Time and money would have been saved if I had owned a spectrophotometer, which is an instrument that measures colors and mixtures of colors in terms of the wavelength of light transmitted by the specimen and also records the intensity of the colors in terms of the percentage of light that is transmitted Until a few years ago spectrophotometers cost more than I could afford; those of the highest performance still do. The advent of inexpensive electronic and optical parts, however, has made it possible for anyone who is reasonably handy to build a serviceable spectrophotometer at home. One simple design that can be assembled for less than $75 is described by R. C. Dennison of Westmont, N.J. He writes:

"The spectrophotometer can be one of the most useful instruments in an amateur's shop, particularly for the analysis of chemicals. I use mine for determining properties as diverse as the color of glasses and plastics, the transmission of light by neutral-density filters and semisilvered mirrors, the percentage of chlorine and other substances in water, the kind of metals that may be present in specimens of rock and the composition of alloys. Essentially the instrument disperses light that is transmitted by the specimen into the rainbow hues of the visible spectrum and measures the intensity of the emerging colors one by one.

"The physical scheme of the instrument is simple. Diverging rays from an incandescent lamp pass through a thin mechanical slit, known as the entrance slit, and through a collimating lens that makes the rays parallel [see Figure 1]. The parallel beam passes through the specimen, where certain colors may be fully or partially absorbed, depending on the nature of the specimen. Colors that remain in the beam enter a prism that disperses them into the orderly array of the spectrum.

"The spectrum is focused by a second lens, called the telescope lens, as a band of rainbow colors on an opaque white screen that is perforated with the exit slit. This slit transmits one narrow band of color or another, depending on its position in relation to the spectrum. The transmitted rays fall on a photoelectric cell and induce in it an electric current that varies in magnitude with the intensity of the colored light. The current is measured by a microammeter.

"The amount of light that reaches the specimen is controlled by the position of a wedge-shaped diaphragm that in effect determines the length of the entrance slit. The photocell, the exit slit, the telescope lens and the prism are assembled on a carriage of sheet metal that is attached at the prism end to a vertical shaft. By turning the shaft the operator can move the exit slit across the spectrum in order to select a desired color. A dial fixed to the shaft indicates the position of the slit in terms of the wavelength of light. The microammeter is calibrated to indicate the percentage of light transmitted by the specimen. All parts are housed in a cabinet equipped with a light-tight lid for shielding the photocell from room light when measurements are made.

"The parts are assembled on the bottom and one side of a steel radio chassis (Bud CB-643) that is 17 inches long, 13 inches wide and four inches deep. A second chassis of the same size is hinged to the first as a dust cover and light shield. The lamphouse is made of cookie-sheet aluminum and is 2 3/4 inches long, 2 1/8 inches wide and 2 1/2 inches high. A hole 5/8 inch in diameter in one wall of the box is partly closed by a pair of double-edged razor blades spaced .0045 inch apart to form the vertical entrance slit. A socket that fits a General Electric No. 93 incandescent lamp is mounted on the wall opposite the slit. The lamp is installed with its filament in the vertical plane. The housing is ventilated by a one-inch hole in the chassis and several quarter-inch holes in the top. A baffle of sheet aluminum inside the lamphouse near the top prevents the escape of stray light.

"The collimating lens is mounted on a rectangle of sheet aluminum 2 3/8 inches square that contains in the center an aperture 5/8 inch in diameter. Two 1/8inch tabs cut from the upper corners of the rectangle are bent over as supports for the upper edge of the lens. The lower edge is supported by a tab of aluminum attached to the rectangle by a machine screw.

"The assembly is mounted on a bracket of sheet aluminum 2 1/2 inches wide and 2 3/4 inches high that is perforated with a centered hole one inch wide. The bracket is attached to the chassis by screws passing through slots in the foot that enable the collimating lens to be moved toward or away from the lamphouse when the lens is focused. The optical axis of all elements of the optical train is 1 1/2 inches above the base.

"The carriage, which includes the photocell, the exit slit, the telescope lens and the prism, is made of sheet brass 1/16 inch thick, 2 1/8 inches wide and nine inches long. One edge is bent up 5/16 inch to provide stiffness. The photocell housing is similar to the one for the lamp but consists of a Bud CU-3000 A Minibox. One wall contains the vertical exit slit, made with razor blades spaced .003 inch apart. They face the photocell. The photocell (RCA Type 7117) was designed for automatically dimming automobile headlights. I bought one at an automobile junkyard for a fraction of the list price.

"The cover of the photocell housing consists of a rectangle of Bakelite 1/8 inch thick to which an Amphenol socket (Type 77MIP11) is attached by screws. Nine 33-kilohm resistors that supply voltage to the photocell are soldered directly to the lugs of the socket. A thin sheet of Bakelite, supported by standoff pillars, is mounted over the lug side of the socket to prevent accidental contact with the high voltage.

"The inverted photocell projects downward into the box. The socket must be oriented so that the photocathode faces the exit slit. (Point the keyway of the socket toward the exit slit.) The completed assembly is attached to the carriage arm by machine screws in the position illustrated. Apply a strip of flat white paint to the razor blades and the front of the housing.

"A 1/4-inch vertical shaft is attached by machine screws to the end of the carriage arm opposite the photocell by means of a flange in the form of a brass gear that happened to be on hand [see Figure 2]. Any equivalent flange would serve as well. The distant end of the carriage is supported by a wheel that rides on top of the chassis. The wheel was made by soldering a short hollow rivet into a 3/4-inch washer. The axle of the wheel consists of a machine screw that enters the rear of the photocell housing near the base on the center line of the carriage.

"The vertical shaft is supported and driven by a modified worm gear from a surplus gunsight. An equivalent mechanism could be constructed with a Millen Type 10000A worm-gear drive of the kind designed for rotating the tuning capacitors of radios. This unit is available from dealers in radio supplies. It is made in gear ratios of 16 to one and 48 to one. A gear of the latter ratio should be used.

"The horizontal shaft of the worm gear extends through the front of the chassis. It should be equipped with a planetary dial that has a drive ratio of eight to one (Lafayette Radio Electronics stock No. 99 H 6029). When this dial is turned through 240 angular degrees, the carriage sweeps the exit slit completely across the spectrum. The dial is designed for only 180 degrees of rotation, but the plastic stops that limit the rotation can be sawed off by disassembling the unit.

"A small table of sheet metal was improvised to support the prism on the optical axis of the instrument above the end of the vertical shaft. The prism (Edmund Scientific Co. stock No. 30,143) is held in place on the table by a clip made of spring brass. The telescope lens is attached to the carriage arm by a fixture similar to the one that supports the collimating lens.

"The amount of light that reaches the specimen is controlled by a triangular diaphragm cut into a plate of brass that moves across the path of the rays emerging from the entrance slit. One edge of the plate is bent at right angles to form a foot 1/8 inch wide. This foot is soldered to a rectangular brass bar, of about the same width and thickness, six inches long. The bar moves in slotted guides made from two small blocks of Bakelite that are attached to the chassis by machine screws. The bar is driven by a mechanical linkage consisting of a Millen No. 10,012 right-angle drive, a bell crank and a slotted brass arm.

"The right-angle drive is mounted on the inner face of the front wall of the chassis. An arm of sheet brass, soldered to the outer end of the rod that supports the triangular aperture, projects down through a slot in the chassis. A slot about l/2 inch long in the arm engages a bell crank that is driven by the right-angle drive. The bell crank, when turned by the right-angle drive, advances or retards the triangular aperture. By advancing or retarding the aperture the microammeter can be adjusted to full-scale deflection without altering the current that is normally present in the photocell when the photocathode is in darkness.

"Specimens are inserted in the light path at a point between the collimating lens and the prism. A fixture improvised from sheet metal is attached to the chassis at this point for supporting a rectangular glass box that is called either a cuvette or an absorption cell. Fluid specimens are placed in the cuvette for measurement.

"Cuvettes are priced at $15 and up because they must be made of reasonably flat, well-annealed glass. Glass of the quality used for photographic plates is adequate and can be bought from distributors of photographic supplies. The inside dimensions of the cuvette should be about 75 millimeters, 35 millimeters and 10 millimeters. The ends, sides and bottom piece can be cut with a glass cutter of the wheel type and assembled with epoxy cement. Solid specimens such as colored glasses, filters and semitransparent mirrors are inserted for measurement in the position normally occupied by the cuvette.

"I followed conventional techniques when building the electronic portion of the instrument. Small components were mounted on stiff sheets of perforated plastic known as Vector board. The transistors specified in the accompanying schematic illustration [Figure 3] were used because they were on hand; they do not necessarily represent either the best or the least costly design. The 2N1702 transistor is mounted on the chassis, which acts as a heat sink, and is insulated from the chassis by a thin mica washer supplied by the manufacturer. The lN1204RA diodes are mounted directly on the chassis, which again acts as the heat sink. (Incidentally, the anodes of diodes that bear the suffix 'RA' connect to the mounting stud and do not need to be insulated from the chassis.)

"The conventional scale of the microammeter was replaced by one calibrated in intervals from O to 100 for indicating the transmission of light in percent. It was also calibrated in units of density, according to the relation: density equals the logarithm, to the base 10, of the ratio 100 divided by the transmission in percent. For example, at 50 percent transmission the density of the specimen is equal to log₁₀ 100/50, or .3.

"In order to align and focus the optical system and calibrate the dial that controls the position of the exit slit I first removed the prism and the collimating lens of the fully assembled instrument. The lamphouse was then positioned to center the wedge of light that emerges from the entrance slit on the optical axis of the instrument. The lamphouse was locked in this position by its mounting screws.

"The simple plano-convex collimating lens (1 3/4 inches in diameter with a focal length of three inches) is mounted with its plane side toward the lamphouse at a point that causes the diverging rays from the slit to become parallel after they have passed through the lens. To locate this position I first focused a pair of binoculars on an object about a mile away. The binocular was then positioned so that the objective lens of one half of the instrument was on the optical axis of the spectrophotometer and faced the collimating lens, looking toward the entrance slit. (A sheet of white paper can be placed between the lamp and the slit to reduce the intensity of the light.) The position of the collimating lens was now adjusted until a sharp image of the slit appeared in the binocular. Any small, low-power telescope can be substituted for the binocular. When the collimating lens was focused, it was locked to the chassis by tightening the mounting screws.

"The telescope lens is assembled in its holder with its plane side facing the exit slit. Again, with a small telescope focused on infinity, adjust the position of the telescope lens on the carriage until a sharp image of the exit slit appears. (Light the slit in front with the beam of a 35-millimeter projector.) Lock the telescope lens in this position.

"Release the setscrew on the rear hub of the dial that drives the carriage. Mount the prism on its table, light the incandescent lamp and slowly rotate the prism back and forth until the spectrum appears on the white surface of the exit slit. Turn the prism back and forth on its table and observe that at one position the angular deflection of the spectrum is at a minimum. The prism is now set at the angle of minimum deviation. Turn the shaft of the worm-gear drive and shift the position of the prism by trial and error until at minimum deviation the center of the yellow band of the spectrum falls on the exit slit. Rotate the shaft to move the slit to the red end of the spectrum, position the dial at the limit of its excursion and lock it to the shaft.

"The dial must now be calibrated to indicate wavelength. If possible, borrow a set of narrow-band interference filters that transmit light of known wavelength. With the exit slit at the red end of the spectrum, insert the interference filter of longest wavelength in the cuvette holder at right angles to the light beam and rotate the dial for maximum photocurrent.

"Record the arbitrary indication of the dial and repeat the procedure for each filter. On linear graph paper plot the arbitrary indications of the dial against the corresponding wavelengths. From these data make a scale for the dial calibrated in millimicrons. The scale will be crowded at the red end.

"If interference filters are not available, the instrument can be calibrated with reasonable accuracy by means of didymium glass. A didymium filter accompanied by a curve of spectral transmittance can be bought from the Arthur H. Thomas Company, P.O. Box 779, Philadelphia, Pa. 19105. The item is listed as didymium filter No. 9104-N20 and costs less than $2. The transmittance curve supplied with the filter displays nine dips and peaks between 400 and 700 millimicrons. Mount the filter and make a series of readings in which the dips and peaks are correlated with arbitrary dial readings. Convert the arbitrary readings to wavelengths by referring to the calibration curve supplied with the filter and make a corresponding wavelength scale for the dial.

"When the photocurrent of the spectrophotometer is plotted against wavelength, the resulting graph takes the form of a bell-shaped curve that peaks at approximately 550 millimicrons and drops to about 5 percent of the maximum reading at 380 and 660 millimicrons. The graph depicts the intrinsic response (1) of the instrument.

"Intrinsic response must be known before an unknown color can be determined. For maximum accuracy of measurement the intrinsic response should be redetermined prior to measuring each unknown specimen. For example, to measure the spectral response of a piece of colored glass, find the wavelength at which the photocurrent is maximum. Remove the specimen and adjust the intensity of the light (by altering the position of the wedge-shaped diaphragm) until the pointer of the meter swings to full scale (the 100 percent indication). Turn the wavelength dial to its limit at the red end of the spectrum. Replace the specimen. Record the meter indication at this wavelength and designate the response R. Remove the specimen and designate the resulting response 1. Make similar pairs of readings at intervals of five or 10 millimicrons across the spectrum to the limit at the blue end.

"When the readings of R and I drop below 20 percent of full-scale meter deflection, as they will doubtless do at the red and blue ends of the spectrum, the instrument will lose some accuracy. This loss can be compensated for by increasing the sensitivity of the instrument. The sensitivity can be increased by turning up the sensitivity control until the limit of usable gain is reached.

"The transmittance (T) of the specimen is equal at each interval of wavelength to R divided by 1. Calculate the transmittance at each interval of wavelength and from the tabulated computations prepare a graph of the spectral response. The response of unknown solutions is also measured. The intrinsic response of solutions is determined by replacing at each interval of wavelength the cuvette containing the specimen with an identical cuvette that contains only solvent.

"Experience in the use of the instrument and confidence in the reliability of the measurements can be gained easily by making graphs of the spectral responses of Wratten filters and comparing the results with graphs supplied by the manufacturer. These inexpensive filters and their graphs can be bought from dealers in photographic supplies. Simple experiments for beginners also include the measurement of various dyes, food colors and other colored solutions [see The Amateur Scientist, SCIENTIFIC AMERICAN; February, 1965].

"A more advanced experiment that demonstrates the usefulness of the spectrophotometer as an analytical instrurnent consists of the analysis of steel for the presence of manganese. The procedure is based on the magenta color that appears when a solution of manganous salts is oxidized. The amount of manganese in the steel is calculated by comparing the absorbency of the colored solution with the absorbency of a standard solution of potassium permanganate. (Absorbency, like density, is equal to the logarithm, base 10, of 100 divided by the transmission in percent.)

"To prepare the standard solution dissolve 72 milligrams of potassium permanganate in water and dilute the solution to 250 milliliters. In this step and all following procedures use only distilled water and reagent-grade chemicals. Weights and volumes must be accurately determined.

"Transfer four milliliters of the standard solution to a clean container and dilute to 40 milliliters. As thus diluted the solution contains 10 milligrams of manganese per liter. Transfer the diluted solution to a cuvette and make a graph of the spectral response [see Figure 4]. Maximum absorbency occurs at 540 millimicrons. My instrument indicated a transmission at this wavelength of 40.5 percent, from which the absorbency was calculated to be .393. (100/40.5 = 2.47. Log₁₀ 2.47 = .393.) From the stock solution make similar dilutions that contain five milligrams and 20 milligrams of manganese per liter and tabulate the absorbencies. A calibration curve for use in the subsequent analysis can now be drawn by plotting the absorbencies of these three measurements against concentration in milligrams [see Figure 5].

"Saw a small piece of steel, weighing about 200 milligrams, from a bar or rod. Place the sample in a 100-milliliter volumetric flask and add five milliliters of water and five milliliters of nitric acid. Warm the solution until the sample has dissolved. While the solution is warm add sodium bismuthate until a slight excess remains. Dilute to 100 milliliters.

"Transfer 10 milliliters of this solution to a clean vessel and dilute to 100 milliliters. Transfer a specimen of the latter solution to a cuvette. Measure the transmittance and calculate the absorbency. Determine the concentration of manganese by referring to the calibration chart. Assume, for example, that the concentration turns out to be 1.5 milligrams of manganese per liter. Before the dilution the concentration was 15 milligrams per liter. The volume of the original (undiluted) solution was 100 milliliters. Therefore it contained 1.5 milligrams of manganese. The specimen of steel weighed 200 milligrams. Hence the steel contains 1.5/200 X 100, or .75, percent manganese.

"Another interesting experiment involves a test for cobalt. To make it you will need the following materials: (1) a few grams of sodium pyrophosphate, which can be made by fusing disodium hydrogen phosphate in a crucible; (2) a 60 percent solution of ammonium thiocyanate, made by dissolving 30 grams of the salt in water and diluting to 50 milliliters; (3) acetone; (4) a standard cobalt solution containing 100 milligrams of cobalt per liter, made by dissolving 49.36 milligrams of cobalt nitrate hexahydrate in water and diluting to 100 milliliters, and (5) another cobalt salt, such as cobalt chloride.

"Transfer 10 milliliters of the standard cobalt solution to a graduated cylinder and add 1/2 gram of sodium pyrophosphate. The sodium pyrophosphate prevents any iron that may be present from discoloring the solution. Add 2.5 milliliters of 60 percent ammonium thiocyanate and mix. Dilute to 25 milliliters with acetone and mix. The clear solution will turn blue.

"Place a specimen of the colored solution in the cuvette, measure the transmission and plot the spectral response. Determine the absorbency at the wavelength of maximum absorption and, by serial dilution and subsequent measurement, tabulate data for plotting a calibration graph, as in the analysis of steel. Check the results by measuring the percentage of cobalt in a solution of cobalt chloride by weight.

"Finally, determine the amount of cobalt in an alloy, such as Aluico. Wrap a small Alnico magnet in cloth and, with a hammer and chisel, break off a few fragments. Weigh a specimen of about 200 milligrams. Dissolve the specimen in 10 milliliters of hot nitric acid. Dilute to 100 milliliters. Transfer 10 milliliters of the solution to a clean vessel and add 1/2 gram of sodium pyrophosphate and 2.5 milliliters of 60 percent ammonium thiocyanate. Mix and filter the solution. Dilute the filtrate with an equal volume of acetone. Measure the absorbency and calculate the percent (in weight) of cobalt in the specimen.

"Caution: Most of these chemicals are toxic. Acetone is highly flammable. Avoid contact with the substances. Do not inhale the fumes of reacting mixtures. Be sure to work in a well-ventilated room."

Bibliography

ANALYTICAL ABSORPTION SPECTROSCOPY: ABSORPTIMETRY AND COLORIMETRY. Edited by M. G. Mellon. John Wiley & SonS, Inc., 1950.

CHEMICAL SPECTROSCOPY. Wallace R. Prode. John Wiley and Sons, 1943.

OPTICAL METHODS OF CHEMICAL ANALYSIS. Thomas R. P. Gibb, Jr. McGraw Hill Book Company, Inc., 1942.

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.