SCANNING-BEAM MICROSCOPE

Until recently, microscopy remained outside the general tendency to expand the capabilities of optical apparatus by equipping it with radio-engineering devices. The microscope with a scanning beam described below[^1] is apparently the first attempt of this kind. The idea of the device is very simple and is borrowed entirely from television technology. Instead of the usual wide-angle light beam, which illuminates the entire field of view simultaneously and creates a diffraction image of the whole object, a thin light beam is used here, successively traversing all points of the field of view and momentarily picking out individual details of the object under observation. In this way the object is decomposed into a number of elements subjected to successive viewing. In doing so, the individual elements of the object act as light filters which, to one degree or another, change the integral luminous flux passing through the field of view. Then, with the aid of a condenser, the luminous flux is directed onto a photoelement, the output of which is fed to an ordinary television receiving device. Thus an enlarged shadow image of the object is reproduced on the screen of the television tube. The degree of magnification in this case depends on the ratio of the sizes of the television screen and the field of view, while the resolving power is determined by the number of lines into which the image is divided, i.e., ultimately, by the area of the transverse section of the light beam in the plane of the object. To obtain the thinnest and most mobile light beam possible (with a diameter of the transverse section in the plane of the object on the order of \(1\mu\)), the authors used as the light source the glow of the screen of a cathode oscillograph (the diameter of the light spot about \(1\ \mathrm{mm}\)), projecting it into the plane—

...brightness of the object with the aid of an ordinary microscope objective (in this case less than one millionth of the light emitted by the tube is concentrated on the object).

The published information gives no basis for judging the optical parameters of such a device. The authors assert that they may be much higher than those of an ordinary microscope, but they give no arguments in support of this assertion. Moreover, comparison with an ordinary microscope would perhaps be premature, since what is involved is the first models, still not possessing optimal conditions. Nevertheless, even in the form described, the device has already made it possible to approach in a new way the solution of one of the important problems of microscopy. We have in mind the problem of counting the number of particles in the field of view and determining them by size². Problems of this kind constantly arise in connection with many applications, for example, in powder metallurgy, the confectionery industry, biology (counting blood cells), meteorology (water droplets in clouds and fogs), labor protection (dustiness), etc. The visual methods usually employed are extremely laborious and are not free of individual errors. At the same time, the problem of counting particles here differs substantially from the usual one, since the particles do not follow one another in a definite sequence, but are scattered chaotically over the field of view.

The solution of this problem by means of a beam successively scanning the field of view is quite obvious. It is achieved by the simple replacement of the television receiver at the output of the photocell by an ordinary counter of electric pulses. However, in such a primitive form the device proves to be of little use because of the counting errors inherent in it. In fact, if the diameter of the light beam exceeds the diameter of the particles, then there is the possibility of failing to distinguish particles that are closely adjacent to one another. Conversely, if the beam diameter is smaller than the particle diameter, then the same particles will be counted several times, since the beam will cut across them again and again as it passes along neighboring lines. To avoid these errors, a whole series of various improvements has been proposed² (a doubled beam, variation of the beam diameter, artificial exclusion of repeated counts in neighboring lines, etc.), but, apparently, a complete solution of the problem has not yet been achieved. As an example we shall give a brief description of one of the methods, using a doubled beam³. In the path of the beam between the objective (condenser) and the object a birefringent crystalline plate is placed; the parameters of the plate are chosen so that the two mutually perpendicularly polarized beams thus formed fall on neighboring lines. Behind the object there is placed, for example, a pile that separates both polarized components and directs them onto two photocells connected opposite one another. If the object blocks only the first component, then the signal from the photocell is fed to the counter and is counted by the latter. If, however, the object blocks both components (two neighboring lines), then the signals from the two photocells compensate one another and the counter does not operate. Finally, if the object blocks only the second component, then the signal from the photocell reaches the counter, but is not registered by it because of the opposite sign of the signal. Therefore, if a particle occupies several lines, it is counted only once, namely when only the first component passes through it. The authors indicate that in this way they achieved a counting rate of the order of a million particles per second, with the error amounting to about 1%. However, while ensuring speed and reliability of counting, such a device does not make it possible to determine the size-distribution function of the particles.

G. R.

References

1. F. Roberts and J. Z. Young, Nature 167, 231 (1951).
2. N. H. Walton, Nature 169, 518 (1952).
3. F. Roberts and J. Z. Young, Nature 169, 963 (1952).