Commercial colorimeters are highly precise instruments that generally cost more than most amateurs would care to spend. An inexpensive version that can be made at home has now been specially designed for amateur construction by Sam Epstein, chief chemist of the Federated Metals Division of the American Smelting and Refining Company in Los Angeles. Epstein has tested the apparatus extensively and reports that it can analyze accurately most substances of interest to the home experimenter. He writes: "Colorimetric analysis is based on the fact that various substances characteristically absorb light more strongly in certain portions of the spectrum than in other portions and conversely transmit more light at some colors than at others. The method can be applied to the analysis of anything that is colored or that can be made to form a colored compound in solution. Solutions of copper sulfate, for example, transmit blue hues more strongly than greens, yellows and reds. Potassium permanganate solutions, on the other hand, transmit reds and blues strongly but absorb greens and yellows, with the result that the transmitted light appears reddish violet.

"Such phenomena are described by the laws of the German physicists Johann Heinrich Lambert and August Beer. Lambert's law states that for a given concentration equally thick layers of a solution will absorb equal fractions of transmitted light of any color. Beer's law states that light is absorbed by a solution-in proportion to the concentration in the solution of the substance in question. Lambert's law can be demonstrated by pouring coffee into a cup: the bottom of the cup becomes progressively darker as the cup is filled. Beer's law can be similarly demonstrated by adding coffee to a cup of water: the solution darkens as the concentration of coffee increases. At first glance these laws may appear to be self-evident and even trivial. As expressed in mathematical terms, however, they serve as reliable guides for the theoretical study of solutions, for developing analytical systems and for the design of colorimetric apparatus.

"The earliest method of colorimetric analysis was based on Beer's law. It consists of matching the color of an unknown solution with one of known concentration and equal depth. The method requires the preparation of a series of solutions of gradually increasing concentration. All solutions, including the unknown one, are placed in test tubes that have fiat bottoms. The tubes are lighted from below and are examined by looking down the tube toward the light source. The intensity of the light transmitted by the unknown solution is then compared serially with the known solutions until the best match is observed.

"By making use of both Lambert's and Beer's laws an improved apparatus was devised. It requires the preparation of only one comparison solution of known concentration. In this apparatus the two solutions are viewed simultaneously through a pair of rods of optical glass. The ends of the rods are polished to optical flatness. The rods are partially immersed in the solutions, which are lighted from below. The depth of the fluids traversed by the light can be varied by raising or lowering the rods. The rods are moved by a mechanism equipped with calibrated drums that indicate the relative intensity of the transmitted light. By means of a simple calculation one can then determine the concentration of the unknown solution.

"The accuracy that can be achieved by these methods is limited by two factors. First, the observer must estimate the point at which the solutions match. Observations tend to vary with individuals and with fatigue and similar subjective factors. Second, the laws of Lambert and Beer are valid only for light of a single color. Manual comparisons are made with white light, which of course is multicolored. For these reasons the construction of colorimeters of the highest accuracy had to await the development of photoelectric sensing devices.

"Photocell instruments of two basic types have been developed. In one- the spectrophotometer-white light is dispersed into the spectral colors by either a prism or a diffraction grating. An adjustable slit between the dispersion element and the unknown sample admits dispersed light of any desired color to the sample. The colored beam transmitted by the sample falls on the photocell with an intensity that is indicated by an associated meter. The most elaborate models include a motordriven dispersing element for scanning the complete spectrum and a synchronously driven paper chart that moves below a calibrated recording pen. The resulting graph, the absorption spectrum of the sample solution, displays the relative transmission of all colors.

"In the similar but simpler instrument known as the colorimeter, a glass or gelatin filter of selected color admits light to the sample. The transmission is indicated by a photocell and its associated meter. In general, filters transmit a broader band of color than a prism or a diffraction grating equipped with an aperture slit. For this reason colorimeters generate somewhat less precise and detailed information than spectrophotometers.

"Professionals usually make the initial determination of a solution's absorption spectrum with the spectrophotometer. This step is essential in setting up an analytical procedure. The spectrophotometer can also be used for subsequent routine work, but such work can be performed adequately with the less costly colorimeter. The amateur merely selects a filter of the color that has been found by means of the spectrophotometer to be most strongly absorbed.

"After an appropriate filter has been selected the colorimeter is calibrated by preparing serial dilutions of the solution to be analyzed. The intensity of the light transmitted by each dilution is measured and recorded in the form of a graph by plotting transmission against concentration. The strength of any unknown solution of the substance can then be ascertained by measuring the transmission and determining the corresponding concentration by referring to the graph.

"A reliable photoelectric colorimeter is easy to build. Its essential parts include a flashlight lamp powered by a storage battery, a rheostat for regulating the intensity of the light given off by the lamp, a set of gelatin color filters, a cadmium sulfide photocell energized by dry cells, a microammeter and an appropriate cabinet for housing the assembly and holding the test sample. (I tried to energize the lamp from an ordinary power line through a stepdown transformer but found that variations in the voltage caused false indications. Hence I solved the problem by substituting the storage battery.)

"The photocell is simply connected in series with the dry cells and microammeter. I find this circuit adequate. The photocell draws less than .0001 ampere, so dry batteries last a long time. One precaution should be observed when assembling the circuit: Hold the terminals of the photocell tightly with long- 19 nosed pliers when soldering the leads. The pliers keep the terminals cool and prevent heat from damaging the cell.

"I use Wratten gelatin filters for transmitting desired colors. The material is sold by photographic-supply dealers in sheets that measure three inches square. The sheets must be handled carefully by the edges to avoid smudges and fingerprints, which are difficult to remove without damaging the filters. Gelatin filters transmit a much wider band of color than the two-inch-square glass filters normally used in colorimeters. On the other hand, the gelatin material is adequate and is priced at only about $1 a sheet, in contrast to $12 for glass filters. Glass filters are stocked by many laboratory-supply dealers. Experimenters willing to make the larger investment should procure filters of the type used in the Klett-Summerson photoelectric colorimeter.

"The color of filters is expressed in millimicrons. (One millimicron is equal to 10 angstrom units. For example, a 515-millimicron filter transmits mostly green light with a wavelength of 5,150 angstroms.) The numbers by which Wratten filters are identified, however, are not related to their color. Instead the transmission characteristics of these filters are designated by number in the Handbook of Chemistry and Physics.

"Unknown solutions are placed in a conventional test tube that is inserted in the aluminum tube-holder of the in strument [see Figure 1]. To adjust the instrument for operation first place any filter in position and use clear water for the test solution. Put an opaque cover over the top of the test tube. Adjust the light intensity so that the meter indicates about 50 microamperes. Rotate the tube-holder back and forth until the meter indicates maximum current. This is the position of best alignment of the holes in the aluminum tube with the optical system. With a pencil make a pair of reference marks on the holder and on the top of the housing so that the holder can be returned to its optimum position if it is accidentally turned.

"Until the experimenter becomes familiar with the operation of the instrument the coarse control should be turned to the position of full resistance (minimum lamp current) before the power is switched on and the filters are changed. Never withdraw a filter when the lamp is lighted unless the sample-holder contains an opaque test tube, which is most conveniently made by wrapping a layer of black Scotch electrical tape around the tube. Otherwise the meter may be damaged by overload. It is a good idea to keep the opaque tube and a filter in place in the colorimeter when it is not in use in order to protect the photocell and lamp reflector from dust.

"An interesting introductory experiment can be made with potassium permanganate. Insert a green Wratten filter (No. 58) in the filter compartment of the instrument, adjust the coarse lamp control for minimum lamp current and switch on the power. Place a sample test tube of clear water in the sample-holder of the instrument and adjust the lamp intensity until the meter indicates exactly 100 microamperes, using first the coarse adjustment and then the fine one. The response of the photocell is somewhat sluggish, so wait about 15 seconds after each adjustment for the meter to reach its final position. Incidentally, test tubes that contain samples for analysis should be thoroughly clean and dry on the outside before they are placed in the holder. Cultivate the habit of holding the sample tubes near the top to avoid smudges and fingerprints in the region through which the light passes.

"Next remove the tube and after 15 seconds record the new, or 'blank,' reading of the meter. Now make up a solution of potassium permanganate of arbitrary concentration and transfer 25 milliliters to a clean test tube. Place the tube in the instrument and measure the transmission. If the measured transmission does not fall between 35 and 45 microamperes, return the sample to the vessel in which it was prepared and either add more chemical or dilute with water as necessary to produce a meter reading of between 35 and 45 microamperes. This is a cut-and-try procedure. You may have to repeat it several times.

"When the pointer at last falls within the desired range, record the exact indication. Remove the 25-milliliter sample, dilute it to half strength by adding 25 milliliters of water, rinse the test tube with three or four small portions of the diluted solution, fill the sample-holder with 25 milliliters of the diluted solution and again measure and record the transmission. The meter will now indicate a somewhat higher current. Repeat the dilution procedure-cutting the concentration in half each time- until the meter indicates between 90 and 99 microamperes. Record the exact meter indication for each dilution. Each time, just before inserting the test tube with the newly diluted solution, check the blank reading on the meter. It should not change appreciably with time, but any slight drift should be compensated by readjustment of the fine control.

"The first solution of potassium permanganate measured-the one that produces a meter response between 35 and 45 microamperes-should be assigned an arbitrary concentration number of 100. The dilutions are then designated successively as 50 percent, 25 percent, 12.5 percent and so on. The transmission values, in microamperes, are now plotted against the dilutions on semilogarithmic graph paper, with the transmission plotted logarithmically. The result, if a No. 58 Wratten filter is used, will resemble the accompanying calibration graph. Solutions that conform to Beer's law, such as those of potassium permanganate, generate curves in the form of straight lines. The curvature observed in the calibration curve A is caused by the wide transmission characteristic of the filter. Compare it with the associated curve B made of the same solution with a highly selective filter that transmitted wavelengths close to 515 millimicrons.

"Many substances produce curved calibration graphs even when they are measured with a sharply selective filter. Solutions of potassium dichromate, for example, do not yield straight graphs. This and kindred substances do not obey Beer's law because they exist in more than one form. In the case of chromium the solution contains both chromate ions (CrO₄--) and dichromate ions (Cr₂O₇--). The fact of nonlinearity does not invalidate the calibration curve for analytical purposes, however. It simply limits the accuracy and range to a greater or lesser extent, depending on the amount of curvature.

"The following experiments illustrate tested procedures that are common to all photometric analysis. Each substance requires unique processing before it reaches the colorimeter. For the most reliable results distilled water and chemicals of reagent grade should be used. The impurities contained by materials of lower grade may distort the results. The most consistent results require the accurate measurement of weight and volume. For the same reason all solutions should be mixed thoroughly.

"For the analysis of copper the colorimetric technique takes advantage of the fact that copper solutions containing ammonia are deep blue in color. To make a typical analysis, first dissolve .5 gram of pure copper wire in 20 milliliters of nitric acid mixed with an equal volume of distilled water. Work in a well-ventilated room. Heat the solution in a covered beaker to 35 degrees centigrade. When the material has dissolved, boil to complete the reaction and to drive out the brown fumes. Cool and dilute the resulting copper nitrate solution with enough distilled water to make 500 milliliters. Mix thoroughly. To 100 milliliters of this solution add 50 milliliters of concentrated ammonium hydroxide and enough water to make 200 milliliters of solution. Mix. Rinse a clean colorimeter tube several times with small portions of the solution, fill it about a third full and measure the transmission with the red filter (after first determining the blank reading with pure water). Repeat this procedure with portions of the original solution in the amounts of 75, 50, 25 and 10 milliliters. Plot the resulting meter readings against the dilutions to produce a standard reference graph of the metal. The graph should resemble the second accompanying curve [below left].

"With the reference graph at hand it is easy to demonstrate that copper sulfate (CuSO₄^.5H₂O), for example, contains 25.5 percent metallic copper. First, examine a small crystal of the salt with a magnifying glass. White areas indicate loss of the water of crystallization. Such samples should be discarded. Select .25 gram of material from a solidly blue crystal and dissolve it in a small amount of water. Add 50 milliliters of ordinary ammonia water to the solution and enough distilled water to make 200 milliliters. Mix. Measure the transmission after adjusting the light intensity to give the blank reading for the copper calibration curve. The amount of copper, as determined by reference to the calibration graph, should be very close to 25 percent of the weight of the crystal.

"Copper is a constituent of many aluminum alloys. Dissolve one gram of thin drillings from a piece of solid aluminum rod by immersing the drillings in 30 milliliters of hydrochloric acid diluted 50 percent with water. (Do not add all the acid at once.) When the violent reaction is concluded, add three milliliters of concentrated nitric acid and boil until the solids have completely dissolved (with the possible exception of silicon, which, if it is present, will remain as a black powder). Cool and transfer the material to a 200-milliliter volumetric flask. Add 20 milliliters of 20 percent citric acid and mix well. (To make a 20 percent solution of citric acid dissolve 200 grams of the acid in one liter of water.) Then add 50 milliliters of concentrated ammonium hydroxide, mix again, cool and dilute with water until the level of the solution is even with the mark on the neck of the volumetric flask. Mix again. The citric acid prevents the aluminum from precipitating as aluminum hydroxide when the ammonia is added. Filter the solution through a highly retentive paper (such as Whatman No. 42) directly into the colorimeter tube. Rinse the tube several times with small portions of the filtrate and finally measure the transmission of a sample. Find the percentage of copper in the aluminum rod by dividing the indicated number of milligrams by 10.

"The analysis of aluminum can also be made colorimetrically. Aluminum combines with an organic chemical, the ammonium salt of aurin tricarboxylic acid, marketed by the Eastman Kodak Company as Aluminon, to form a red compound known as lake. Measurements are made with the No. 58 green Wratten filter. The analysis requires both a standard solution and a color developer solution.

"Ordinary aluminum foil is pure enough to be used as a standard. Dissolve one gram with 30 milliliters of a 50 percent dilution by volume of hydrochloric acid in a covered beaker. Add the acid in small amounts. Heat until the solution becomes clear and then cool it. Dihlte to 1,000 milliliters with water and mix. Pipette out 10 milliliters of this solution and again dilute to 1,000 milliliters and mix. This is the standard solution. It contains .01 milligram of alumimlm per milliliter.

"The color-developer is prepared by dissolving exactly .5 gram of Aluminon (Eastman organic chemical No. P4468) in 500 milliliters of water. Add .25 ounce of plain Knox gelatin to a beaker containing about 100 milliliters of water and place it in a pan of boiling water until the gelatin dissolves. Dilute to 500 milliliters. Dissolve 250 grams of ammonium acetate in enough water to make a 500-milliliter solution. Mix the three solutions, add 40 milliliters of glacial acetic acid and about three grams of sodium benzoate (to prevent the growth of mold) and mix thoroughly. Store the solution in a dark bottle.

"Mix exactly 15 milliliters of color developer in a 100-milliliter volumetric flask with 10 milliliters of water and immediately place the flask in boiling water. After exactly five minutes remove the flask, cool it to room temperature with running tap water and dilute the solution to 100 milliliters with water. Mix well. Use this solution to determine the blank value.

"Pipette one milliliter of the standard aluminum solution into a 100-milliliter flask and add nine milliliters of water and 15 milliliters of color-developer. Mix. Place the flask in boiling water and repeat the procedure followed in the blank determination. Make a transmission reading on the solution. Complete the calibration curve by making up five other solutions containing two, three, four, five and six milliliters of the standard solution. Reduce the amount of water added before heating by one milliliter each time. The graph should resemble the accompanying curve A [Figure 6]. Also shown is the improved slope, curve B, obtained by the use of a 515-millimicron filter.

"The antacid ingredient in buffered aspirin is aluminum hydroxide. A tablet can easily be analyzed for its aluminum content. Dissolve a tablet in 30 milliliters of 1:1 hydrochloric acid by boiling for a few minutes. Filter the solution into a 200-milliliter volumetric flask and wash the paper three or four times with 20-milliliter portions of hot water containing a few drops of hydrochloric acid, allowing the wash water to run into the flask. Cool to room temperature and dilute to the mark. Pipette one milliliter into a 100-milliliter flask, add nine milliliters of water, the color developer and so on as before. Find the aluminum content of the solution. The analyzed sample solution of one brand of aspirin indicated a content of .0325 milligram of aluminum. The sample represented only one two-hundredth of the total actual amount. The tablet therefore contained about 6.5 milligrams of aluminum.

"Although a research chemist chooses the proper filter for a substance by referring to its absorption spectrum, the lack of a spectrophotometer need not prevent the amateur from making valid determinations. In general, the following combinations will yield fairly accurate results. Use a green filter for red solutions, a red one for blue solutions and a blue one for yellow solutions. A closing note of caution: Many metallic salts and other chemicals are toxic. Handle them accordingly."

Bibliography

CHEMICAL ANALYSIS OF METALS. American Society for Testing Materials, 1961.

COLORIMETRIC METHODS OF ANALYSIS, Vols. I-IV. Foster D. Snell and Cornelia T. Snell. D. Van Nostrand Company, Inc., 1948-1954.

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.