlorate may be cited, should never be added. It is desirable to choose those com bustion agents which give the least solid residue.' Lastly, the constituents should be in a sufficiently finely divided state to permit of rapid combustion, without, however, using the magnesium in the form of too fine a powder, in which condition it is rapidly oxidized and may, in certain cases, cause the spontaneous combus tion of the mi xture.
One of the best magnesium flash powders is obtained from a mixture of magnesium and perchlorate of potash, which are powdered separately before mixing, the proportions being then chosen according to the desired result. The most rapid combustion of a second for a charge of from 15 to 30 gr.) is obtained with equal quantities of the two constituents, but the intensity of the light is improved if the proportion of perchlorate is reduced. A mixture of 30 gr. magnesium with 15 gr. of perchlorate (time of combustion about .2th of a second for the same charge as above) is to be preferred because of its higher actinic value, in all cases where extremely rapid combustion is not necessary.
Excellent flash powders, generally with slower combustion for the same actinic value, can be prepared by replacing the magnesium by aluminium (not the impalpable aluminium powder used for metal paints), the most suitable combustion agent in this case being either potassium or barium permanganate (15 gr. of metal to 3o gr. of permanganate).
Powders of exceedingly high actinic power and burning with a very small percentage of smoke, have been prepared by mixing cerium, zirconium, and thorium (much more costly metals than magnesium) with their respective perchlorates.
301. Testing of Flash Lamps and Flash Pow ders. Tests of flash lamps and flash powders should be designed to indicate actinic value and also speed of combustion. It may also he useful in the case of powders to determine the volume of the flash, which will give some idea of the diffusion of the light, and, if necessary, of the required dimensions of smoke traps or diffusers. Various methods have been outlined for these tests by A. Londe (iqoi), J. 1. Crabtree (i916), and others.
One of the simplest methods for the deter mination of the actinic value is to expose to the flash one half of a plate placed behind an even neutral grey density (transmitting about one-thousandth of the incident intensity). The other half is then exposed behind a sensito metric wedge (a neutral grey filter of gradually increasing density) at a known distance from a light-source of known intensity. Having devel oped the two halves of the plate together, a comparison is then made of the exposures necessary to produce equal density on each half, and from this comparison the actinic value of the flash can be expressed in terms of that of the known light-source.
In order to determine the duration of the combustion, and, at the same time, the variation in actinic intensity at various stages of the combustion, the powder is ignited behind an opal glass, and the flash photographed with a high frequency cinematograph camera. The exactness of the measurement increases with the speed at which the camera is run, provided the frequency of the images is accurately known. From a photometric comparison of the densities registered on the various sections of the film with a scale of densities made on the same film under known conditions, the actinic value of the flash at the moment of registration of each image can be determined.' The dimensions of the flash may be deter mined by photographing it behind a screen (formed of squares of known size) against a black background. By placing a mirror at an angle of 45° close to the flash, two images from directions at right angles to one another (on somewhat different scales) may be obtained simultaneously.