Photometric Determination of Beryllium in Beryllium-Copper Alloys A n Aluminon Method C. L. LUKE
AND MARY E. CAMPBELL Bell Telephone Laboratories, Znc., Murray Hill,X. J .
been destroyed. Cover and ignite at approximately 1000° C. for a t least 1 hour. Cool over a good desiccant and weigh quickly. Reignite to constant weight. Grams of Be0 X 0.3605 = grams of Be. Transfer 10.0 ml. of the above solution to a 1-liter volumetric flask, add 10 ml. of hydrochloric acid, and dilute to the mark with distilled water.
OLLOWING a recent investigation by Luke and Braun on F the photometric determination of aluminum ( 2 ) ,the present authors developed a similar method for the direct photometric determination of beryllium in 2% beryllium-copper alloys. Thioglycolic acid was used to suppress the interference of copper and iron. Aluminum interfered, but because most of the commercial beryllium-copper alloys contain only small amounts of this metal, it is possible in some instances to ignore the interference. For accurate work, however, a more specific method is needed. Ethylened?minetetraacetic acid has been recently recommended as a complexing agent in the separation of beryllium as the hydroxide in the presence of such elements as aluminum, manganese, chromium, lead, bismuth, copper, cadmium, zinc, cobalt, nickel, and iron ( 4 ) . I t has also been used to complex interfering metals in a photometric method for beryllium (5). It occurred to the authors that this masking agent might be used in place of thioglycolic acid in the photometric method mentioned above. This has been proved to be true and, as a result, a very satisfactory method has been developed for direct photometric determination of beryllium in 2% beryllium-copper alloys.
PREPARATION OF CALIBRATION CURVE
Transfer 0-, 1-, 2-, 4-,6-, 8-, and 10-ml. portions of standard beryllium solution (approximately 0.01 mg. of beryllium per ml.) to seven 100-ml. volumetric flasks. Add 1 ml. of cop er chloride solution (2 mg. of copper per ml.) to each flask a n 8 dilute the solutions to 75 ml. Add 2 ml. of complexing solution and swirl. Add 15.0 ml. of aluminon-buffer composite solution, swirl, dilute to the mark without delay, and mix. Immediately transfer 30ml. portions of the solutions to absorption cells and allow to stand undisturbed and away from direct sunlight for exactly 20 minutes from the time of addition of the aluminon-buffer composite solution. Quickly read the per cent transmittancy of the solutions at 515 mp, using distilled water as the reference solution. Prepare a calibration curve on semilog paper by plotting milligrams of beryllium against per cent transmittancy-the latter on the log scale.
APPARATUS
ANALYSIS OF SAMPLE
Photoelectric Photometer. An Evelyn photometer with a 515 nifi filter ww used in the present investigation and the procedure was witten with this instrument in mind. REAGENTS
Copper Chloride Solution (2 mg. of copper per ml.). Dissolve 0.54 gram of cupric chloride dihydrate in 100 ml. of distilled water. Aluminon-Bder Composite Solution. Prepare as directed by Luke and Braun ( 2 ) .
Table I.
10.
1 2
3
4 5 .
6 7 8 9
Complexing Action of Ethylenediaminetetraacetic Acid Metal Added None 2 mg. Cu 5 mg. Cu 10 mg. Cu 15 mg. Cu 0.0372 mg. B e 0.0372 mg. Be 0.0372 mg. B e 0 . 0 3 7 2 mg. Be
Complexin Solution Adied, M1. 2
Transrnittancy,
%
83.8 83.5 83.0 82.3 46.5 40.5 41.5 42.3 44.5
Ethylenediaminetetraacetic Acid Solution (complexing solution). Mix 2.5 grams of ethylenediaminetetraacetic acid with 30 ml. of distilled water plus a drop of 0,0570 alcoholic solution of methyl red. NeutraIize with 1 to 1 ammonium h droxide to the yellow color of the indicator and warm gently to dissolve the last traces of solid. Cool and dilute to 100 ml. with distilled water. Standard Beryllium Solution (approximately 0:Ol mg. of beryllium per ml.). Dissolve 9.82 grams of beryllium sulfate tetrahydrate in 100 ml. of 1 to 3 hydrochloric acid. Filter if necessary, and dilute t o 500 ml. One milliliter of this solution contains approximately 1 mg. of beryllium. Standardize as follows: Transfer 50.0 ml. of the solution to a 250-mI.. beaker and dilute to 100 ml. Add 10 ml. of complexing solution and then add, slowly and with continuous stirring, 10 ml. of 1 to 1 ammonium hydroxide to reci itate the beryllium. Allow to stand a t room temperature f% 3 tours or overnight. Filter on Rhatman No. 41 pa r and wash well with hot 1 to 99 ammonium nitrate solution wQch has been made alkaline to methyl red with 1 to 1 ammonium hydroxide. Transfer the paper and precipitate to a platinum crucible and heat at a low temperature until carbon has
Transfer 0.2000 gram of the sample to a 125ml. chemically resistant conical flask. Add 5 ml. of hydrochloric acid and 5 ml. of hydrogen peroxide (30%) and cover immediately. Cool if the reaction is too violent. When solution is complete, boil down to about 2 ml. to destroy the excess hydrogen peroxide. Cool, transfer to a 500-ml. volumetric flask, dilute to the mark, and mix. Transfer 5 ml. of this solution to a 100-ml. volumetric flask and dilute to 75 ml. with distilled water. Add 2 ml. of complexing solution and then 15.0 ml. of aluminon-buffer composite solution and proceed aa directed for preparation of calibration curve. With the aid of the calibration curve determine the number of milligrams of beryllium present in the portion taken for analysis. DISCUSSION
The addition of a small amount of the complexing solution prevents interference from moderate quantities of such elements as copper, aluminum, zirconium, titanium, manganese, iron, cobalt, nickel, and zinc in the photometric beryllium determination (see Table 11). The beryllium is but very slightly complexed (see Table I). Under the conditions specified in the method, not much more than about 10 mg. of copper can be tolerated (see Table I). If more than this amount is present in an analysis, the amount of complexing solution will have to be increased. The complexed copper absorbs slightly at 515 mp, but this interference can be eliminated by adding an equivalent amount of copper to the standards.
Table 11. Analysis of Synthetic Mixtures NO.
1 2 3 4 5 6 7
a 9
10 11 12 13 14
1056
Beryllium Bdded, Mg. 0.83 5.58 7.44 0
3.72 5.58 7.44 3.72 3.72 3.72 3.72 3.72 3.72 3.72
Other Metals Added
Beryllium Found, Mg. 1.0 5.6
5.0;hg. A1 5 . 0 mg. A1 5.Omg.Al 5 . 0 mg. A1 2 . 5 mg. Zr 2 . 5 mg. T i 2 . 5 mg. Mn 5 .O mg. Fe 5 . 0 mg. Co 5.Omg.Zn 5 . 0 rng. Ni
;.3 3.6 5.6 7.4 3.7 3.6 3.7 3.6 3.6 3.7 3.7
... ...
105?
V O L U M E 2 4 , NO. 6, J U N E 1 9 5 2 After the addition of the aluminum-buffer composite solution, there is a slow but continuous increase in the intensity of color of the beryllium lake. The rate of increase is very slow after 15 minutes, but a stable reading is not attained even after 60 minutes. For this reason the time interval from the addition of the aluminon-buffer composite solution until the memrement of the transmittancy must be uniform for all standards and samples. In contrast to experience with aluminum (d), the calibration curve for beryllium is very nearly a straight line. A further difference in the behavior of beryllium is that heating is not required to develop the color of the lake. In agreement with recent work ( I ) , it is found that an excess of ammonium hydroxide must be added to obtain complete precipitation of beryllium as hydroxide. EXPERIMENTAL
Measured portions of a standard beryllium chloride solution (approximately 0.01 mg. of beryllium per ml.) or a standard copper chloride solution (2 mg. of copper per ml.) were placed in 100-
ml. volumetric flasks. The mixtures were diluted to 75 ml. and measured amounts of complexing solution were added. A 15.0ml. portion of aluminon-buffer composite solution was added to each solution and they were then diluted to the mark and measured photometrically as directed for reparation of calibration curve. The results are recorded in Tagle I Measured portions of a standard beryllium solution (approximately 1 mg. of beryllium per ml.) plus 0.2-gram milled portions of pure copper metal were placed in 125mI. conical flasks. In some instances, appropriate amounts of various metals in the form of chloride or perchlorate solutions were also added. The mixtures were then analyzed for beryllium as directed in the method. The results are recorded in Table 11. LITERATURE CITED
(1) Am. SOC.Testing Materials,
“A.S.T.M. Methods of Chemical Analysis of Metals,” p. 138,1950. (2) Luke, C. L., and Braun, K. C., A N ~ LCHEM., . in press. (3) Meek, H. V.,and Banks, C. V., Ibid., 22,1512 (1950). (4)Pribl, R., and Kucharsky, K., Collection Czechoslov. Chem. Conamzins., 15, 132 (1950).
RECEIVED for review October 17,1951. Accepted May 13, 1952.
Millicoulometric Redox Titrations LOUIS iMEITES Sterling Chemistry Laboratory, Yale University, .Vezu Hacen, Conn.
M L - C H work has been devoted‘to the development of a method of analysis in which the substance to be determined is “titrated” with a reagent that is generated electrolytically with 100% current efficiency. Szebelledy and Somogyi (do), who originated the technique, used an internal indicator to signal the end point and a silver coulometer in series with the cell to measure the quantity of electricity consumed. Lingane ( I d ) described a hydrogen-oxygen volume coulometer which is more convenient to use and slightly more sensitive. Unfortunately, neither coulometer issufficiently sensitive for use with very small quantities of electricity. More recently the coulometric method of measuring the current consumed in the titration has been replaced by the use of a constant-current generator (3, 18), and the calculations are based on the time during which this known current is applied to the cell, so that one might more properly speak of a “chronometric” titration (81). At the same time amperometric ( 2 , 8, 13, 16, 19, 2%’)and potentiometric (4, 6, 7, 9, 17, 21) methods for the detection of the end point have found wide application. Many of the coulometric titration procedures described in the recent literature have been designed for the titration of 0.1 to 3 niilliequivalents of sample. These procedures are of great importance because they have shown that many types of titrations are possible by coulometric methods, defined the general precautions that must be taken t o ensure satisfactory accuracy, and demonstrated that coulometric methods are capable of giving results as satisfactory as those achieved by ordinary volumetric techniques. But when as much as 0.1 me. of sample is available, it seems probable that the practicing analyst will prefer a familiar volumetric or potentiometric titration with a standard solution; the problems and techniques involved in the preparation and storage of these solutions are well understood and a coulometric titration appears relatively complicated. If only a microequivalent, or less, of the sample is available, the balance is turned in favor of the coulometric procedure. Even if reagents of good quality are used, it would be difficult in most cases to justify the assumption that the very dilute standard solutions needed could be accurately prepared determinately or by dilution, and so a standardization on a micro scale would be necessary. It would equally be necessary to repeat the blank determinations whenever a new lot of any reagent was used. Furthermore, few standard solutions as dilute as would be re-
quired could be expected to maintain constant titers for longer than a day or two. Thus the conventional methods are a t a severe disadvantage in the titration of very dilute samples, and it is here that coulometric methods should find their widest application. The present paper describes the results of work prompted by this line of thought. APPARATUS
The solution to be titrated, usually about 100 ml., was placed in a 200-ml. tall-form lipless beaker provided with a rubber stopper to exclude air. A platinum electrode, consisting of a helix of bright 14-gage platinum wire about 6 mm. in diameter and 10 mm. high, waa immersed in the solution. This is the generating electrode (the cathode in the present work); the other electrode in the electrolysis circuit waa a saturated calomel electrode. This was connected to a small beaker containing saturated potassium chloride solution via a 4% agar-saturated potamium chloride bridge, and the potassium chloride solution in turn waa connected to the electrolysis cell by a bridge like that described by Laitinen (IO) filled with 2 N sulfuric acid. This arrangement effectively prevents contamination of either the solution or the reference electrode. Placing the agar bridge directly in the solution to be titrated gave erroneous results due to attack of the agar by the acid solutions. These two electrodes are connected to a 1.5-volt dry c$l in series with a 100,000-ohm decade resistance box for approximate adjustment of the current. The exact value of the generating current is not a matter of interest, and its variation by even ic 10% during the titration does not lead to a significant error. The oxidation tential of the solution was measured wjth a Beckman Model g p H meter, using the 5-inch external platinum and saturated calomel electrodes furnished by the manufacturer. With very small amounts of oxidant, an appreciable error may be caused by the presence of dissolved oxygen ( 4 , 9), and therefore a slow stream of tank carbon dioxide was passed through the solution for about IO minutea before the titration and during the titration to deaerate the solution. Somewhat unexpectedly, the use of unpurified gas caused large negative errors in the determination of oxidizing agents, indicating the presence of an oxidizable impurity, possibly sulfur dioxide. This was removed by passing the gas through an efficient gas-washing bottle filled with a weakly acidic potassium permanganate solution. It would probably be safer to remove oxygen also by passing the carbon dioxide through a vanadium(I1) perchlorate solution (.lJ115), but in this work no error appeared to result from the omismon of this step. The purified gas was led into the solution through a Corning No. 39533 12C gas dispersion cylinder. Magnetic stirring w a ~used, and all titrations were made a t rmm temperature. All chemicals used were reagent grade. Known amounts of the various oxidants were secured by weigh-