The pattern of the recorded lines is determined both by the kind of atoms in luminous gas that emit and absorb radiant energy and by the atoms' energy levels and velocities. A unique spectrographic pattern is associated with each kind of atom. The pattern is influenced in minor but significant ways by the atom s environment.

In terms of the frequency at which electromagnetic waves vibrate, the solar spectrum spans more than 30 octaves, including the single octave of visible light. Spectroscopists have measured, identified and tabulated some three million spectral lines. With this information and an instrument for recording ultraviolet spectra amateurs can identify the metals in ores, learn the relative velocities of particles in glowing gases and vapors, measure colors precisely and perform many related experiments. An inexpensive ultraviolet spectroscope has been built at home by Roger Hayward, the retired architect and optical designer who illustrates this department. Hayward writes:

"Figuratively, at least, I have spent some of my most delightful hobby hours chasing rainbows. It all grew out of a pre-high-school desire to build an instrument that was good enough to show separately the two closely spaced yellow lines that appear in a spectroscope when a sheet of asbestos paper soaked in salt brine is held in the flame of a Bunsen burner. The instrument I planned was to consist of a slit through which light from the Bunsen burner would be admitted to a lens. The lens would make the diverging rays parallel and direct them into a glass prism that, like a drop of rain, would split the light into its constituent colors. The emerging colors would enter a second lens, through which I hoped to see a band of colored lines, each an image of the slit. In particular I hoped that the yellow emission lines of sodium in the salt would appear.

"A glass prism that dangled from the chandelier in our dining room was liberated to do the splitting of colors. A lens from a magnifying glass was appropriated for making the light rays parallel. A lens from another magnifying glass was used as the telescope lens for examining the spectrum. The slit was made from a tin can.

"When these parts had been assembled in a wooden box, I fired up the burner and had a look. I could make out a single yellow line with fuzzy edges, but no amount of adjustment made it split into the sodium pair. Finally I gave up in disgust, partly at my own ineptness and partly because of the inadequacy of the illustrations in my father's old high school physics text.

"The memory of that frustrating experience goaded me for nearly 30 years. Eventually I decided to try again and also to practice the art of making informative illustrations. In the meantime my interest in spectroscopy had expanded. I learned that the spectroscopic information of most interest lies outside the visible spectrum, much of it in the ultraviolet region to which glass is relatively opaque.

"Two basic schemes have been devised for dispersing ultraviolet radiation with minimum loss. In one of them the rays are reflected from a polished metallic surface, such as aluminum, that has been ruled with closely spaced parallel lines. This forms a diffraction grating, which causes reflected waves to interfere with one another selectively so that the angle at which they emerge from the rulings increases in proportion to their length. At the time I made my spectrograph diffraction gratings were not available at a price I could afford.

"The second scheme is to use a prism cut from a substance that is transparent to ultraviolet radiation. Quartz is such a substance. Fused quartz would work, but at that time it could not be made in large blocks of the necessary optical quality. The resolution of a spectrograph-the ability of the instrument to separate closely spaced spectral lines- increases with the size of the prism. I wanted a prism with at least two inches on a face, which meant I had to cut the prism from a natural crystal of quartz.

"This requirement introduced another complication. Crystalline quartz is an optically active substance. It polarizes light, meaning that it rotates the plane in which the light waves vibrate. It is also birefringent: unless the light travels along a path that parallels the optic axis of the crystal an entering ray is split. It emerges from the other side of the crystal as two rays that produce a double image.

"The effect of birefringence can be minimized by cutting the prism from the quartz in a direction such that the rays travel parallel to the optic axis of the crystal. The effect of polarization can be minimized by placing behind the crystal a mirror that reflects the rays back through the quartz. Waves that are rotated in one direction during their forward transit through the prism are untwisted by the same amount during the return trip.

"In 1938 I bought a two-pound quartz crystal from the Brazilian Importing Company in New York for $5. (No doubt it would cost much more today.) The facets came to a point at one end, which indicated that the optic axis was parallel to the length of the piece.

"Quartz is difficult to cut, so I took the crystal to a shop that had a diamond saw. The technician simply held the crystal by hand and ran it against the blade, using kerosene as a coolant. After cutting out a block roughly in the form of an equilateral triangle we examined the piece critically. The corners were not filled out. It was clear that if we took thin slices off the sides, the prism would present a better appearance. So the technician took off the two slices by hand. As I recall I paid $10 for the job.

"We had just guessed at the angle of the apex. It turned out to be approximately 63 degrees. Prisms of commercial spectrographs are usually made with 60-degree angles, at which the loss of light by reflection from the faces is minimal. The extra loss at 63 degrees is trivial.

"The faces of the roughly cut blank were ground flat and smooth against a surface made of hexagonal ceramic tiles stuck to an eight-inch steel disk with hot pitch. For the grinding compound I used a slurry of Carborundum grains in water. The rough grinding was begun with No. 80 grit and continued with successively finer grits through No. 600, as explained in Amateur Telescope Making: Book One, by Albert G. Ingalls, which is published by SCIENTIFIC AMERICAN. I held the prism by hand and ground it back and forth across the ceramic surface, much as amateurs make telescope mirrors. To polish the faces I covered the ceramic blocks with pitch and used a slurry of rouge in water. The prism was supported by a wooden block during polishing.

"I cannot remember why I elected to grind and polish the prism by this unorthodox procedure instead of mounting it conventionally in a plastic disk along with pieces of quartz to fill out the circular array. I suppose I was just impatient to see the finished product. Anyway, I knew that if the faces did not come out flat, I could always mount the piece conventionally and make the necessary corrections. One of the secrets of my success in handwork of this kind is that I have cold hands. The heat transmitted to the piece by my fingers did not distort the surfaces. People with warm, fleshy hands usually have difficulty getting good optical figures when they attempt handwork, but skinny hands like mine make out well.

"Having finished the prism, I made four mirrors. One had a flat surface for reflecting rays through the prism, one was spherical and two were cylindrical. The cylindrical mirrors collect light from the source and direct it through the slit of the instrument. The spherical mirror receives converging rays from the slit, transforms them into parallel rays, directs the parallel rays through the prism, receives the dispersed rays and reflects them to a focus on a strip of photographic film located immediately above the slit [see Figure 1].

"The glasses were ground, polished and figured by techniques somewhat similar to those used for making telescope mirrors. A commercial shop applied reflecting films of aluminum to the polished glasses. Incidentally, disks of cast iron that are used as weights on barbells can be made into handy flat surfaces 011 which to grind mirrors. They are generally available from dealers in sports accessories. The hole in the center is not much bother. I had the face of a disk ground flat on a Blanchard grinder by a local machine shop.

"The instrument is housed in a lightproof pine box with Masonite ends. A tube of 16-gauge steel 1 1/2 inches in diameter forms the backbone of the mechanism for supporting and adjusting the position of the optical parts. One end of the tube slides into a bronze casting, locked by clamps, that forms a table for supporting the spherical mirror.

"The table has a sliding top for altering the distance between the mirror and the other optical elements of the instrument [see Figure 2]. A threaded hole in a lug that extends downward on one side of the sliding tabletop engages a screw that can be turned by a knob on the front of the instrument. This control adjusts the focus of the spectral lines on the film.

"Two other knobs and similar screw mechanisms operate a pair of rocker arms that bear against the rear edges of the spherical mirror, which is supported in an annular cell by spring clips. By manipulating the screws I can adjust the angular position of the mirror The bottom of the casting includes a foot that is fastened to the bottom of the box with a screw.

"The design of the mechanism for mounting and positioning the prism and the flat mirror is somewhat more elaborate. The prism must be rotated to bring the various regions of the spectrum into view and to center desired spectral lines on the photographic film. To accomplish the rotation certain optical properties of the prism must be taken into account in the design of the mounting fixture.

"A ray of light that traverses a prism is bent where it enters the prism and bent still more in the same direction when it emerges into the air. The magnitude of the total angle through which a ray is bent depends on the nature of the crystal and also on the angle between the ray and the face of the prism through which it enters. The ray is bent least when the angles between the ray and the front and rear faces of the prism are equal. Then the path of the ray is symmetrical with respect to the faces of the prism. When the prism is so positioned, it is said to operate at minimum deviation.

"Short waves are bent progressively more than long ones. The waves are dispersed according to their length. Some transparent substances disperse waves of differing lengths more than other substances do. It turns out that the dispersion of quartz is relatively low. For this reason I decided to use a 60-degree prism instead of one of 30 degrees' as is customary in instruments of this type.

"This decision made it imperative that I operate the prism at minimum deviation because, as mentioned, quartz is birefringent. If a quartz prism is operated at other than minimum deviation, the v oppositely polarized rays do not cancel completely, and two images appear in the focal plane. Of course, not all rays can be made to pass through the prism at minimum deviation because some waves are bent more than others, depending on their length. Even so, the prism can be placed so that the rays of most interest traverse it on a path that closely approaches minimum deviation.

"To achieve this condition it is essential when scanning the spectrum to rotate the flat mirror at exactly twice the rate of the prism. A system of differential gears would do the trick, but I was worried about the smoothness of gear motion. Gear teeth tend to introduce periodic accelerations.

"I decided to generate the desired motion by rotating a disk on a set of three conical rollers linked with yokes. One revolution of the disk would cause a half-revolution of the roller assembly [see Figure 3]. The movement of the rollers is restrained by a pair of circular grooves. One groove is in the bottom of the disk and the other is in a circular baseplate.

"The disk, which is a flat worm gear, carries the mirror. A hole through the gear admits the foot of a prism table to the roller assembly, to which the table is attached by screws. The width of the hole restricts the rotation of the disk to 30 degrees, but in use the table turns only 26 degrees. Two of the rollers are yoked to the prism table. The third roller acts as an idler. A worm engages the teeth of the gear and is rotated by a shaft that terminates in a calibrated dial on the front of the instrument.

"Most of the metal parts are castings of red brass. They could have been made from strip, plate and rod stock of other metals. I undertook the project just after I had written a chapter on molding and casting for the book Procedures in Experimental Physics, by John Strong and others. It therefore seemed to me that I should design the apparatus to be built with castings made from my own patterns. There are 18 patterns all told, one for a wavelength drum that I later discarded. The bill from the foundry, as I recall, amounted to about $20. I sometimes wonder what the same job would cost today.

"The control panel consists of a brass plate that covers an opening in the front of the cabinet. It fits loosely in a groove in the edge of the Masonite front. In turn the plate supports the shaft bearings and control knobs, the slit assembly and the film carrier. When the spectrograph was completed, I learned that the spherical or autocollimating mirror could have been equipped with less fancy controls. The position of the mirror has not been altered since its initial adjustment more than 25 years ago!

"Ports for the slit assembly and the film carrier are identical. They are separated by a narrow dovetail strip. The optic axis of the system passes through the center of the strip. For this reason the slit and the film carrier can be interchanged.

"Access to the optic axis can be had by removing the dovetail strip, which is attached to the panel by four machine screws. By removing the strip and the prism it is possible to direct light along the optic axis by means of a half-silvered mirror. One can then look through the half-silvered mirror and adjust the positions of the flat and spherical mirrors so that light is returned along the optic axis. The mirrors are then in proper orientation.

"The dial that rotates the flat mirror is calibrated in minutes of angular arc. A setscrew locks it to its shaft in the zero position when the flat and spherical mirrors have been aligned. Each full turn of the minutes dial advances an associated degrees dial 1/30 of a turn.

"The motion is transmitted from the minutes dial to the degrees dial through a Geneva movement and a pair of 1: 3 reduction gears. This feature was an afterthought. I could have used a Veeder counter to record degrees, of course, but I had fun filing out the Geneva wheel by hand. I undertook the job mostly to see if I could do it. The cup-shaped degrees dial and the large flat minutes dial are made of Lucite.

"The film carrier was designed to take pieces of cut roll film, Type No. 120, 2 1/4 inches wide and 2 1/3 inches long. Strips 3/4 inch wide can also be used. It is easier to make the film carrier lightproof if the dark slide that protects the emulsion when the carrier is removed from the instrument cannot be fully removed. I fastened a limiting stop to the inner end of the slide; it prevents the slide's removal. The slide was made from sheet brass 1/32 inch thick. The frame of the carrier is milled from stock brass strip one inch wide and 1/2 inch thick. The rim of the cover, which includes a flat spring for pressing the film in place, is also made of sheet brass and is silver-soldered to the top plate. It could be made of other metal.

"The body of the slit assembly was milled from the same stock used for the frame of the film carrier. The width of the slit should be adjustable so that the width of the spectral lines and the amount of light that enters the instrument can be controlled. The center of the slit should not be displaced when the width is altered, because the spectral lines would be similarly displaced.

"The slit of my instrument is formed by two strips of 16-gauge stainless steel that slide in V grooves of the body. Springs that bear against the outer ends of the strips move the strips together. A pair of eccentrics that bear against the inner ends of a pair of pushrods, which are coupled in turn to the ends of the strips, force the strips apart [see Figure 5]. A mechanism designed to force the strips together, instead of separating them, might damage the inner ends, which are honed to sharp edges.

"The frame of the slit assembly also supports a pair of strips made of phosphor bronze, the inner ends of which are cut at an angle of 45 degrees. These strips can be adjusted to admit rays through any part of the slit. In one position, for example, they may admit rays from an unknown source to the upper half of a piece of film. After the unknown spectrum has been recorded the position of the strips can be readjusted for recording a known spectrum on the lower half of the film as a comparison. The position of the diagonal strips can be similarly changed for registering a series of exposures of various time intervals.

"The most precise and convenient method of determining the wavelength of an unknown spectral line is to compare its position with that of a line of known wavelength. The spectrum of iron, which is particularly rich in lines that have been accurately measured and tabulated, is used routinely as the comparison spectrum. I generate emission from iron by means of a spark gap. "The gap consists of a pair of cold-rolled steel electrodes 1/4 inch in diameter; they are sharpened to edges 1/32 inch wide and 1/4 inch long. The rods are supported in alignment by an electrically insulating fixture of plastic so that the inner edges are parallel and spaced 1/32 inch apart. Alternating current is applied to the electrodes by the smallest available neon-sign transformer. The one I use develops about 2,000 volts and perhaps 20 milliamperes..

"I connect a high-voltage capacitor across the gap. I made the capacitor by sandwiching sheets of aluminum foil between glass plates and connecting alternate sheets to the respective electrodes of the spark gap. Odd-numbered foils were connected to one electrode, even-numbered foils to the other.

"The glass plates are about one millimeter thick and of the quality used for supporting photographic emulsion. They are five inches long and 2 1/4 inches wide. The sheets of foil are 4 1/4 inches long and 1 1/2 inches wide.

"A 350-micromicrofarad capacitor of the high-voltage type used in amateur radio transmitters would serve nicely. Without the capacitor in the circuit the spark is faint and bluish and radiates many spectral lines that arise from gases and chemical radicals in the atmosphere. With the capacitor it becomes a brilliant bluish white and the light is not contaminated by the atmosphere.

"For operation the instrument requires a final adjustment. The prism must be mounted on its table and oriented to the position of minimum deviation. Two accessories are needed: a source of monochromatic light and an eyepiece that includes a pair of cross hairs. For the source I use a mercury lamp; a sodium flame could be substituted. The eyepiece is trained on the center of the field normally occupied by the photographic film. The mechanism for adjusting the eyepiece is a mounting plate that slides into the grooves that hold the film carrier.

"Light from the mercury lamp is focused on the slit. With the prism table stationary I rotate the prism by hand, thereby moving the spectral line at a wavelength of 5,461 angstrom units (the most brilliant green line of mercury) into the field of view. As the prism is rotated still more in the same direction the green line advances at a constantly decreasing rate and finally comes to rest. At this point the prism is set for minimum deviation. If the prism is rotated still more in the same direction, the green line resumes its movement but in the reverse direction. The prism is lightly clamped to its table in the position of minimum deviation by means of a knurled screw.

"To record the iron spectrum I replace the eyepiece with the loaded film carrier, open the slit to about .004 inch and flood it with light from the spark gap. Th time of exposure must be determine experimentally. Several exposure intervals can be tried by periodically advancing one of the two diagonal jaws to increase the length of the spectral lines b increments. A second spectrum, such a that of mercury, can then be recorded adjacent to the spectrum of iron. Th brilliant green line, which is known be located at 5,461 angstroms, is easily identified on the developed film and can serve as a starting point for identifying the iron lines by reference to published data.

"The spectral lines of iron are numerous and closely spaced. Their positions are most easily measured by means of a traveling microscope, which consists of a mechanical carriage with a calibrated micrometer screw for transporting a low-power microscope across the recorded spectrum at precisely measured intervals of length. A traveling microscope I made was described in this department in August, 1954. The accompanying photograph [below] shows a small length of the ultraviolet spectrum of iron that was made with my spectrograph and analyzed with the traveling microscope."

Bibliography

FUNDAMENTALS OF OPTICS. Francis A. Jenkins and Harvey E. White. McGraw-Hill Book Company, Inc., 1950.

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.