IN THIS department in July and August, 1942, data for building a concave grating spectrograph of the Rowland circle type were given by R. L. Watrous. Another form of that type now has been described by Wm. S. von Arx in the Journal of Chemical Education (Easton, Pa.) Vol. 19, pages 407-10. As it is the policy of thus department to make available to its readers, by reprinting, matter which relatively few of them would otherwise see, the following is extracted from the longer paper. Von Arx, an amateur telescope maker and author of the chapter on Stellar Photography in "A.T.M.A.," originally contributed it to the professional journal named, not in that capacity but in his capacity as a majoring student in science at Brown University. He also drew the sketches. The extract:

There is a new design of grating spectrograph manufactured by Adam Hilger Ltd., of London, under the name "Technal." It has exceptional properties of ruggedness and simple design which are so inherent as to be preserved even when homemade. It is possible to build a modest version of the "Technal" spectrograph in about two weeks of evenings. In no case should the cost exceed $50. Half this amount should be adequate.

First, let it be made clear why the concave grating is preferable to the more familiar prism as a dispersing medium. Prisms introduce irrational dispersion-non-uniform separation of equal wavelength intervals in different spectral ranges-which makes the interpretation of spectrograms unnecessarily difficult for the beginner. Furthermore, the prismatic instrument must always contain three component parts -the collimator, prism, and camera, each of which involves at least one pair of optical surfaces. For analysis in the ultra-violet range, these parts must be made of quartz, which is very expensive. The grating spectrograph, on the other hand, not only produces linear dispersion but may contain no lenses whatever and only one spherical reflecting surface if the concave grating is employed. With these, an aluminized reflecting surface is all that is required for efficient operation in the ultraviolet. This simplicity carries a twofold advantage; it reduces the initial cost of the instrument and makes it easier to keep in adjustment. Another advantage in the use of grating dispersion is the wide range of dispersions available in the higher orders of spectra. While the intensity of these higher orders is usually considerably less than that of the first order, the high intensity of the carbon arc, which is almost invariably used for qualitative analysis of non-conducting samples, allows them to be used for more precise analysis of the complex spectra characteristic of the transition group of elements. The concave grating does possess a few disadvantages, the worst of these being astigmatism. But this can be effectively controlled either by means of properly designed slit illumination systems, or by employing the minimum astigmatism mountings, of which the "Technal" is an example. These reduce the stigmatic error to such a small figure that it becomes unimportant in the normal working ranges.

The "Technal" design has no inventor's name associated with it as yet, but J. S. Dowell, of Adam Hilger, Ltd. has described the mounting and the Hilger interpretation of its design, which he credits to Cotton and Richards.

The mechanism involves only three levers of fixed length, a pivoting hinge, and a short track along which the grating moves under control of the lever system. The arrangement of parts is shown diagrammatically in the main sketch, the grating being at A, the slit at B, and the plate at C. The line BA represents the center line of the spectrograph bed and the levers BR and AR are in length one half the radius of curvature of the grating and have the grating and slit-plate elements rigidly fixed to their ends. Where they meet at R they are pinned together so as to articulate. A track which is parallel to the bed BA is placed under A, and is long enough to allow the grating element on the radius arm AR to travel along the bed BA in the direction of the double-ended arrow for a few inches.

It is evident, since BR and AR radii of the Rowland circle of the grating, that by moving the joint R toward or away from the line of the bed A the range of wavelengths recorded on the plate C will change. Furthermore, the parts will always remain on the circumference of the Rowland circle and will therefore be in correct focal relation to each other at all times. If a light source is located at S on the line of the bed AB prolonged, the incoming light-ray must always fall fully upon the grating at A no matter what wavelength range is being photographed. If R is moved away from the bed it will be found that the plate will record the longer wavelengths or higher orders of dispersion of the grating since the central image CI of the slit is on the opposite side of the radius arm AR at an angle 2 with the bed. The angles are all equal, as is evident from the geometry of the mechanism. They usually have values ranging from 0 to perhaps 10 or 12 degrees.

The angles are varied by moving the joint R from outside the instrument by means of the lever QPT, which in pivoted on the bed of the instrument at P and connected with the joint R by a short toggle RT. The pivot is somewhat nearer T than Q in order to provide a small mechanical advantage and greater precision of motion of R. The outer end Q of the lever sweeps a scale upon which the wavelength ranges for each setting are marked. Not more than half a dozen standard settings need marked upon the scale. They may be determined by experiment. It is evident that, once adjusted, the entire optical system is completely controlled by the motion of the lever QPT and with complete assurance that all optical parts are properly oriented with respect to each other for perfect focus.

All images formed by a concave spherical surface suffer astigmatic distortion except that one image which falls exactly in line with the light source. Astigmatism increases slowly at first as one travels from this point in any direction in the focal plane, but increases rapidly beyond angular departures which are in excess of a very few degrees. Those lines nearest the slit will be most nearly stigmatic and those farther away will show increasing distortion. It is for this reason that the slit in the "Technal" mounting is placed as close to the ultra-violet end of the plate as possible. Since the far ultra-violet sensitivity of plates is always somewhat lower than that of the near ultra-violet and visible blue it is desirable that no light should be wasted in that region. Ideally, the slit should be placed in the very center of the plate, but this is difficult mechanically and would cause great inconvenience in operation.

For spectrochemical analyses of compounds containing iron or other elements of the transition group which 6ive exceedingly complex spectra, two minimum specifications must be observed regarding the dimensions of the optical system of the spectrograph: (1) dispersion of at least 16 angstroms per mm and (2) sufficient resolving power to separate completely two lines of equal intensity not more than 0.4 angstroms apart. In grating instruments this requires a focal length of about one meter, 15,000 lines to the inch, and a ruled surface at least 30 mm wide. The Central Scientific Company of Chicago sells a Wallace replica grating having these minimum specifications for a little more than ten dollars. These gratings are of fairly good quality initially but may be improved by changing the shape of the factory-made mask to be somewhat longer, thereby exposing more ruled surface and increasing the resolving power, and somewhat narrower, in order to compensate for the irregularities in the collodion replica. The precise shape of the mask must be determined by experiment. It is a long and exasperating job but eminently rewarding in the end.

The slit and the plate holder (inset sketch) are the most difficult parts of the instrument to construct and should be given double their share of careful planning and workmanship. A fixed slit of moderately narrow width is recommended. The plate must be curved to the circumference of a circle whose radius is one half the radius of curvature of the grating-the Rowland circle. The "plates" may be strips of 35mm motion picture film if it is expected that only one or two samples will be run at a time; or 8 x 10-inch cut film sliced down the middle to the standard spectrographic size 4 x 10 if more extensive work is anticipated. On these 4 x10-inch plates it is possible to record at least 16 well-separated spectra with their iron comparisons. The spectra need not be more than three millimeters high if no comparison spectra are juxtaposed but should be half again as high with comparison spectra, so that about one millimeter of the end of each line can interfinger with the comparison lines. This simple device increases the accuracy of plate measurement, since lines of nearly the same wavelength are more easily classified as coincident or separate.

The astigmatism of gratings causes the ends of the spectral lines to be ragged in appearance, the brightest line being longest. In order to conserve space and trim up the spectra, it is necessary to build an extra slit just in front of the plate in its holder. This slit is at right angles to the principal slit of the instrument and is preferably constructed to have variable width. The slit's function is simply to limit the height of the lines and make their outer terminations sharp. A pair of brass-edged foot rulers is admirable for the purpose. They may be made adjustable by coupling the two at their ends in the manner of the navigator's parallel rule.

The light-tight housing around the optical parts of the instrument may be made of sheet metal or of ply-wood screwed to a light wooden frame. The housing should have a door or hatch in it near the grating end of the case so that the grating is accessible for adjustment whenever necessary. A simple flap shutter placed in front of the slit is a convenient accessory. The plateholder motion scale will be found to be more useful if it is graduated in metric units. The metric scale of a 6" celluloid pocket ruler glued to the plate-holder track with cellulose acetate cement serves admirably.

The design of the bellows between the light-tight housing and the articulating plate-holder track presents something of a problem. Spectrograph bellows are usually of such an odd size and shape that they must be specially made. Bellows cloth costs about one dollar per yard. Bellows may be installed in the usual accordion pleat fashion built up of two layers of bellows cloth with cardboard stiffening pieces cemented between-rubber cement is recommended-or, since the span is always very short, simply cut to fit the gap and allowed to fold as it will without reinforcement inside. The latter is quite satisfactory for small instruments, the natural stiffness of the cloth itself being sufficient.

When the inner mechanism of the instrument is complete, the entire inside of the case and the parts enclosed should be painted dead black. Ordinary blackboard paint is admirably suited to this job, containing enough varnish to stick equally well to metal and wood. Since the instrument is likely to be used in a dimly lighted room, the outside of the instrument is best painted a light gray so that it is more easily seen. It is convenient to mark the rulings on the plate-holder motion scale with zinc silicate paint to which a trace of a uranium salt has been added. A dot of the same on the plate holder makes it far easier to set the plate holder without turning on bright lights in the room. Should the phosphorescent paint be too dim for ready visibility to those coming in from brighter quarters, a small argon night light may be used to excite the fluorescence temporarily.

END OF extract from von Arx's article. While this presents the main data, readers who undertake the actual job probably should obtain the full paper, which is a little longer and which also describes an accessory called the Hartmann diagram for use in photographing two spectra side by side without moving the spectrograph between exposures.

Von Arx is now at Yale University.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.