ANALYTICAL CHEMISTRY
1072 Table IV.
Effect of Cation Mixture on Radiant Power of Copper
Composite Stock Soh. Added, % (v./v.)
Copper, Present
7 per
1 0
7.9 6.0 4.0 1.9 0.9 8.0 6.0 4.1 2.0 0.9
so
so
8 0
5"
6 0 4 0
2 0 1 0 8 0 6 0 4 0
10"
2 0 15*
40 2 0
6.0 3.9 1.9
1 0
1 1
10 b
8 8 4 9 1
0 0 0 0 0
7.9 6.0 3.9 1.9 0.9
16 b
so
8.0 6.1 4.1 1.8 0.9
6 0
ti0
4 0 9 0 1 0
40 b
8 0 6 0 4 0 ' 0 1 0
a
b
M1. Found __
Read the copper equivalent of T, from the standard curve. Subtract the amount of copper occurring as a contaminant in the composite stock solution to obtain the amount of copper recovrred. Table IV illustrates the data obtained by utilizing this technique. It is apparent that radiation interferences produced by the addition of large concentrations of foreign ions can be corrected by determining the excess emission at a wave length of 325.1 m9. The results obtained with 80% methanol as solvent xere comparable to those obtained with water. LITERATURE CITED
Cholak, J., Hubbard, D. LI., ISD. EX. CHEM.,A N A L .ED. 16, 728 (1944).
Curtis, G. SV., Knauer, H. E., Hunter, L. A , Am. SOC.Testing Materials, Tech. Pztbl. 116, 67 (1952). Dean, J. A , , AIBAL.CREY.27, 1224 (1955). Dean, J. A., Lady, J. H., Zbid., 27, 1533 (1955). Dean, J. A., Lady, J. H., Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, February 1956. Dean, J. A., Thompson, C., ANAL.CHEX 27, 42 (1955). Dippel, W.A , , Bicker, C. E., Furinan, H. W., Zbid., 26, 553
8.0 6.0 4.0
(1954).
Fink, A. Mikrochim. Acta 1955, 314. Gerber, C. R.. Ishler. iY.H.. Borker, E., AA-AL.CHEM.23, 684
2.0 1.0
(1951).
SO% methanol solvent.
Griggs, 11. A , , Johnstin, R.. Elledge, B. E., IND. EKG.CREM.,
Water solvent.
ANAL.
ED. 13, 99 (1941).
Jordan, J. H., Jr., Petroleum Refiner 33, 158 (1954). Kingsley, G. R., Schaffert, R. R., J . Biol. Chem. 206,807 (1954). Lady, J. H., Ph.D. dissertation, University of Tennessee,
until the null point is obtained on the meter. With this instrument the wave length was found to be 325.1 mp. Measure the interfering radiation of each copper solution containing the intcrfering cations. Calculate the radiant power due to copper by the following equation:
.lugust 1955.
Parks, T. D., Johnson, H. O., Lykkeii. L., IND. ENO.CHEW, ANAL.ED.20, 822 (1948). Robinson, A. R., Newman. K. J., Schoeb. E. J., AIBAL.CHEM.22 1026 (1950).
Tc
=
T ~ . -T (T3tc.1 - Ts)
Raring, - C. L.. Zbid., 21. 425 (1949). (17)
where
T, T.
Weichselbaum, T. E., T-arney. P. L.. 3Iargraf. H. W.. Ibid., 23, 684 (1951).
= radiant power due only to copper = background of pure solvent
RECEIVED for review October 19. 1955. Accepted April 16, 1956.
Dif f e rential Spectrophotomet ric Method for Determination of Uranium C. D. SUSANO, OSCAR MENIS, end C. K. TALBOTT Analytical Chemistry Division, O a k Ridge National Laboratory, O a k Ridge, Tenn.
A rapid spectrophotometric method is described for the determination of uranium in high concentrations with a precision equal to that of conventional volumetric or gravimetric methods. The relative absorbance of uranyl ion is measured at a wave length of 418 mp against a highly absorbing reference standard. From the difference in absorbance, the concentration of uranyl ion in excess of that in the reference standard solution is determined with a high degree of precision. A method for the detertnination of the optimum concentration of the reference standard is presented. The effects of trace impurities are evaluated. In the optimum range of 20 to 60 m g . of uranium per ml., the precision is within 0.3yG.
A
LARGE number of methods have been described (9) for the gravimetric and volumetric determination of macro (luautities of uranium. Several apectrophotometric methods,
dependent on colors developed by the addition of chromogenic reagents, have been applied to the determination of uranium, particularly in low concentrations. However, few procedures for uranium determination, based on the color of the uranyl ion, have been published. A colorimetric method was used by Scott and Dixon (11) for the determination of uranium in leach liquor. Rodden (IO) noted that a differential spectrophotometric technique was used by Brackenbury for the estimation of uranium in alkali peroxide solutions. Recently, a method in which uranium is determined spectrophotometrically in perchloric acid n a s reported (8). While the present paper was being reviewed, Bacon and Milner of the British Atomic Energy Research Establishment reported the results of a similar study ( 1 , 2 ) . They make use of differential spectrophotometry for the determination of uranium in the metal, binary and tertiary uraniumbase alloys, and uranium oxide. Ordinary spectrophotometric methods dependent on the absorbance of uranyl ions fail t o yield satisfactory precision for macro amounts of uranium, because the absorbance scale must
V O L U M E 28, NO. 7, J U L Y 1 9 5 6 0 700-
0 600
-1-
-J
II
0 500 -
~ 0 4 0 0 -
u z m a v)
m
a
0 200-
0 100
I\
INSTRUMENT - BECKMAN SPECTROPHOTOMETER, MODEL DU CUVETTES- SILICA, 1OO.cm.PATH LENGTH TEMPERATURE 2 5 . 2 t 0.1”C. S L I T WIDTH - 0.06 TO 0.07mm.FOR WAVE LENGTHS FROM 600 TO 400 m p 0.18 TO 0 2 5 mrn.FOR WAVE LENGTHS FROM 400 TO 3 3 4 mu REFERENCE SOLVENT H 2 0 U R A N I U M CONC. - 1084 mg./ml.-A I 5 4 2 ma/ml. - 6 10.84m&rnl. -C
method. Hiskey demonstrated that the relative error of the method, dc*/c*, and the concentration of the standard reference solution, CI, are related as indicated hy Equation 1:
-
-
Iil 1
-
0300
fi
1073
when PZ
= concentration of the more roncentrated solution (un-
c1
=
9
known) concentration of the more dilute solution (reference standard)
of transmittance of the more concentrated solution to that of the reference solution a‘ = absorbance coefficient for a point on the absorbanceconcentration graph for cell of 1-cm. light path d C p = relative error of method = ratio
11
’
CI
dm) =
relative instrument error
IJIl
0-
0
30
400 500 600 WAVE LENGTH, f n p
Figure 1.
Absorbance spectrum of uranyl sulfate
cover too \vide a concentration range. T o circumvent this difficulty, i t differential spectrophotometric procedure has been applied to the estimation of uranium in uranyl sulfate solutions. By the use of the differential spectrophotometric techniques which involve measurement of relative rather than absolute concentrations, the precision may be considerably improved. I n fact, the results may often be made to equal the precision attained by classical gravimetric or volumetric methods. Several applications of differential spectrophotometry and the theoretical aspects of this techniqiie have been discussed by Hiskej- (6, 7 ) and Bastian (3, 5 ) . The development of the differential spectrophotometric method for uranium required a study of the absorption spectrum of uranyl sulfate, the selection of suitable reference standards, and the effect of trace amounts of diverse ions and excess acid. These st,udies resulted in the development of a procedure which makes possible the rapid and reliable estimation of uranirini i n concentrations of 20 to 60 mg. per ml. in the final solution. ABSORPTION SI’ECTRI‘11 OF URAUYL SULFATE SOLUTlONS
Figure 1 represents the absorption spectrum of uranyl sulfate solutions over the wave length range, 320 t o 550 mfi. Uranyl and sulfate were determined by gravimetric methods in a stock solution of uranyl sulfate used for the preparation of the four test solutions. The ratio of the two ions was thus determined to be 1 to 1. The Becknian Model DU spectrophotometer used in making the measurements was equipped with a tungsten lamp for the visible spectrum. h liquid-cooled mounting block and thermospacer (Beckman attachments Nos. 2360 and 2021) were used to maintain the temperature of the sample and reference solutions constant at 25 + 0.1’ C. The sensitivity setting was kept constant t’hroughout the series of measurements. T o accomplish this, the galvanometer zero adjustment was made by varying the width of slit. The Tvave length scale was calibrated by means of the lines of a mercury arc sourre.
For dilute solutions, the absorbance of which conforms to Beer’s law, the absorbance coefficient is, of course, independent of the concentration. For more concentrated solutions, the absorbance-concentration graph becomes nonlinear and the absorbance coefficient decreases with increasing concentration. For cells of 1-cm. light path the absorbance coefficient, a’, for a point on the nonlinear portion of the curve equals the slope of a tangent to that point. I n applying Equation 1, it is assumed that a’ is constant over the concentration range c1 to c2. No serious error is introduced by this assumption, so long as . i c is kept small. If the concentrations of the standard reference solution, c1, and of the unknown, c2, are nearly equal, 1 2 / 1 1 will very nearly equal unity, and log 12/11will be practically zero. Under these conditions, for practical purposes, dc*/c*a 0: - l/a’cl. From this, it is appaient that the relative error of the method, dc2/c2, will be a minimum R hen a’cl is a maximum. The optimum concentration of reference solution will be that concentration which gives a’cl a maximum value. T o establish this optimum concentration for the determination of uranium, a series of uranyl sulfate solutions of known uranium content, ranging from 0 to 60 mg. of uranium per ml., was prepared. One of these solutions, with uranium content c1, was used as a reference solution to set the spectrophotometer to zero. The relative absorhance, d r , of a somewhat more concentrated solution of concentration CZ, was then measured. The value of a’ was calculated as -ir/cZ - c1. The process was continued T+ ith reference solutions of increasing uranium content, the Inst containing GO mg. per ml. Finally, a’cl values were computed for each pair of liquids of the series These data are presented 111 Table I.
Table I.
A method described by Hiskey ( 7 ) Tvas follomd in establishing the optimum concentration of reference solution to be used in the determination of uranium by the differential spectrophotometnc
c1,
Ac,
lIg./MI. 0
Mg./Ml. 33.4 25 12 25 05 16 85 16 70 16 04 12.44
8.28 8.35 1G.55 16.70 21.77 %,75 30.9 36 1 -11.2
The absoi,bance of uranyl sulfate was found to be a masimuni at ii wive length of 118 m p . All measurements were, therefore, taken a t this peak. SELECTION OF REFERENCE STANDARDS
Computation of Optimum Concentration for Uranyl Sulfate Reference Standard
20.6
15.4 26.5
46.4
51.5 61.8 ci
=
51:: 41 2
dr 1.75 1.24 1.33 0.882
0.855 0.761 0,589 0.947 0.711 0.776 0 : 743
0.522
Slit Width, Mm. 0,022 0.080 0.084 0.110
0.120 0.165 0.22 0.53 0.68 0.88 1.07 1.29 1.63
a‘ 0.052 0.049 0,053 0.052 0.051 0.048 0,048 0.046 0.046 0.038
o:oi4 0.013
a‘c1
0 : 40G
0,443 0.861 0.852 1.045 1.188 1.42 1.66 1.15
...
0.72 0.80
concentration of more dilute standard reference solution.
Ac = difference between concentration of sample solution and CI. Ac = ca - e l when cg is concentration of more concentrated solution.
A r = relative absorbance of more concentrated solution measured against les8,concentrated solution.
a
=
Ar/cn
-
CI.
ANALYTICAL CHEMISTRY
1074 From this table, it is apparent that a'cl is a maximum when c1 is of the order of 35 mg. per ml. This, then, is the optimum concentration for the reference solution. However, in order to use a solution of this concentration as a reference solution to set the instrument to zero, it was necessary to use a wide slit, of the order of 0.6 mm. This is not desirable because of the loss of resolution and greater possibility that absorbance bands of other components of the system will overlap the band used for uranium determination. For this reason, a somewhat lower concentration of reference solution, 20 mg. per ml., was selected for use. At this concentration, the width of slit required is only about one fourth that required a t 35 mg. per ml., while the value of a'c and, consequently, that of the relative error have been changed by no more than one third.
calibrated and the same glassware was used in the preparation of reference and test solutions. Temperature differences may introduce errors by changing the volumes or by altering the absorbance characteristics of solutions. The first effect may lead to concentration errors; the second may produce a change in the slope of the calibration graph and, consequently, in the calibration factor. In case the absorbance characteristics change with temperature, the magnitude of the error due to temperature difference will depend on the temperature sensitivity of the particular colored solution used. Bastian ( 4 ) demonstrated this relationship for copper perchlorate, potassium dichromate, and potassium permanganate, which differ widely in temperature sensitivity.
ESTIMATION OF URANIUM CONCENTRATION IN UNKNOWN SOLUTIONS
Table 111. Reliability of Differential Spectrophotometric Method in Determination of Uranium in Uranyl Sulfate Solutions
The uranium in uranyl sulfate solutions was calculated from
Uranium, hIg./Ml. Present Found
test data by means of an equation of a calibration graph, as described by Young and Hiskey (14). The equation used is as follows:
29.1 30.50 33.4 35.0 35.4b 36.4 40.4 46.7 46.7 50.2C 58.4C
+
c2 = F X A r cL (2) when c2 = Concentration of uranium in unknown solution, mg. per ml. Ar = relative absorbance (difference in absorbance of unknown solution and reference standard solution) cI = concentration of uranium in standard reference solution, mg. per ml.
F
Ac
= factor = average of a number of - values
Ar
The factor, F , is determined by measuring the difference in absorbance, Ar, of a series of reference solutions and solutions of known and somewhat higher concentration, dividing the difference in concentration, Ac, by the difference in absorbance, and averaging the results (see Table 11).
Table 11. Calibration Factor for Determination of Uranium by Differential Colorimetry Uranium, Concentration Range, Mg./Ml.
Slit Width, Mm.
22-38 22-38 22-42 22-50 22-44 22-50 31-52
0.165 0.165 0.165 0.165 0.165 0.165 0.53
No. of Standards 6 6 9
11 l3 14 4
Av. Coefficient of variation, 76
Average Factor, F
Coefficient of Variation,
20.5 20.4 20.5 20.8 20.1 20.2 20.8 20.5 1
2 1
%
The concentration of uranium in the standard and unknown solutions may differ by as much as 2 to 4 mg. per ml. without seriously affecting the accuracy of the determination. The term, F X Ar, is assumed to equal cZ-cl; so long as the error in this assumption does not exceed 0.4 mg. per ml., no larger error will be introduced into the determination of the concentration of the unknown. CONTROL OF VARIABLES
In precision spectrophotometry, all factors that affect the absorbance to a significant degree must be carefully controlled. Errors due to mismatched cells, dilution effects, and lack of temperature and acidity control often become significant. Accordingly, in this study, all differential absorbance measurements were made with cells that had been carefully matched while filled with highly absorbing solutions. In order to reduce calibration errors to a minimum, all volumetric glassware was
29.0 30.4 33.3 35.1 35.2 36.3 40.3 46.8 46.7 50.0 58.2
Difference Mg. -0.1 -0.1 -0.1
+o.
1
-0.2 -0.1 -0.1 +0.1 0 -0.2 -0.2
% 0.3 0.3 0.3 0.3 0.6 0.3 0.3 0.2 0
0.4 0.4
Sample contained y quantities of Cr,+s, Fe *++, X i + +, and MnOd-. Sample 0.25N with respect to sulfuric acid. Uranium concentration of reference standard differed from that of sample by more than 4 mg. per ml. D
b
In the determination of macro quantities of uranium in uranyl sulfate, errors due to temperature difference were avoided by carrying out all operations a t essentially a constant temperature. All dilutions of sample and standard reference solutions were made a t the same temperature, thus avoiding concentration errors. Relative absorbance measurements were made at essentially the same temperature a t BThich the solutions were processed. It was observed that if no cooling system n-ere used, the temperature of the cell compartment of the spectrophotometer increased 2" per minute. Two different methods were used to maintain the temperature of the sample and reference solution equal and essentially constant. In some tests, a fresh standard reference solution was used with each test sample and relative absorbance was measured rapidly (N
