sequently, t h e fluoride-uranium molar ratios are not significant and are not given. Although t h e amounts of fluorinc and uranium v a r y greatly from sample to sample, good agreement was found between the uranium and fluorine precisions for uranium tetrafluoride and uranyl fluoride. A statistical analysis of the data in Table I11 by means of the expression given previously yields the following results at the 95% confidence level. Elemerit U F
s,
70
1.0.07 1 0 22
L.E., % 1 0 13 1.0.441
ACKNOWLEDGMENT
The author wishes to express his appreciation to J. R. Day and T. H. Barton of the Y-12 Analytical Laboratory in Oak Ridge for their assistance in collecting the data. H e is also indebted to R. C. McIlhenny for his help in preparing the manuscript. LITERATURE CITED
(1) Gahler, A. R., Porter, G., ANAL. CHEX.29, 296 (1957). (2) Hibbits, J. O., U. S. Atomic Energy Commission. BEC Proiect ReDt. Y-883 ( J d v 2. 1952). (3) Susano, C. D.,’ White, J. C., Lee, J. E., ANAL.CHEAT. 27, 453 (1955).
(4) Warf, J. C., “Analytical Chemistry
of the Manhattan Project,” S a tional Xuclear Energy Series, 1st ed., Div. VIII, Vol. 1, pp. 728-30, McGraw-Hill, Yew Tork, 1950. ( 5 ) Warf, J. C., Cline, IT. D., Tevebaugh, R. D., A N A L . CHEhl. 26, 342 (1‘384); Manhattan Project Repts. CC-1981 (Oct. 10, 1944), CC-1983 ( S o v . 10, 1944), CC-2723 (June 30, 1945). RECEIVEDfor reviev May 15, 1957. Accepted July 22, 1957. 8th Annual Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy. Based on work performed for the Atomic Energy Commission by Cnion Carbide h’uclmr Corp. at Oak Ridge, Tenn.
Determi nuti on of Ura nium Dioxide in Stainless Steel X-Ray Fluorescent Spectrographic Solution Technique LOUIS SILVERMAN, WILLIAM W. HOUK, and LAVADA MOUDY Atomics International, A Division of North American Aviation, Inc., Canoga Park, Calif.
b This paper outlines a rapid method for the determination of uranium dioxide in stainless steel b y direct fluorescent x-ray analysis after chemical solution of the sample in perchloric acid. A scintillation counter is used to detect the intensity of the radiation. Modifications in the apparatus are needed to ensure stabilization of the counter. Strontium is used as internal indicator. The determination o f uranium i s unaffected b y the presence o f large amounts o f iron, chromium, or nickel a t the dilutions described. The counting time for four pairs of counts (uranium and strontium) is about 12 minutes. A standard deviation corresponding to 0.5 to 1% of the uranium dioxide present was observed on synthetic samples ranging from 15 to 25% uranium dioxide.
I
fabrication of fuel elements for a n experimental reactor, uranium dioxide and powdered stainless steel (304) are mixed and then encased in stainless steel plates. The test for the homogeneity of uranium dioxide in the finished fuel elements presented a problem. Determination of the amount of uranium dioxide in the finished plates was complicated by the amount of stainless steel present. Gravimetric and volumetric procedures involved long and tedious separations. Direct colorimetric procedures were unsuccessful because of the interference of iron (5) and nickel (6). Fluorescent x-ray techniques were N THE
1762
ANALYTICAL CHEMISTRY
examined as a mpans of solving the problem. Direct radiation of the uranium in steel-encased sample by xrays \vas inapplicable, as the stainless steel case prevented the penetration of the x-rays to the uranium layer. Thus, the sample would have to be altered. The simplest procedure was to prepare an aqueous solution of the sample. Birks and Brooks ( 1 ) analyzed uranium in solution by x-ray fluorescence, but their technique required evaporation of aliquots of the solution and analysis of the dry residue. Pish and Huffman ( 2 ) determined uranium in aqueous and in organic solutions extracted from raw materials. Predominantly large amounts of iron, chromium, and nickel mere not present. This paper describes a rapid method for the determination of uranium dioxide in stainless steel by chemical solution of the sample and by direct fluorescent x-ray analysis of the resulting perchloric acid solution, lvithout further separations or processing. APPARATUS AND REAGENTS
The apparatus consists of a Korth American Philips x-ray spectrograph Type 12049 and a milliampere stabilizer Type 52204 with a FA60 tungsten target x-ray tube and a lithium fluoride analyzing crystal. An additional voltage stabilizer was added to the equipment to regulate the po\ver to the electronic panel. The x-ray tube was operated off the original voltage stabilizer a t 50 kv. and 30 ma. This was necessary to prevent overloading the voltage stabilizer when all the circuits of the ap-
paratus were in operation. A s c i n t i l b tion counter and a scaler were used t o detect and record the counts. The solution cell (4) and the solution techniques (3) have been described. It is important that the space between the scintillation counter and the collimator be shielded with lead to p r e vent pickup of scattered radiation by the counter. Figure 1 shows a diagram of this shielding.
Figure 1.
Lead shielding of system
Reagent grades of hydrochloric, hydrofluoric, nitric, and perchloric acids were used to dissolve the samples. Chemically pure strontium nitrate was used to make the internal standard stock solution of 100 mg. per ml. of solution. EXPERIMENTAL PROCEDURE
Standard Working Curve. Perchloric acid solutions of synthetic mix’tures of uranium dioxide and stainless steel were prepared. The internal standard, strontium nitrate, was added t o t h e solutions to give a concentration of 2 mg. of strontium nitrate per ml. in the test solution. The
time required to record 204,800 counts \vas measured for the R a line of strontium and the La line of uranium. The ratio of counting times of strontium to uranium was calculated and plotted against the concentration (milligrams per milliliter) of uranium to form the working straight-line curve for uranium. Sample Preparation. The sample of the uranium dioxide-stainless steel compact weighing about 1 gram was dissolved in 20 ml. of aqua regia. Several drops of hydrofluoric acid and 15 ml. of perchloric acid were added and the solution was evaporated until dense fumes of perchloric acid were produced. The solution was cooled, and diluted, and 2 ml. of the standard strontium nitrate solution TJ-ere added hy pipet. The perchloric acid solution was transferred t o a 100-ml. volumetric flask and diluted to the mark with water. Approximately 10 ml. of this solution was used to fill the solution cell. The time required, 40 to 70 seconds, to record 204,800 counts was measured for the K a line of strontium and the La line of uranium. The ratio of counting time of strontium to uranium (usually four pairs of counts) was calculated and the concentrations of uranium were obtained from the previously prepared working curve. RESULTS AND DISCUSSION
Uraniuni dioxide can be dissolved in nitric acid, but the presence of the stainless steel necessitated the use of aqua regia t o put the samples in solution. Hydrofluoric acid was used to break up the insolubles and perchloric ucid was used to boil out the hydrochloric, hydrofluoric, and nitric acids. The internal standard was required to approach the accuracy needed in the analysis. Without strontium the ratio of uranium to background would give results with an accuracy of about 95% ( 3 ) . With strontium, the accuracy improved t o approximately 99.57, of the amount of uranium dioxide (15 i 0.06yc) present. Any observable interelement effect is eliminated by adequate dilution of the sample ( 3 ) . The lead shield as shown in Figure 1 improved the ratio of uranium to background by a factor of 2. The standard working curve of ura-
Table I.
Comparison of X-Ray and Chemical Methods
UOZ, % Lab. h-0. 1
2 3 4 5 6 7
8
9
X-ray 17.6 15.8 15.4 16.2 16.5 15.2 15.0 15.8 16.0
Wet
chemical 17.3 15.9 15.5 16.3 16.2 15.2 15.1 16.0 15.8
nium dioxide versus the counting time ratio of strontium to uranium is a straight line. I n Table I, several wet chemical checks of the x-ray method are shown. The results of a series of measurements on several standard solutions are shown in Table 11.
Table II.
Sample,
U& 15.0 17.0 20.0 22.0 24.0
25.0
No. of Reading Pairs 9 9 9 9 9 9 9 9 9 9
Average Tsr/ Tu02 0.705 0,705 0.747 0.808 0,807 0.848 0.848 0.889 0,888
0.907
Table 111.
The correction of the uranium and strontium counting times for background counts did not improve the precision of the results. Doubling the concentration of the stainless steel in the solutions had no effect on the intensity ratios of uranium to strontium.
Accuracy of Results
Std. Dev. in Ratio 0,0030 0.0021 0.0037 0.0037 0.0036 0.0037 0.0025 0.0041
0.0042 0.0043
Std. Dev., ff
10
0.426 0.298 0.495 0,458 0.446 0.3:E
0.285 0.461 0.473 0,474
Std. Dev. Expressed as % UOn 0.064 0.045 0.084 0.091 0,089 0.096 0,065 0,110 0.113 0,118
UO, in Solution, Mg ./an. 1 . 5 i 0.006 1 . 5 i 0.004 1 . 7 =t0.008 2 . 0 i 0.009 2 . 0 3= 0.009 2.2 =t0 009 2 . 2 i 0 006 2 . 4 =t0.011 2 . 4 3= 0.011 2 . 5 i 0.012
Routine Counting Statistics (Four Pairs)
Time to Record Difference 204,800 Counts, Ratio from Seconds Ts,/ dv. S, UOZ Tuoz Ratio 15% Uoz STANDARD SOLUTIOX 49.90 71.30 0,700 0.005 50.42 71.14 0.709 0.004 50.43 71.54 0.705 0.000 50.87 71.98 0.707 0.002 Av. ratio = 0.7052 Range = 0.009 Std. dev. = 0.0030 Std. dev., yo = 0.63 Std. dev. as %I,UOZ = 0.097, 50.05 71.30 0.702 0,002 50.10 71.07 0.705 0.001 50.23 71.70 0.701 0,003 50.57 71.96 0.707 0.003 Av. ratio = 0.7037 Range = 0.006 Std. dev. = 0.0029 Std. dev., 70 = 0.41 Std. dev. as %UOZ = 0.0670 17% UO, STANDARD SOLUTION 49.62 66.85 0.743 0.003 49.45 66.28 0.746 0.000 49.60 66.22 0,749 0.003 49.64 66.70 0.744 0.002 Av. ratio = 0.7455 Range = 0.006 Std. dev. = 0.0029 Std. dev., Yo = 0.39 Std. dev. as %U02 = o.oiyo 20% UO, STANDARD SOLUTION 49.00 60.40 0.811 0.003 48.94 60.48 0.809 0,001 49.03 60.51 0 810 0.002 49.10 60.75 0.808 0.000 Av. ratio = 0.8077 Range = 0.009 Std. dev. = 0.0044 Std. dev., yo = 0.55 Std. dev. as %U02 = 0.11yo 4’7.87 59.05 0.811 0.001 47.75 59.10 0.808 0,002 48.08 59.43 0.809 0.001 59.52 0.811 0,001 48.28 Av. ratio = 0.8097 Range = 0.003
Time to Record 204,800 Counts, - Seconds
S,
uo,
Ratio Tp,/ TUO~
Difference from AV.
Ratio
‘SOLUTION 20y0 UOz STANDARD Std. dev. = 0 0015 = 0 18 Std. dev., % = 0 047, Std. dev. as 22y0 GO, STANDARD SOLUTIOS 0.001 0 847 44 58 52 63 0.007 0 855 44 94 52 55 0.005 44 49 52 80 0 843 0.000 44 92 53 00 0 848 = 0.8482 Av. ratio = 0.012 Range = 0.0059 Std. dev. = 0.69 Std. dev., yo = 0.15% Std. dev. as %UO2 0.849 0.001 45.09 53.10 0,851 0.001 53.02 45.10 0.849 0.001 45.32 53.35 0,850 0 000 45.62 53.65 Av. ratio = 0.8497 = 0 002 Range Std. dev. = 0 0010 = 0 12 Std. dev., % = 0 03% Std. dev. as %GO,
0.887 50.05 56.4,5 0.887 50.05 56.40 0.890 50.35 56.60 0.895 50.63 56.60 Av. ratio Range Std. dev. Std. dev., % Std. dev. as y0U02 0.892 56.43 50.32 0.886 56.82 50.36 0.887 50.59 57.01 0.888 50.84 57.26 Av. ratio Range Std. dev. Std. dev., yo Std. dev. as y’U0,
VOL. 2 9 , NO. 12, DECEMBER 1957
=
= = = =
0.003 0.003 0.000 0,005 0.8897 0.008 0.0039 0.44 0.11yo 0.004 0.002 0,001 0.000
= 0.8880 = 0.006 = 0.0029 = 0.33 = o.osyo
1763
The variation of intensity of the LY line of uranium in the range of 1.5 to 3.0 mg. per ml. rvas of such magiiitude that a scintillation counter was required to obtain the required precision in these analyses. For routine analytical work, four pairs of counts are usually sufficient (Table 111). Voltage Regulation. The voltage regulator supplied m-ith t h e x-ray equipment is not capable of regulating t h e current required by t h e instrumentation. \Then t h e x-ray basic unit is operated a t 50 k r . a n d 30 ma., i t alone draws the maximum rated load from t h e regulator. An additional voltage stabilizer was, therefore, necessary t o regulate t h e voltage input required by the electronic panel. T h i s was a n important factor in ob-
L
taining t h e precision in counting of which t h e equipment is capable. Radiolysis. When a n aqueous solution is exposed t o certain types of radiation (gamma, neutron) gas is formed even though t h e solution is cooled. T h e formation of hydrogen and peroxides in t h e present case is caused by t h e radiolytic action of the x-rays. This can readily be proTed by placing a colorless solution of titanium sulfate in the x-ray beam and noting the development of the yellow color of titanium peroxide. -4 rough calculation shows that 1,000.000 roentgens of Yradiation are produced in 3 minutes a t 1 em., (path to the sample) The formation of hydrogen causes a slight increase in counting time, up to 1 second, on successive counts of the same solution if counting times are prolonged; lion-ever,
the ratio of intensity of uranium to intensity of strontium remains the same. LITERATURE CITED
(1) Birks, L. S., Brooks, E. J., ANAL. CHEW23, 707-09 (1951). (2) Pish. G.. Huffman. -4.A.. Ibid.. 27. \
,
1875-8 (1955).
,
I
’
(3) Silverman, L., Houk,
W. W.,North
American Aviation, Inc., Special Rept. 1788, Narch 15, 1957. (4) Silverman, L., Houk, W.IT., Taylor, W., Sorelco Reptr. 1, 118 (1954). (. 5,) Silverman. L.. ‘Moudv. L.. ANAL. CHEM.28, 45-7 (1956). ’ (6) Silverman, L., Moudy, L., Hawley, I)., Ibid., 25, 1369-73 (1953). RECEIVEDfor review January 28, 1957. Accepted August 22, 1957. Sixth Annual Conference on Industrial Applications of X-Ray Analysis, Denver, Colo., August 1957. Based upon studies conducted for the Atomic Energy Commission under contract AT-1 1-1-Gen-8.
Separation of Strontium from Calcium with Potassium Rhodizonate Application to Radiochemistry H. V. WEBS
and W. H. SHIPMAN
Chemical Technology Division,
U. S.
Naval Radiological Defense laboratory, San Francisco 24, Calif.
Strontium can b e separated from calcium with the organic precipitant, potassium rhodizonate. A single precipitation can separate 50 mg. of strontium, determined as oxalate, from calcium chloride in quantities up to 1 1 grams, with strontium recovery and calcium removal efficiency greater than 80 and 99%, respectively. The extent of recovery, which i s reproducible, is a function of the calcium, strontium, and rhodizonate molar ratios. The two procedures given involve simple and rapid manipulations and employ easily handled chemicals, in contradistinction to the conventional procedure which employs relatively large volumes of 75% nitric acid.
T
Hh + C P ~ R A T I O S of
milligram quantities of strontium from gram quantities of calcium frequently arises in the determination of strontium-89 and -90 radioisotopes in calcareous soil and biological specimens. The conveiitional wet procedure most often used in the radiochemical scheme is predicated upon the differential solubility of strontium and calcium in 75% nitric acid (3). O r i n g to the limited solubility of calcium in concentrated nitric acid, rel-
1764 *
ANALYTICAL CHEMISTRY
atively large volumes of the reagent and repeated precipitations are often necessary to effect separation. Because manipulations with this acid are both hazardous and time-consuming, an alternate procedure was sought. Salts of strontium, but not calcium. react in neutral solution with sodium rhodizonate to form a stable brown-red precipitate ( 1 ) . Use of this reagent to separate these elements was investigated. Preliminary experiments determined the dependence of precipitation upon calcium, strontium, and potassium rhodizonate concentrations. Procedures TI ere developed, and with the aid of a strontium-85 tracer the optimum conditions of separation and recoverv n-ere &:Lblislieci. INSTRUMENTS A N D REAGENTS
Strontium-85 ganinia activity \\-as determined in a scintillation n-ell counter. Potassium Rhodizonate Solution (Eastnian Kodak). Because the reagent is stable for only several hours, a 0.2% aqueous solution was freshly prepared just prior to use. Strontium-85 Tracer. The source of strontium-85 was a year-old sample of neutron-irradiated strontium nitrate.
The purity of the tracer was estahlishcd by gamma spectral analysis. Strontium Solutions. 4 solution of strontium-85 tracer and strontium chloride carrier (Baker, reagent grade) n as prepared to give 3000 c.p.m. per nil. and 12 mg. per nil. of strontium, determined as the oxalate. To avoid high counting rates, another solution of stable strontium m s prepared for diluting the active carrier solution when quantities of the carrier exceeding GO nig. n-ere desired. Standard Cnlciuni Solution. Calcium chloride (Raker, reagent grade) was dissolved to a specified volume 11-ith distilled water; the concentration of this solution (3.9731) n as determined by titration n ith standardized (ethylenedinitri1o)tetraacetic acid. Appropriate dilutions of the stork solution lyere mndc as required. All other chemicals n-ere of c i t h reagent grnde or C.P. quality. CONCENTRATION RELATION
Dependence of precipitation upon the calcium, strontium, and precipitant concentrations was determined in a preliminary experiment. Systems prepared to contain 1-arying concentrations of the cations to be separated m-ere ti-
