Colorimetric Determination of Ruthenium with p-Nitrosodimethylaniline

mended procedure is simple, requires little of the operator's time, and is more sensitive than other methods in the literature. THE colorimetric deter...
2 downloads 0 Views 386KB Size
Colorimetric Determination of Ruthenium with p-Nitrosodimethylaniline J. E. CURRAH’, ALICE FISCHEL, W. A. E. McBRYDE,

AND F. E. BEAMISH Department of Chemistry, University of Toronto, Toronto, Ontario, Canada

The growth i n importance of the chemistry of ruthenium has brought with it demands for methods of determining very minute amounts of the metal. From several organic compounds tested for color formation with ruthenium salts, p-nitrosodimethylaniline has been selected for more detailed study. This reagent i n aqueous solution will develop a bottle-green color when warmed with a solution of ruthenium chloride or bromide buffered to pH 4.1. Maximum absorption occurs at 610 mp when this solution i s compared with a reagent blank. Ruthenium may be separated from the other platinum metals by conventional distillation procedures and then determined in the distillate. The recommended procedure is simple, requires little of the operator’s time, and is more sensitive than other methods in the literature.

T

HE colorimetric determination of ruthenium has been the

subject of several investigations in recent pears. Breckenridge and Singer ( 4 ) applied 5-hydroxyquinoline-8-carboxylic acid; Ayres and Young ( 2 ) and DeFord ( 5 ) independently studied thiourea as a colorimetric reagent; Ayres and Young (3) also examined rubeanic acid (dithio-oxamide) for this purpose. Sandell ( I O ) discussed these and other methods. llarshall and Rickard ( 7 ) described a procedure in which the color of potassium ruthenate was made the basis of a colorimetric determination but this was applicable t o higher concentrations than any of the foregoing methods. In this paper p-nitrosodimethylaniline in aqueous solution is recommended as a colorimetric reagent of high sensitivity for ruthenium. This substance has already been successfully applied by Yoe and Overholser ( I d ) to the colorimetric determination of palladium. Ogburn (8)reported that various nitroso compounds, including p-nitrosodimethylaniline, gave no visible reaction with ruthenium salts. However, the reaction described in this paper requires heating to develop the color, and Ogburn may have overlooked this. APPARATUS AND SOLUTIONS

Optical Instruments. Absorption measurements were made mainly with a Klett-Summerson photoelectric colorimeter. The variation of absorbancy with wave length R-as examined with a Beckman Model DU spectrophotometer. I n order to establish a relationship between the logarithmic scale of the Klett instrument and conventional absorbancy units, some comparisons were made with the same solution in 2-cm. rectangular cells in the Klett colorimeter, in 1-em. rectangular cells in a Lumetron Model 402 EF colorimeter, and with the same filter in each instrument. The filter used was a Lumetron monochromatic KO.RI 610, having maximum transmittance a t 610 mp. It was found that, under these conditions, 519 Klett units equaled unit absorbancy. Most of the measurements were made n-ith the solutions contained in cylindrical Klett tubes, and these were carefully matched in the following manner. The absorption of samples of a large batch of green chromium(II1) chloride solution was measured five times in each of a dozen Klett tubes. The entire sixty measurements m r e made in random order. The concentration of this solution was selected to give a n absorption of about 200 Klett units when the 610 mp filter was used. The results were examined by conventional statistical methods to compare the variance between tubes with that within tubes. On the basis of this examination it was possible t o select sets of tubes for the follom-ing experiments for which the variance of measurements from tube to tube was not significantly greater (to a probability of 5 % ) than that of measuremrnts within any one tube.

Standard Ruthenium Solution. A stock solution of “ruthenium chloride” was prepared by distillation of ruthenium tetroxide follomd by refluxing in 6 M hydrochloric acid, according t o the procedure of Rogers, Beamish, and Russell (9). The stock solution was standardized by the thionalide method described by these authors.

The thionalide precipitation continues to be successfully applied t o the determination of ruthenium in this laboratory. Recent statements by Flagg (6) that results on semimicro quantities tend to be low by as much as 10% have never, so far as the authors are aware, been supported by experimental evidence. Such statements cannot be reconciled with the observations of Thiers, Graydon, and Beamish ( I I ) , made with the aid of radioactive ruthenium, that, within the prescribed acidity range, precipitation of 6 mg. of ruthenium was complete t o less than 0.6 microgram. Solutions of concentration suitable for colorimetric Kork w r e prepared by volumetric dilution from the stock solution. Two such solutions, which were used for many of the results in this paper, contained 44.6 and 4.46 micrograms of the metal per ml. Buffer Solutions. Buffer solutions containing acetic acid and sodium acetate were made up in concentrated form, so that 1 ml., when diluted to 25 ml., gave the desired stabilized pH. The buffer in the recommended procedure x i s 4 Jf in acetic acid and 1 M in sodium acetate, and this gave a p H of 4.1 when diluted 1 t o 25. Measurements of p H were all made with a Beckman Model G meter. Color Reagent. p-Nitrosodimethylaniline (D.P.I. KO. 188)

t 1,

0.340

30

, 40

,

,

,

,

,

50

60

70

80

90

T I M E OF H E A T I N G MINUTES

Figure 1. Effect of Time of Heating on Color Intensity 0.89 p.p.m. of ruthenium Klett-Summerson colorimeter with tube cells. lines represent ranges of results.

1 Present address, Canadian Industries, Ltd , RlcJIasterville, Quebec, Canada.

1980

Vertical

1981

V O L U M E 2 4 , NO. 1 2 , D E C E M B E R 1 9 5 2

I

0.4001

>

U Z

0360

0 in

m 4

1

0320

1 8

e

LO VL

I2

14

16

OF SOLUTION

Figure 2. Variation of Color Intensity with Volume of Solution Heated 0.89 p.p.m. of ruthenium Klett-Summeraon colorimeter with tube cells

1

0.400

periods. For heating times longer than 50 minutes, increase in color is relatively small, as shown in Figure 1. Volume of Solution Heated. The intensity of color for a fixed weight of ruthenium and fixed volumes of buffer and reagent, and for a fixed time of heating, was greater when the volume of liquid heated was kept small. This is evidently due to the fact that the approach to maximum color development is slow, and is hastened by having the concentrations of reactants as great as possible. .The volume specified in the procedure given below is 8 ml., which allows for 5 ml. of ruthenium solution. I n Figure 2 i t is seen that less color develops when the same quantities of reactants contained in a larger volume are heated together. Amount of Added Reagent. On a molar basis, a fair excess of reagent over ruthenium is added. A larger excess than that called for in this procedure would produce more intense coloration, but poorer precision, on-ing to precipitation in and on the walls of the heating tube. pH of Solution. It was found that strong acid prevented the development of much color, and strong alkali brought on precipitation of a black material, possibly a hydrous oxide of the metal. For maximum color formation the solution should be buffered to a p H about 4 a t the time it is heated. The variation in color intensity Tvith p H is given in Figure 3.

z 6

2

s:m

0.360

4

>

3 0

Figure 3.

4.0

I \

I

5.0

C

Variation of Color Intensity with pH

I

0.89 p.p.m. of ruthenium Klett-Summerson colorimeter with tube cells

a as prepared in aqueous solutions by dissolving 150 mg. per 100 nil. of solution. The solution was heated in a boiling water bath for about 25 minutes and then cooled under running water.

The dissolving of reagent is practically complete, and it appeared unnecessary to filter the solution. It is essential that this solution be cooled before using; otherwise the color development is incomplete. Solutions thus prepared have been found stable over a period of 2 weeks, and good agreement in color development was found betn-een different solutions made up by the above procedure. The solution was not light-sensitive, and was stored a t room temperature Alcoholic solutions of the reagent, such as used for the determination of palladium, f:ded t o produce a strong color with ruthenium; this was interpreted to mean that the ruthenium may have been reduced by the alcohol t o an inactive oxidation state. Solutions were also prepared from p-nitrosodiethylaniline

(D.P.I. XO. 522) and p-nitrosodiphenylamine (D.P.I. S o . 1127). Osmium Solution. -in osmium solution containing very close to 200 mg. of the nietal per liter of 1 M hydrochloric acid was

prepared and standardized by hydrolytic precipitation in the usual manner. Fiftyfold dilution of this gave a solution containing 4 micrograms per nil. COLOR REACTION

When an aqueous solution of p-nitrosodimethylaniline is heated with a solution of ruthenium chloride or bromide, a b o t t l e green colored solution is formed. The reagent itself has a ye!low color in aqueous solution. The intensity of the green color depends on several factors, which must be controlled in order t o obtain uniform results. Time of Heating. At higher temperatures (90' to 100" C.) there is some tendency to get precipitation and hence less color. At lower temperatures color development is slower. The procedure adopted was to heat for 50 minutes a t about 70" C. For shorter periods of heating the color development is not complete and results tend to be lower and not stable over 24-hour

600

500

WAVE

700

LENGTH

mlJ Figure 4. Variation of Absorbancy with Wave Length A . Reagent solution, water blank B. C.

Osmium, 0.8 p.p.m., plus reagent Ruthenium, 0.89 p.p.m., plus reagent

D. C - A

The absorbancy of the colored solution and that of the blank compared with pure water are given in Figure 4. The difference betneen these, which is shon-n as a broken curve in this figure, is obtained when the reagent solution alone is used t o set the colorimeter a t zero. -iceording t o this, a broad band of absorption, centered about 600 mg> distinguishes the ruthenium from the reagent solution. In the photometric measurements a filter with transmittance centered at 610 mg was used. Beer's law was found to hold throughout a wide range of absorbancy, as shown by the data in Table I. The solution in stoppered flasks \vas found t o remain unchanged in absorption for periods of 21 hours. The related compounds p-nitrosodiethylaniline and p-nitrosodiphenylamine were tested for activity toward ruthenium under these conditions. The former produced a somewhat less intense color, but the solution of the reagent denatured, so that progressively less color n-as obtained with each batch after the fmt day. N o marked color was obtained in the second case. Both compounds are less soluble in water than the dimethylaniline compound, and it was necessary t o prepare an acetonewater solution of these, from which there was considerable tendency t o get precipitation.

1982

ANALYTICAL CHEMISTRY

The possibility of osmium occurring together with ruthenium in an analysis required examination of the reaction of the reagent with osmium solutions. .4 chloride solution of osmium, of concentration about equal to that of the ruthenium, was tested by the same procedure. The absorption compared t o a blank was about 20% of that of a ruthenium solution of equal concentration. The absorption curve is included in Figure 4. -4s good separation of ruthenium from osmium is usually achieved by conventional distillation procedures, this interference is not a serious disadvantage. Also, as ruthenium is nearly always isolated by distillation before it is determined in an analysis, no extensive investigation of interfering elements was undertaken. It was established, in preparation for the distillation procedure t o be used, that sodium chloride or bromide in concentrations up to 0.2 -liand small amounts of magnesium chloride did not affect the color intensity, It was also established that sulfate and nitrate ions up to 0.5 III caused no significant change in color, but that perchlorate ion did result in a noticeable lessening of color intensity.

Table I. Concentration, P.P.Xf 0.17 0.34 0 . 5 1 (7)

.

0.68 0.86

1.03 1.38 1.55 1.72 2 07 2.58

Conformity to Beer’s Law Absorbancy 0.075 0.152 0.233 0.310 0.393

0.466 0.624 0.707 0.762 0,942 1.19

Absorbancy, P.P.M. 0.44 0.45 0.451 0.456 0.457 0.454 0,452 0.456 0.455 0.455

0.461

Recommended Procedure. The sample of ruthenium, containing between 6 and 60 micrograms of the metal, is made approximately neutral and adjusted to a volume of 5 ml. It is treated with 1 ml. of a concentrated acetic acid-sodium acetate buffer solution (see above) and 2 ml. of the reagent solution. The resulting solution is then heated in a water bath a t 70” i 4” C. for 50 minutes. This heating can most conveniently be done in a 25 x 150 mm. borosilicate glass test tube with a pouring lip. The solution is then cooled under running water, transferred to a 25 ml. volumetric flask, and diluted to this volume with water. A blank solution is prepared containing only buffer and reagent, and this is used to set the zero on the colorimeter. One blank solution may be used for several determinations made over a period of several hours. The size of the ruthenium sample recommended is based on the advisability of keeping the absorbancy of the colored solution between the limits 0.112 and 1.105. The somewhat arbitrary choice of these limits is based on the fact that, when Beer’s law holds, the quantity dlnc/dT (1) exceeds 5 for solutions whose absorbancies lie outside these values. Since Beer’s law holds in this case, the absorbancy is a linear function of concentration, and these error limits may be translated into concentration values. Where the final solution is diluted to a different volume, or different absorption cells or another instrument is used, theee limits for the weight of ruthenium taken should be adjusted accordingly. Sensitivity. The sensitivity when expressed by the notation of Sandell (10) appears to be the most concise x-ay of expressing and comparing the results of different colorimetric procedures. The authors found that 2,’smicrograms per sq. em. corresponded t o unit absorbancy in the Beckman spectrophotometer a t 610 mM, and 2.9 micrograms per sq. cm. equaled unit absorbancy in the Klett-Summerson colorimeter \!-here the sample was contained in rectangular cells. There was no practical decrease in sensitivity with the simpler instrument. This indicates that this reagent is more sensitive than either thiourea or rubeanic acid for the colorimetric determination of ruthenium. The authors

were unable, perhaps through inexperience with the methods, to match the sensitivity claimed by Ayres and Young for either of these two reagents. DlSTlLLATIONS

The colorimetric procedure was found suitable for determining ruthenium when isolated by distillation. The apparatus used was patterned after that of Thiers, Graydon, and Beamish (11) or that of DeFord ( 5 ) . The former apparatus was employed for distillations in which ruthenium tetroxide was produced either by chlorine and sodium hydroxide, or sodium bromate and sulfuric acid, and received in 3% hydrogen peroxide. ,The procedure of DeFord is simpler and appeared to yield more satisfactory results for very small amounts of ruthenium. In this the tetroxide was formed by perchloric acid and a small amount of sodium biqmuthate, and received in 6 -lI sodium hydroxide. Hydrogen peroxide distillates were evaporated with hydrobromic acid in order to destroy excess peroxide. The ruthenium absorbed in sodium hydroxide was isolated as a hydrated oxide, which was then dissolved by warming with 6 M hydrochloric acid. In both cases some adjustment in p H was necessary before the reagent and buffer were added, owing to the high acidity of these solutions. This was accomplished by adding a saturated solution of sodium bicarbonate until the distillate was nearly neutral; the solution m s then made up to known volume and aliquot portions were taken for colorimetric determination. The distillation procedure of DeFord ivas applied t o the distillation of as little as 200 micrograms of ruthenium, and this colorimetric procedure applied to one tenth of the distillate. When such small amounts as this were being handled, scrupulous attention must be paid to assure no contamination of reagent,s by nitrate. The following results are representative of distillations on this scale: Ruthenium distilled Ruthenium sought (colorimetric) Rutheniumfound (replicate distillations)

223 y 22.3 y 21.8, 22.2, 21.9, 21.4,22.1 y

Although these result’s are, on the average, about 2% low, the difficulty was shown to be connected with thedistillation by analyzing other portions of each distillate by the thiourea procedure (6). Agreement between the colorimetric procedures was taken to indicate that the loss occurred during the distillation. ACKNOWLEDGMENT

The authors wish to thank R. E. Thiers and Kay Fallis Balmer who, while students in this laboratory, performed various experiments and distillations n-hich assisted in the development of this procedure. LITERATURE CITED

(1) Ayres, G. H., ANAL.CHEM.,21, 652 (1949). (2) Ayres, G. H., and Young, F., Ibid., 22, 1277 (1950). (3) Ihid., p. 1281. (4) Breckenridge, J. G., and Singer, S.A. G., Can. J . Research, 25B, 49 (1947). ( 5 ) DeFord, D. D., “Chemistry of Ruthenium,” Oak Ridge, U. S. Atomic Energy Commission, Document NP-1104 (December 1949). (6) Flagg, J. F., “Organic Reagents,” S e w York, Interscience Publishers, 1948. (7) Marshall, E. D . , and Rickard. R. R., ANAL.CHEM.,22, 796 (1950). (8) Ogburn, S.C., Jr., J . Am. Chem. Soc., 48, 2493 (1926). EXG. (9) Rogers, IT. J., Beamish, F. E., and Russell, D. S., IND. CHEY.,ASAL.E D . , 12,561 (1940). (10) Sandell, E. B . , “Colorimetric Determination of Traces of Metals,” 2nd ed.. New York. Interscience Publishers, 1950. (11) Thiers, R. E., Graydon. W. F., and Beamish, F. E., Ax.4~. CHEM..20,831 (1948). (12) Yoe, J. H., and Orerholser, L. G., J . .4m. Chetn. SOC.,63, 3224 (1941). RECEIVED f o r review June 27, 1952.

Accepted September 26. 1952.