Separation of Tellurium from Selenium and Other Fission Products by Precipitation with 1,lO-Phenanthroline SIR: 1,lO-Phenanthroline and a number of similar compounds have been used in analytical chemistry for many years ( I , 2). Some time ago we found 1,lO-phenanthroline t o be a suitable precipitant for separating tellurium from selenium and other fission products (3). Though tellurium can be separated from selenium by distilling the selenium (4-7), selectively reducing selenium (8, 9), using ion exchange chromatography (I0-12), or precipitating tellurium with tetraphenylarsonium chloride (IS), or all of these, the precipitation of tellurium with 1,lo-phenanthroline offers a rapid and equally effective separation method. Furthermore, from nitrate-free solutions, bromide and fluoride do not interfere with the formation of the chloro or bromo type of tellurium-phenanthroline salts. In a 6 M HC1 solution, tellurium and phenanthroline form a canary-yellow precipitate whereas a n orange colored one forms in 6 M HBr solutions. These precipitates are readily dissolved with water and d o not form when the acidity is as low as 3 M in HCl or HBr, which is the ideal acidity for reducing tellurium to the metallic form. Therefore, the process of precipitating tellurium with 1,lo-phenanthroline can readily be incorporated into most of the common tellurium separation procedures. However, it should be noted that tellurium(V1) should be reduced t o tellurium(1V) by digesting (fuming) with hydrobromic acid. In addition, to assure interchange of radioactive isomers of tellurium with inert tellurium carrier, the digestion with hydrobromic acid should be followed with a metal reduction step with sulfur dioxide t o form tellurium metal before proceeding with the formation of Te-phen&, salts. Then, after the precipitate has been dissolved in water, it can be reformed by bringing the acidity up to 6 M and/or adding 1 2 M HCl and a few drops of a 6 M HCl solution saturated with 1,lo-phenanthroline. The latter technique is preferred. If reprecipitation of tellurium is not desired, the water solution can be adjusted to 3 M HCl and tellurium reduced to the metal with sulfur dioxide. The approximate composition of the tellurium-halophenanthroline materials is given in Table I. These relative values suggest the empirical formula to be Te(phen)p&, where phen is 1,lO-phenanthroline and X either chloride or
GAMMA-RAY SPECTRA
1 L
w -
+ a
a -
0
zc -
z 3
8W
1 + a
J -
w -
a -
-
L
0
(1) W. W. Brandt, F. P. Dwyer, and E. C. Gyarfas, Chem. Rev., 54, 960 (1954). (2) N. I. Lobanoz and V. A. Smirnova, Russ. J. Znorg. Chem., 10, 453 (1965). (3) R. R. Rickard, D. R. Matthews, and E. I. Wyatt, Tenth Conference on Analytical Chemistry in Nuclear Technology, Gatlinburg, Tenn., Sept. 1966. (4) V. Lenher and D. P. Smith, Ind. Eng. Chem., 16,837 (1934). ( 5 ) W. R. Schoeller, Analyst, 64, 318 (1939). (6) W. W. Meinke, U.S. At. Energy Comm. Rept., AECD-2738 (1949). (7) L. Winsberg and L. E. Glendenin, “Radiochemical Studies: The Fission Products,” C. D. Coryell and N. Sugarman, Ed., Book 3, Part IV-9, McGraw-Hill, New York, N.Y., 1951, p 1443. (8) F. E. Beamish, J. J. Russell, and J. Seath, IND. ENG. CHEM., ANAL.ED., 9, 174 (1937). (9) E. Keller, J. Amer. Chem. SOC.,19,771 (1897). (10) K. A. Kraus and F. Nelson, Amer. SOC.Test. Mater., Spec. Tech. Publ., 195,25-27 (1958). (11) U. Schindewolf, Geochim. Cosmochim. Acta, 19, 134-8 (1960). (12) L. Wish, ANAL.CHEM.,32, 1920 (1960). (13) H. Z. Bode, 2.Anal. Chem., 134,100(1951).
600 800 (000 PHOTON ENERGY, keV
400
1200
4400
Figure 1. Separation of tellurium from selenium
Table I. Partial Composition of Te:X-Phen Complexes Complex
Tellurium
Weight per centa Chlorine
Te-C1-Phen Te-Br-Phen
17 13
...
a ~~~
200
Bromz
31
... 52
Assay made after precipitate was dried to constant weight at
110 “C.
bormide. There is likely to be some error in this approximation because the values listed in Table I were derived from tracer studies and neutron activation analyses. Tellurium and chloride were determined in the chloro salt by neutron activation analyses, where 1z9m,lz9Tetracer and a standard reference solution of tellurium were coupled to estimate tellurium, and indirectly bromine, in the bromo salt. The results of the tracer method used to estimate the tellurium content of the bromo salt suggest the halogen to tellurium ratio t o be constant, that is -6/l. How selective is the phenanthroline precipitant for tellurium ? The selectivity of the phenanthroline reagent for tellurium was first tested with %e and 1 3 T e tracers. The gamma-ray spectra of the mixture ( A ) and the precipitate ( B ) are shown in Figure 1. The use of phenanthroline t o ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
877
separate tellurium from selenium may find some special application. A more practical application involved a direct precipitation of tellurium from a solution of a reactor-fuel specimen. The gamma-ray spectra of the composite and precipitate are shown in Figure 2. The latter spectrum clearly shows the 230 keV photon of tellurium with some loaRu (490 keV) and 1321, the daughter of tellurium. From the spectra, it is seen that tellurium was separated from many fission products sufficiently pure for analytical purposes. Zirconium, niobium, and protactinium can be complexed with fluoride and thus retained in the supernatant liquid. When looking for trace amounts of selenium, it may become necessary to remove gross amounts of tellurium. We have used several precipitations with phenanthroline in order to reduce the number of distillations, selective reductions, and hydroxide scavenges that would otherwise be required to obtain a few milligrams of selenium free of tellurium, and other ions. This works expecially well when trying to obtain ’9Se free of other radionuclides found in nuclear fission products.
R. R. RICKARD E. I. WYATT
t-
I (00
I 700 PHOTON E N E R G Y , k e V I
I
500
500
I 900
1
Figure 2. Separation of tellurium from mixed fission products
Oak Ridge National Laboratory Post Office Box X Oak Ridge, Tenn. 37831
RECEIVED for review September 24, 1971. Accepted January 6, 1972.
Criterion for Judging the Acceptability of Analytical Methods SIR: E. A. McFarren, J. R. Lishka, and J. H. Parker (I) recommended the following characteristic as a criterion of the acceptability of analytical methods total error
=
+
d 2s 100 P
where d is the absolute value of the mean error, s the standard deviation, and p is the correct value. Here the absolute value of the mean error is substituted in relation (A) without regard to its statistical significance-Le., irrespective of whether it represents the actual systematic error or is statistically insignificant. In this latter case, the absolute value of the mean error would only be the result of accidental variation of individual results around the true value. However, in the criterion (A), this oscillation within the limits of variability is already characterized by the value of the standard deviation, s. T. J. Farrell ( 2 ) also used the criterion (A) in his remark on the study ( I ) . I therefore suggest the use of a similar criterion total error
=
100
da
+ 2s
~
P
(1) E. A. McFarren, R. J. Lishka, J. H. Parker, ANAL.CHEM., 42, 358 (1970). (2) T. J. Farrell, ibid., 43, 156 (1971). 878
ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
where, however, dA is the vzlue of the absolute mean error which is statistically significantly different from zero. When the absolute value of the mean error does not significantly differ from zero, dA = 0 is substituted and criterion (B) converts to 100 2 4 , ~ . Statistical significance of the difference of d~ from zero must first be tested by means of Student’s t-test on the significance level CY = 0.05 for the number of degrees of freedom v = n - 1 (3). Here n is the number of determinations from which the absolute mean error and standard deviation were obtained. The significance level for testing the statistical significance of the mean error must be prescribed strictly. The reason is, that if selection of the significance level were left to the discretion of individual authors, the criterion (B) could not become a generally acceptable means of comparison. In assessing the acceptability of analytical methods, the criterion (B) takes into the consideration the variability of results (2s) as well as the value of the mean error, provided of course that the latter is statistically significant (dA). This also satisfies the demand expressed by Ch. Eisenhart (4)--i.e., that when the uncertainty of results obtained by measurement is being expressed, cases in which the results (3) K. Eckschlager “Errors, Measurements and Results in Chemical Analysis,” van Nostrand-Reinhold Co., London, 1969. (4) Ch. Eisenhart, Science, 160,1201 (1968).
