1%. Why the top point of each curve corresponding to 3.241-em. depth (25ml. volume) is too large to fall on the curves is not known. Nevertheless, the constant slope of 1.0 and the straight-line relationship which occurs when depths vary from 0.1296 to 1.296 cm. and concentrations range from to 2.366 X mole 0.02366 X per liter, show that the theoretical
relationship between fluorescence and depth is followed for tenfold change in depth over a very wide range of light absorption, even though total absorbance never exceeded 0.01. The instrument design has therefore largely eliminated those effects which could result from varying the depth of solution and affect the relationship between fluorescence intensity and solution depth.
LITERATURE CITED
(1) Fletcher, M. H., AYAL. CHEM.35, 278 (1963). (2) Fletcher, M. H.9 May, Irving, Anderson, looq,J.Pt.W., 12, U. 93-5S. (1954). Geol. Survey Bull.
(3) Mllkey, R, G,, Fletcher, M. H., J . Am. Chem. SOC.79, 5425-35 (1957).
R~~~~~~~ for review J~~~ 11, 1962. Accepted December 5, 1962.
Solvent Extraction Method for Zirconium-97 Use for Evaluating Critical Nuclear Incidents WILLIAM J. MAECK, S. FREDRIC MARSH, and JAMES E. REIN ldaho
Atomic Energy Division, Phillips Pefroleum Co., ldaho Falls,
b Radiochemical analysis is an important means of determining the magnitude of critical nuclear incidents. O n the basis of experience with three incidents in which large levels of fission products were present prior to the incidents, three fission product CeI43, and ZrgT-are nuclides-Mo99, recommended as monitors. A method is described for Zrg7 based on TTA extraction of zirconium activity followed b y milking and counting of the Nbg7 daughter activity. Relative standard deviation is 1.1 %.
A
means of determining the magnitude of a critical nuclear incident is radiochemical analysis of the postcritical material for selected fission product nuclides. Desirable nuclear properties of nuclides for this purpose are (1) a high fission yield, (2) a well-known, preferably simple, decay scheme, (3) the absence of gaseous precursors in the decay chain, and (4) a half life of 1 to 100 hours. The nuclides should be nonvolatile, even at elevated temperature. Reliable methods of separation and dctermination are needed. From the standpoint of the radiochemist, critical nuclear incidents can be classified in two types, characterized by either the presence or absence of fission product. prior to the incident. For example, in a reactor that is temporarily shut down or in a processing plant in which U*35 or Puz39 is recovered from spent fuel, copious quantities of fission products are present. This dictates that the half life of the selected criticality-measuring nuclide be sufficiently short that significant quantities N IMPORTANT
292
ANALYTICAL CHEMISTRY
are not present just prior t o the incident. However, the half life should be a minimum of about 1 hour, because a critical incident invariably contaminates the environs and samples are not obtainable until hours or even days afterwards. The upper half-life value is governed by the age of the fission products in the precritical material. For instance, spent fuel usually is cooled a t least 6 months before processing, and a nuclide with a half life of 100 hours is essentially decayed. Lyon, Reynolds, and Eldridge (3) have suggested nine nuchdes for the radiochemical evaluation of critical incidents. These are, listed according to decreasing half life, 65-day Zr95, 51-day Srsg, 33-day Ce141, 12.8-day Ba", 67-hour M099, 33-hour CelJ3, 17-hour Zrg7, 9.7-hour Sr9l, and 84minute Ba139. On the basis of experience in this laboratory with the evaluation of three incidents, the SL-1 reactor (8) and two in processing vessels in the Idaho Chemical Processing Plant (1, 9 ) , all characterized by the precriticality presence of large levels of fission products. six of these nine nuclides were not suitable. The four longest lived nuclides-Zrg5, SrSg.Ce141, and Bal4--especially the first three, were present in large amounts prior to the incidents. Two nuclides, Ba139 and Sr91, which have long-lived gaseous precursors were spewed from the environs as a result of the high temperature, rapid expansion, and, in the case of the SL-1 reactor, rupture of the fuel cladding. These were found in air samples several miles from the incident areas. I n the decay chain, over 80yoof the Ba'39 is formed through 41-second Xe139 ( 2 ) , and about 60%
of the Srgl is formed through 10-second KrQ1. Also, 60% of Ba140 is formed through 16-second Xe14 (2). Of the three nuclides suitable for the evaluation of the type of incident being discussed, applicable radiochemical methods for molybdenum-99 and cerium-143 have been reported ( 5 , 7 , I I ) . The main purpose of this paper is to describe the development of a method for ZrQ7. In addition to its 17-hour half life, this nuclide has desirable nuclear properties of a 5.9% U*36 thermal fission yield and no significant gaseous precursors. EXPERIMENTAL
Apparatus a n d Reagents. Extractions were made in 25-ml. screw-top test tubes (Kimax K45066A), with Teflon stopcocks sealed to the bottom of the tube, on a 33-r.p.m. extraction wheel (spinnerette model, S e w Brunswick Scientific Co., Ken. Brunswick, N. J.). The strip solutions were collected in 1-inch-diameter x 4-inch cylindrical plastic tubes (Lermer Plastics, Inc., Garwood, S . J.) and counted with a 3 X 3 inch NaI (Ti) crystal coupled to a Nuclear Data 512 channel analyzer. Reagent grade chemicals and the 2-thenoyltrifluoroacetone (TTA) (Peninsular Chemical Research, Inc., Gainesville, Fla.) were used without purification. To prepare the zirconium carrier, dissolve 6.0 grams of ZrOCl? 8H20 in 6 X HKOa. Boil until the solution clears and nitrate decomposition ceases. Cool and dilute to 500 ml. with concentrated "03 and water to make the final concentration 6-11 in nitric acid. Prepare the niobium strip solution by dissolving 131 mg. of KsNbsOlg 16H20 in 500 ml. of water, adding 17 ml. of 30% HzO~,and diluting to 1 liter with 6M HClOI.
Table I. Analysis of an Irradiated Natural Zirconium Nitrate Sample Showing Method Reproducibility
lo9 atoms of
Zr97 at to in sample aliquot
Milking 1 2 3
1 2 3 4 5 7 . 4 5 7.27 7.18 7 . 3 0 7 . 5 7 7.37 7 . 2 9 7.27 7 . 3 4 7 . 4 2 7 . 3 2 7 . 3 7 7 . 4 0 7 . 4 3 7.46
Av. 7.363 Std. dev., single detn. 0.080 Rel. std. dev., % 1.09
of Zr97 atoms present a t to according to general Equation 1, where in this case the final term is zero, and t = tl - to. Correct the number of Zr97 atoms a t t o back to the time of the fission event and divide by 0.059, the fission yield of Zr97 (Z), to give the number of fission events which occurred. X*(Sb-97) =
~
A2
ivlo(Zr97)(e - x d
- e-x?f)
STRIP PHASE
+ iV*'(e-X2t)
(1)
RESULTS AND DISCUSSION
t
I
'
- A1
,
200
Figure 1.
,
400
1
1
600 800 ENERGY ( K E V )
IO00
I
1200
Distribution of fission products in method
Procedure. Add an aliquot of the sample to a 50-ml. centrifuge tube containing 1 drop of zirconium carrier, 1 drop of 1 M potassium iodide, 3 drops of 27M hydrofluoric acid, and 2 ml. of 11.6;M perchloric acid. I n a hood, heat gently a t first, then to fumes of perchloric acid. Fume to near dryness ( 5 0 . 1 ml.). Add 2 ml. of 11.6M perchloric acid and fume to approximately 0.5 ml. (If a precipitate remains, boil with 3 N perchloric acid until dissolution is complete. Continue heating until the volume is reduced t o approximately 0.5 ml.) Cool and quantitatively transfer the sample to a 25-ml. separatory tube with three 3-ml. portions of 3111 perchloric acid. Add 2 drops of 30y0 hydrogen peroxide and 10 ml. of 0.5,'M TTA in xylene, and extract for 10 minutes. Discard the aqueous phase.
Scrub the organic phase for 10 minutes with 10 ml. of the niobium strip solution. Discard the aqueous phase. Kote the time a t the end of the scrub, designating it as to. Add 10 ml. of the niobium strip solution and mix continuously on the extraction wheel for a t least 1 hour. K'ote the time a t the end of the strip, designating it as ti. Drain the aqueous strip into a 50-ml. plastic tube, dilute to a standard volume with water, and perform a gamma spectral analysis. r o t e the time of counting as tz (tl - tz should 2 1 0 minutes to allow decay of 1-minute Nb97m). Integrate the area under the 665-k.e.v. photopeak by summation (99y0 branching ratio through this level) and correct the count rate a t tz back to the time of separation, tl. Convert the Nb97 count rate to atoms and calculate the number
Steinberg (IO), in his monograph on the radiochemistry of zirconium, states that direct counting of Zrg7 activity can involve numerous corrections. Generally the separated zirconium fraction :s beta-counted in equilibrium with the 74-minute Nb9' and 60-second Nbg7" daughter activities. With Z P 5 present, isotope resolution requires the obtaining of half-life data and beta-absorption measurements. A simpler, more reliable scheme seemed to be to separate the zirconium nuclides quantitatively, let niobium daughter activity grow in, milk and count the Xbg7 activity, and compute the Zr9' atom abundance. For the initial separation of zirconium activity, a TTA extraction method (6) was selected that previously had been developed for the quantitative separation of Zr95 from samples containing young (