LXIII. The Production of Soft Röntgen Radiation by the impact of Positive and Slow Cathode Mays, By Sir J. J. Thomson, O.M., F.M.S [Communicated by the Author, having been read before the Physical Society, June 20, 1914.] A gap of about eight octaves separates the softest characteristic Röntgen radiation yet investigated, that of aluminium, and the shortest waves in the Schumann region, those recently discovered by Professor Lyman; the latter have a wave-length of about 9 x 10~ 6 cm., the former one of 3*6 x 10 ~ 8 cm. Very little is known about any radiation of intermediate wave-length, and yet the study of such radiation is essential for the determination of the structure of the atom. By its aid we might hope to gain a knowledge of the distribution of the electrons in the atom, to determine, for example, how many rings of electrons there are in the atom and the number of electrons in each ring. We associate the K and L types of radiation with the vibrations of the two rings nearest the centre of the atom, and the visible spectrum with those of the outermost ring. By using Röntgen radiation ranging in hardness from that characteristic of aluminium to the hardest we can produce, we can detect the existence and study the properties of the two rings nearest the centre. By using the light of the visible and ultra-violet region we can find out a good deal about the outermost ring, but to study the intermediate rings, and thus get even an approach to the constitution of the atom, we require radiation intermediate between Schumann and ordinary Röntgen rays, a type of radiation which has not hitherto attracted much attention. I wish in this paper to describe some experiments recently made in the Cavendish Laboratory on methods of producing and studying radiation of this type. In the first place I wish to describe a new method of producing this radiation, for I find that when positive rays impinge against a solid, radiation of this type is produced. The apparatus is shown in fig. 1. The positive rays are produced in the bulb A, they pass through a tube about 2 mm. in diameter and 5 cm. long in the cathode C, and strike against a platinum disk P. B is a side tube, at the end of which is the arrangement described in my book on 'Positive Rays,' by which a photographic plate could be exposed to the radiation in the tube: a slit was [header] 621 placed in front of the plate so as to get a definite image. The plates used were Schumann plates or plates specially prepared for me by the Paget Plate Company; these are not so sensitive to the radiation as the Schumann plates, and require a longer exposure ; when, however, the radiation is [figure redacted] strong enough they are more convenient to work with. L and M are two pairs of parallel plates 6*5 cm. long and 1 mm. apart, placed so that any radiation coming from A and striking the plate P must pass between L, while any radiation from P must pass through M before it reaches the photographic plates. These plates could be connected to a large battery of small storage-cells, and a potential difference of 1000 volts established between the plates in either pair. When the positive rays were striking against the plate an exposure of a Schumann plate for an hour gave a dense photograph of the slit, showing that something was coming down the side tube which could affect a photographic plate. There are many well recognized types of radiation produced in A, and it is necessary to make further experiments to see if the photograph is produced by these or by some new type of radiation. It might be suggested, for example, that the effect had nothing to do with positive rays passing through the cathode to P, but was produced by ultra-violet light or Röntgen rays generated in the discharge-tube, which passed through the aperture in the cathode, and was then reflected from the platinum plate. If this were the case, since neither ultra-violet light nor Röntgen rays are deflected by an electric field, the intensity of the photograph should be the same whether the potential difference between the plates in L was zero or 1000 volts. It was found, however, that the photographic effect almost disappeared when there was a potential difference of 1000 volts between the plates, showing that the effect is due to the positive rays. The effect, though reduced to a small fraction of its former value by putting on 1000 volts, is yet not entirely eliminated. The small residual 622 [header] effect is due, I think, to the fact that the positive rays, except at the very lowest pressures, do not; remain constantly positively charged, but alternate from the charged to the uncharged condition : thus a few of them might remain without charge all the time they were between the plates and thus escape deflexion ; after passing through the plates they might re-acquire a positive charge before striking against P. I think this is more probable than that the residual effect is due to ordinary Röntgen rays or ultra-violet light produced in A, for the rays, as we shall see, are too easily absorbed by white fluorite to be ultra-violet light, and by thin mica or collodion to be ordinary Röntgen rays. Another proof that the effect is not due to stray radiation from A, is that it disappears entirely if C is made anode instead of cathode. One of the effects of the impact of positive rays against a metal plate is to make the plate emit slow cathode rays, and it might be thought that the effect on the photographic plate was due to these rays, starting from P and travelling down to the plate. If this were the case, then putting a potential difference of 1000 volts between the plates in M ought to stop the effect entirely. I find, however, that the photographs are just as dark when 1000 volts are on the plates as when they are at the same potential. This seems a conclusive proof that the radiation which affects the plate is not a corpuscular radiation or a form of positive rays, but is analogous to light or Röntgen radiations, that in fact Röntgen radiation is produced by the impact of positive rays against a solid. This radiation is unable to penetrate even the thinnest films I have been able to procure of substances such as collodion, mica, paraffin-wax, aluminium, or white fluorite: when part of the slit was covered with one of these films it entirely stopped the radiation through that part of the slit. This radiation can be reflected, for if a slit of the kind shown in [figure redacted] fig. 2 is put in front of the photographic plate, it is found that the plate is affected not only underneath A but also on [header] 623 the part to the left of the opening B. Part of this reflected radiation is corpuscular as it is affected by a magnet ; a part of it is not so affected, and so must, like the incident rays, be a form of Röntgen radiation. I should estimate the velocity of the positive rays at about 2 x 10 8 cm./sec; the impact of cathode rays possessing the same energy as these would generate a very hard type of Röntgen ray; the type of Röntgen ray generated by the impact of positive rays more nearly resembles that produced by cathode rays with the same velocity but much less energy than the positive rays. I now pass on to consider the production of soft Röntgen radiation by the impact of slow cathode rays. The arrangement used is shown in fig. 3. is a Wehnelt cathode — a [figure redacted] thin strip of platinum foil with a patch of barium oxide deposited on it by burning away a speck of sealing-wax. The anode A is a piece of brass rod with a hole bored through it through which the cathode rays pass on their way to the target B, a copper plate which is at the end of and in metallic communication with a cylinder of wire gauze. The variation in the speed of the cathode rays was produced by putting between the gauze and the anode an electromotive force tending to stop the rays. Thus if Vj is the potential difference between the anode and the cathode, V2 that between the anode and the gauze, the energy of the cathode rays when they strike the target is proportional to V1 — V2 . This method of varying the energy of the rays was found to work better than altering the potential between the cathode and the anode, as the emission of cathode rays from C was much more regular with a constant potential difference between the anode and cathode. To detect the radiation coming from the target, a camera similar to that used in the previous experiment was placed at the end of the side tube T. A slit was placed in front of the 624 [header] plate, and half of it covered by thin slices of paraffin-wax, collodion, mica, glass, or fluorite so as to estimate the penetrating power of the radiation. A magnet was placed between the target and photographic plate so as to deflect from the latter any corpuscular radiation from the target. The vacuum was made as low as possible by means of charcoal and liquid air ; it was so low that no luminosity could be detected between the anode and the target. The plates used were Schumann plates ; the Paget plates were not sensitive enough to detect the radiation from the slowest cathode rays, though they gave good photographs when the rays fell through an effective potential of more than 100 volts [Since this paper was read I Lave, by using a more copious supply of cathode rays, been able to get photographs at the lower voltages with Paget plates.]. The times of exposure, which varied from 1 minute to 2 hours, were chosen so as to make the energy in the cathode rays striking against the target during the time of exposure constant: thus with cathode rays which had fallen through 20 volts, the time of exposure would be ten times that for those which had fallen through 200 ; the latter gave quite dense photographs with an exposure of 2 minutes. I have obtained photographs with cathode rays whose energy ranged from 10 to 600 volts, and there would be no difficulty in getting those corresponding to higher voltages by using larger batteries to produce the main discharge. These photographs are not due to ordinary light coming from the discharge-tube, for (i.) they are not obtained when the beam of cathode rays is deflected by a magnet from the target, and (ii.) the rays which produce them are unable to penetrate exceedingly thin films of glass. To test whether they were due to a corpuscular radiation from the target two methods were employed. First, a magnet was placed between the target and photographic plate, so as to deflect the corpuscular radiation from the plate; this did not affect the photographs. The second method was to place between the target and the photographic plate a pair of parallel plates similar to those used in the first experiment with positive rays, and apply to them a potential difference of 1000 volts. The intensity of the photographs was not diminished when all the radiation which struck the plate had passed through this strong electric field, which would have stopped any charged particles. With regard to the penetrating power of this radiation, when it is produced by cathode rays with less energy than 40 volts, I have never been able to detect any photographic [header] 625 effect behind a film of collodion thin enough to show the colours of thin plates, paraffin-wax 4 [mu] thick, mica, or thin fluorite [With larger currents from the Wehnelt I have been able to detect the photographic effect of the 40 volts rays behind thin collodion and mica, and also to detect, the photoelectric effect and ionization due to the rays which had passed through the films.]. When the energy corresponds to 80 volts the effect behind the paraffin and collodion is appreciable, while with 200 volts and more there is very considerable penetration of the collodion and paraffin by the rays. The great opacity of very thin films suggests that the frequency of the rays may be within tho limits of those vibrations which, according to the usual theory of dispersion, are totally reflected by a medium. According to this theory, if the medium has only one free period of frequency n, it is impervious to light whose frequency p is between the limits given by the equation [formula redacted] and [formula redacted] when N is the number of electrons per unit volume, e and m are respectively the charge and mass of an electron. This relation applies when the wave-length is large compared with the distance between the molecules; and the limits of opacity depend on the degree of closeness with which the molecules are packed : for example, in a gas they depend upon the pressure. In the case of Röntgen rays when the wave-lengths are small, or even comparable with the distance between the molecules, the case is different. The effect of matter in this case is not so much to increase the refractive index as to scatter the radiation, and this scattering will be greatest when the atom is impervious to the radiation. If we apply the equation to the atom itself, it indicates that the atom would be impervious to and therefore scatter strongly rays whose frequency is between limits which depend on the density of the electrons within the atom, and not on the closeness with which the atoms are packed. When there are more frequencies than one intrinsic to the atom, there will be several regions of great opacity separated by intervals of comparative transparency. I am indebted to my assistants Mr. Everett and Mr. Engle for the assistance they have given me in these experiments. [footer]