XIV. Radioactivity of the Atmosphere. By S. J. Allans. M.Sc., Demonstrator in Physics, McGill University [Communicated by Prof. E. Rutherford, F.R.S.]. [Plate XIV.] In a previous paper [citation redacted] some experiments were described dealing with the rate of decay and penetrating power of the excited radioactivity obtained from the atmosphere on a negatively-charged wire. It was found that the activity decayed according to an exponential law with the time, falling to half its value in 45 minutes. Its penetrating-power was slightly greater than that of the excited activity from radium or thorium. Its absorption by solids followed an exponential law with the thickness, and the radiation was cut down to half its value by *001 cm. of aluminium. The amount of excited activity that could be obtained from [header] 141 the air at any time was found to he strongly influenced by weather conditions. A cold, clear, windy day gave the largest amount and a warm dull day the least. In the experiments described in the present paper the radioactivity was obtained from a closed room which gave a constant amount from day to day. The electrical method of measuring the radiations has been used throughout the experiments. The electrometer employed was of the ordinary Thomson quadrant type, fitted with a needle of light construction, which was kept connected to one pole of a battery of 300 volts. A reflecting mirror and a millimetre-scale indicated the movement of the needle. Each scale-division corresponded to 1/500 of a volt P.D. of the quadrants. A small quantity of uranium served to standardize the readings of the instrument. Increase of Excited Activity with Time. It has been shown in a previous paper [citation redacted] that the excited activity derived from the air decays according to an exponential law, the rate of decay being given by the equation [citation redacted] where I is the excited activity at any time t, I the maximum value, and [lambda] a constant. If the excited activity produced on a negatively-charged wire is due to a constant supply of positive radioactive carriers whose activity decays according to the above equation, then the intensity of the activity It after an exposure for a time t will be given by [citation redacted] where I0 is the maximum value and [lambda], the same constant as before. The following experiment was made to verify this view : — About 60 feet of copper wire was suspended in a large attic and kept charged to a constant negative potential of about 20,000 volts, by means of a Wimshurst machine driven by an electric motor. In parallel with the charged wire was arranged an adjustable spark-gap to regulate the constancy of the potential of the wire. After a certain time of exposure the wire was taken down and wound lengthwise on an iron frame. This frame was placed inside a cylindrical vessel of zinc and connected to one pair of quadrants of the electrometer, the other pair being earthed. The zinc cylinder was connected to one pole of the battery and the other pole was to earth. Between the iron frame 142 [header] and the cylinder was arranged a guard-ring, to prevent any leak around the ends of the electrodes. The rate of movement of the electrometer-needle was taken as a measure of the amount of excited activity present. The wires were always tested five minutes after removal from the attic. The results obtained from these experiments are given in the following Table, the second column giving the amount of excited activity on the wire five minutes after it was taken down. Table I. Time of exposure. Excited activity produced. [table redacted] The results are shown graphically in fig. 1 (PI. XIV.). From an examination of these results it will be seen that the excited activity increases with time according to the equation given above, rising to half value in about 60 minutes. It has been shown that the excited activity on a charged wire decays to half value in 45 to 48 minutes. The results are thus only approximately in agreement with theory, since it is difficult under the experimental conditions to obtain more than rough results. Rates of Decay under various Conditions. In the previous paper it was shown that the excited activity on a charged copper wire always had the same rate of decay, wherever and whenever produced. It was, however, deemed advisable to examine the rate of decay of the excited activity produced under as many different conditions as possible. Iron and lead wires were tried and each gave the same rate of decay as the copper wire. Experiments were then made to see if by transferring the excited activity from the copper wire to such substances as leather and felt, by means of ammonia, any difference could be observed in the rate of decay. For this purpose about 150 feet of copper wire was suspended in the attic and kept charged for about three hours a constant negative potential of 20,000 volts. A piece of leather, about 7 cms. square, moistened with ammonia, was then rubbed over the wire, and by taking care to rub a fresh [header] 143 part of the leather over the wire every five feet, a large quantity of excited activity could be transferred from the copper wire to the leather. This radioactive leather was then tested n the following apparatus: — Two parallel zinc plates were placed horizontally over one another, and insulated. The upper plate was connected to one pair of quadrants of the electrometer, the lower plate being connected to one pole of the battery, the other pole of which was earthed. The radioactive leather was placed on the lower plate, and the rate of movement of the needle taken as a measure of the quantity of radioactivity present. About 100 volts P.D. between the plates was sufficient for saturation. The apparatus was inclosed in a metal box connected to earth, which acted as a guard-ring. It was found that with thin close-grained leather the excited activity decayed to half value in about 45 minutes, whilst with a thicker and more absorbent leather it took 48 to 52 minutes. A piece of felt, moistened with ammonia, was also rubbed over the wire, and tested in the same way. Only a small quantity of excited activity could be observed in this, but it was found to decay much more slowly, falling only to half value in 60 minutes. Another thick spongy piece of felt gave a rate of decay even slower. If, however, the felt was reduced to ashes and the residue tested, much more radioactivity was observed than the felt itself showed. The reason for this seems to be that the ammonia dissolves off the matter which causes the radioactivity and carries it into the interior of the felt. Some of the radiation is thus absorbed in the felt before it can reach the surface. The rate of decay of the activity of the ashes from the felt is about the same as for the copper wire. In PI. XIV. fig. 2, curve 1. shows the decay of the activity of the unburnt felt, and curve II. that of the ashes from the felt. The difference in the rate of decay in the two cases is probably due to the fact that the penetrating part of the radiation which passes through the felt has a rate of decay different from that of the whole radiation. A piece of cotton wool, moistened with ammonia and rubbed over an active copper wire, then reduced to ashes and tested, gave a very large amount of radioactivity. The rate of decay is the same as for the excited activity on copper. This is a good method of obtaining a large quantity of excited activity in a concentrated form. It produces more ionization than uranium bulk for bulk. When a piece of copper wire which has been made radioactive 144 [header] was partly dissolved in ammonia and the solution evaporated, the residue gave the same rate of decay as the excited activity on the solid copper. The activity excited on a piece of rubber also gave the same rate of decay. It has been shown by C. T. R. Wilson, [citation redacted] and independently by the author [citation redacted], that freshly fallen snow, when evaporated down to dryness, leaves behind a residue which is temporarily radioactive. The rate of decay of this radioactive residue was examined. Snow was gathered from a thin sheet on the surface during a snow storm, evaporated down to dryness in a shallow dish, and placed in the parallel plate apparatus described above. A litre of snow, when evaporated, produces about the same effect as one-fifth of a gramme of uranium. The rate of decay was found to follow an exponential law, and the activity fell to half value in from 30 to 32 minutes. Thus there is a distinct difference between the rate of decay of the radioactivity on snow and that obtained from the air. The penetrating power, however, was found to be about the same for both. In fig. 3 are two curves showing the rate of decay of the radioactivity from snow. C. T. R. Wilson has shown that freshly fallen rain, when evaporated, leaves a radioactive residue. This residue was tested in the same way as that obtained from snow, and it was found that the rate of decay was the same, the activity fell to half value in about 32 minutes. Wilson states that the activity of the rain-water he examined fell to one-quarter value in one hour. In fig. 4 are shown two curves for the decay of the radioactivity from rain-water, taken on different days. Fig. 5 shows the curves of decay of the radioactivity from snow and of the activity excited on a lead wire and of the excited radioactivity from a copper wire transferred to felt, all plotted to the same scale for the sake of comparison. Absorption of the Excited Activity by Solids. A number of experiments were made with various solids to see if the absorption-density law held for the excited activity. The excited activity on a copper wire was transferred to a thin piece of leather moistened with ammonia, and the leather was placed between the parallel plates of the testing apparatus. Readings were taken when the leather was bare and when covered with thin layers of various solids. From the curve of decay and these readings the percentage absorption could [header] 145 be calculated. The results for aluminium foil of average thickness '00038 cm. are shown in fig. 6. The lower one is for the excited activity transferred to leather and the upper one for that on a lead wire. The ordinates give the percentage of rays unabsorbed, and the abscissae the number of layers traversed. It will be seen that there is a small difference in the penetrating power of the two. In fig. 7 are plotted two curves, showing the penetrating power of the excited activity from (1) a thick piece of felt, and (2) a thin piece of leather? Curve II. follows closely a G.P. with the thickness up to about 1 layers, the radiation falling to half value after passing through about 2 layers of aluminium. After 10 layers there is a marked difference in the curve, 5 additional layers having only a small effect. With 25 layers *8 per cent, of the rays was still unabsorbed. The portion of the curve between 10 and 25 layers follows roughly a G.P. These results show the presence of a more penetrating kind of radiation. Curve I. for the felt shows the same effect, except that there is a greater percentage of penetrating rays than in the case of the leather. This difference can be explained on the assumption that there are two kinds of radiation given off — one, the [alpha] radiation, being practically all absorbed in about 10 layers of the aluminium foil, and a more penetrating [beta] radiation. Now, the greater part of the excited activity on the felt would be carried into the interior by the ammonia, and hence the radiation would have to penetrate through a considerable thickness of felt before reaching the surface. The [alpha] rays will thus be largely absorbed, arid the radiation at the surface consist of a greater percentage of [beta] rays, which pass through more easily. This also explains why such a small amount of energy is given off from the felt compared with that given off from the leather, since in the former case the [alpha] rays are mostly absorbed. The a rays represent about 99 per cent, of the total energy radiated ; the [beta] rays are half absorbed in about 15 to 18 layers of the aluminium. The excited activity from the ashes from radioactive cotton was tested in the same way, and showed the presence of [beta[ rays, though not in so large a percentage as the leather and felt. The three cases are compared in the following table. If the radiation consists of rays of a homogeneous character, then the intensity I, after passing through a distance d of the [footer] 146 [header] Table II. No. of layers of Aluminium. Percentage of unabsorbed rays. [table redacted] absorbing material, will be given by the equation [formula redacted] I0 is the intensity at the surface before any absorbing layer is laid on, [lambda] being the coefficient of absorption of the material considered. If the absorption is proportional to the density, then the ratio [lambda]/density should be a constant, A number of substances were tried to test this point. It was difficult to get the solids in sufficiently thin layers to give enough ionization to work with. The [beta] rays produced too little ionization for accurate measurement. The experiments were performed in the same way as already described, the solids tested being mica, celluloid, paper, aluminium, brass, tinfoil, silver, and Dutch-metal. The value of [lambda] for each substance could be easily calculated from the curves of absorption. The results are shown in the following table. Substance. [lambda]. Density. [lambda]/ Density [table redacted] From an examination of these results it will be seen that [header] 147 for the light substances and for aluminium the absorption is nearly proportional to density, but for the heavier metals there is a wide divergence. These results are similar in character to those of Rutherford and Miss Brooks [citation redacted], who examined the [beta] radiation from uranium, and found that the ratio [lambda]/density was the same for density Absorption in Gases. A series of experiments were also made on the absorption of the excited activity by air, coal-gas, carbonic-acid gas, and hydrogen. For this purpose a special apparatus was constructed, the general arrangement of which is shown in fig. 8 (PL XIV.), and is similar in principle to that used by Rutherford in his experiments on uranium radiation [citation redacted]. It consists of a cylindrical brass vessel, closed at the top by an air-tight cover, and at the bottom by a mercury trap. It is divided on the inside into two chambers by means of a horizontal partition, which has a circular hole cut in it, covered with a sheet of aluminium "00038 cm. thick. The partition was insulated from the sides of the cylinder, and connected to one pole of the battery, the other pole being earthed. Immediately below the partition was a circular table,, which could be moved up and down by means of a screw passing through the bottom of the cylinder. At the top of the upper chamber was suspended an insulated disk, connected to the electrometer. The radioactive leather was placed on top of the table immediately underneath the aluminium foil. The radiation given off by this leather penetrated through a certain layer of air or any gas with which the cylinder might be filled, and thence through the aluminium foil into the upper chamber, where it could ionize the gas and produce a movement of the needle of the electrometer. The brass cylinder was earthed and acted as a guard-ring, preventing any leak along the sides. The radiation in passing through the layer of gas before reaching the upper chamber, would be absorbed to an extent depending on the thickness of layer traversed. This thickness could be regulated by means of the screw. The volume of gas in the upper chamber remaining constant the ionization produced there would always be a measure of the strength of the radiation unabsorbed after passing through 148 [header] a given thickness of gas. A fixed distance, about 6 mms., between the leather and the aluminium, was always taken as a basis from which to calculate the percentage of unabsorbed rays. If the radius of the active surface is large compared with its distance from the aluminium foil, it can be readily shown from the ionization theory that the following equation holds : [formula redacted] I0 is the intensity of the radiation after passing through a distance [xi] of the gas, and [lambda] the coefficient of absorption of the gas considered. The percentage of the radiation unabsorbed is calculated in the same way as for solids. The results of the experiments are shown plotted in fig. 9 (PL XIV.), the ordinates giving the percentage of rays unabsorbed after passing through a certain distance, and the abscissas the turns of the screw-head, each turn corresponding to 1*27 mm. These results are compared with those of aluminium in the following table. Substance. Radiation reduced to half its value in [lambda]. Density. [lambda]/ Density [table redacted] It will be seen that the absorption by gases follows the order of their densities, and is almost proportional to density for air and carbonic-acid gas. Increased Conductivity of Air mixed with Water Spray. J. J. Thomson [citation redacted] describes some experiments in which the conductivity of air was increased by passing it through a water-pump into a large vessel, where it was tested. He also found that when a brass rod was suspended in this vessel and kept charged for a number of hours to a high negative potential, it had acquired a certain amount of excited activity [Note. — The effects observed by J. J. Thomson have since been shown by him to be due to a radioactive emanation present in the tap-water of Cambridge.] [header] 149 In view of the importance of these results, experiments were undertaken to see if the Montreal tap-water derived from the River St. Lawrence showed similar properties. For this purpose a large cylindrical zinc tank, diameter 102 cms. and height 150 cms., was used. In the centre of this was suspended a brass cylinder 5 cms. in diameter, which passed through an ebonite plate at the top of the tank, and was connected to the electrometer. The outer cylinder was connected to a battery of 300 volts. Between the two cylinders was arranged a guard-ring connected to earth. A rubber tube passed from the bottom of the tank to an ordinary water pump, from which a return tube entered the top of the tank. The natural leak of the tank filled with ordinary air was first observed, and found to vary from 4 to 5 divs. per sec. The water pump was then started, and the moist air circulated through the tank, while readings of the conductivity were taken every minute. The conductivity of the air in the tank immediately began to increase, and reached a maximum in about five minutes, reaching in one test 25 divs. per sec, or nearly six times the natural leak. When the water pump was stopped this increased conductivity at once began to decrease, and reached the natural leak in about six to eight minutes. The maximum varied from time to time, but was always from four to six times the natural leak. This modified air, when passed through pumice-stone saturated with sulphuric acid before reaching the tank, only gave 8 divs. per sec. as the maximum, but as soon as the pumice-stone was removed gave 20 divs. per sec. It was found that the quicker the air was drawn through the tank the greater was the conductivity produced. Passing the air through a cotton-wool plug destroyed a large portion of the conductivity. It was also found that when the moist air was passed through a spiral tube immersed in liquid air, or a tube heated to redness, the increase of conductivity previously observed was completely absent. The experiment was tried of allowing a quantity of liquid air to evaporate inside the tank, but no increase of conductivity could be observed. A brass rod was suspended in the tank, and kept charged to a high negative potential for several hours, whilst the air charged with water-spray was circulating through. It was then removed and tested in another vessel, but no signs of any excited activity could be detected. I think we may conclude from these experiments that the increased conductivity is not caused by an emanation in the water-spray, since it will not stand the tests to which an emanation may be subjected. Neither is there any 150 [header] appreciable excited activity produced on a rod suspended in it. It takes a far greater volume of air than the tank held to produce any measurable amount of excited activity from the air, unless some radioactive substance, such as thorium or radium, is present. There is certainly an increase of conductivity produced, which dies away quickly, and which i& undoubtedly caused by the mixture of the water-spray with the air. The water from the tap, when evaporated down to dryness and tested, gave no signs of any radioactivity. Conclusion. From these results we may conclude that the excited activity from the atmosphere behaves in many respects like the radioactivity from thorium and radium. It contains, as they do, an easily-absorbed [alpha] radiation, and a more penetrating [beta] radiation. The [alpha] radiation is probably responsible for the greater part of the total energy radiated, and it is completely absorbed in about *004 cm. of aluminium and 10 cms. of air. The ft rays are cut down to half value in *007 cm. of aluminium, and completely absorbed by '06 cm. The [beta] rays probably consist of negatively-charged particles, similar to cathode rays, and projected with great velocity. The ionization produced by them is too small to test whether they are deviable in a magnetic field. The difference in the rates of decay of the excited activity obtained under different conditions seems to point to the fact that the radioactivity of the atmosphere is of a very complex nature. The radioactivity of snow and rain must be derived from some radioactive matter in the air which adheres to the surface of the snow-flake or rain-drop, and is brought down with it in its descent. A possible explanation of the difference observed in the rate of decay of the radioactivity from snow and rain, and that of the excited activity on a wire, may be based on the view that the radioactive matter in the air is of different kinds, having different rates of decay. Snow and rain may owe their activity to one kind while the negatively-charged wire removes all the active carriers to its surface. The rate of decay of the charged wire might thus be the resultant of several different rates. In conclusion, I wish to thank Prof. Rutherford for his kindly interest in the work. McDonald Physics Building, McGill University, Aug. 1903.