maximum variation in flux of 3% from 
one end of the can to the other. In 
further experiments, quantities of so- 
dium carbonate were irradiated in the 
top, middle, and bottom of acan. The 
specific activities obtained indicated a 
flux variation of less than 1%. 
The most likely source of error from 
attenuation within the can would 
be self-shielding by the aluminum 
alloy samples. In order to determine 
whether or not this was an important 
factor, the array of disks used in an 
analysis was simulated by cylindrical 
coils of aluminum foil (0.001 in.) 2-cm 
in height and l-cm in radius. The 
coils were irradiated and the flux varia- 
tion within them determined by count- 
ing }4-in. circles punched from various 
positions along a radius. The varia- 
tion with radius at half the height of 
the coil was about 5%/cm. 
From the foregoing variations, it can 
be shown that the average flux on a 
disk in the middle of the cylinder 
differs from the average flux on a disk 
at the end of the cylinder by less than 
1%. 
Sodium separation. Aliquots from 
the Na,CO; and NaHCO; standards 
are carried through a process identical 
to the following one for obtaining 
radiochemically pure sodium from the 
irradiated aluminum alloys. The alloy 
is dissolved with HCl and H.O; in the 
presence of 5-mg of carrier Na (as 
NaCl). The residue after the high- 
silicon alloys are dissolved is collected 
and the Si volatilized as SiF,. The 
remaining residue is then dissolved and 
added to the HCl solution. Precipita- 
tion of the Na as sodium zinc uranyl 
acetate follows. The precipitate is 
filtered, then redissolved in hot water. 
As a general clean-up, heavy metals 
are then precipitated as sulfides. This 
step also eliminates the possibility of 
Zn contamination by exchange with 
the reagent, or Cu contamination by 
co-precipitation. The sodium is re- 
precipitated, and the resulting salt is 
used directly for the counting meas- 
urements. The figures on page 30 
show the effectiveness of the purifica- 
tion procedure. 
Activity determination. The count- 
ing measurements were carried out 
using an end-window Geiger tube. 
Enough counts were taken to obtain 
less than 1% statistical error. To 
eliminate self-absorption errors, all B 
radiation was filtered out with a 600 
mg/cm? Al filter and only y rays were 
counted. To eliminate backscattering 
errors, the samples were counted in 
identical containers. Counting meas- 
urements were made at rates no higher 
than 10,000 counts per minute, and ex- 
perimentally determined coincidence 
corrections were applied. 
Sodium Content of Aluminum Alloys 
Previously Neutron Standard 
Aluminum assigned activation Number of deviation 
Alloy % sodium % sodium determinations (%) 
3A 0.0014 0.0017 2 0.00007 
0.002 0.0024 4 0.00016 
0.008 0.0107 2 0.00000 
0.0006 0.0007 1 —_ 
0.005 0.0064 2 0.00010 
0.032 0.0352 4 0.00083 
0.018 0.0199 2 0.00000 
0.023 0.0267 2 0.00007 
0.016 0.0189 2 0.00016 
7.5% Si 0.002 0.0015 1 _ 
0.032 0.0307 2 0.00000 
220 0.001 0.0005 2 0.00000 
0.006 0.0061 2 0.00007 
0.011 0.0125 2 0.00045 
0.026 0.0265 4 0.00040 
2024 0.017 0.0168 4 0.00025 
0.013 0.0130 2 0.00040 
0.006 0.0063 2 0.00010 
0.0025 0.0027 2 0.00000 
7075 0.006 0.0054 2 0.00016 
198 
The activity from the uranium in 
the sodium zinc uranyl acetate salt, 
equivalent to about 0.0003% sodium, 
was determined by means of a blank 
and corrected for. The relative activ- 
ities of the primary standards and the 
unknown samples were read from the 
decay curves at a particular time. 
Results and Conclusions 
Since in positions 14-Z-19 through 
14-Z-21, the super-purity aluminum 
gave from 0.0001 to 0.0002% sodium, 
the blank resulting from the transmu- 
tation reaction can be no larger than 
about 0.0002%. Conversely, the so- 
dium content of that sample is be- 
tween the limits 0.0000 and 0.0002% 
but is indeterminate. A correction of 
about 0.0001 % (44 the observed blank) 
was made in each set of determinations. 
The uncertainty in the blank results in 
a limiting accuracy of +0.0001% 
sodium. 
Precision and accuracy of method. 
In the table, the results of several anal- 
yses are given. Many analyses were 
performed in multiplicate. In the 
range below 0.01 % sodium, the average 
standard deviation was 0.00009%; 
from 0.01-0.02% sodium, the average 
standard deviation was 0.00020%; 
and from 0.02-0.04% sodium it was 
0.00042%. It is apparent from the ob- 
served precision that random errors 
that might be associated with variation 
in flux over the can dimensions, or with 
self-shielding by the aluminum alloy 
samples, and statistical errors associ- 
ated with the counting measurements, 
did not exceed 1%. 
It is seen from the table that activa- 
tion analysis gives results in excellent 
agreement with the previously as- 
signed values for the standards. The 
technique of activation analysis ap- 
pears to be capable of producing 
sufficiently precise and accurate results 
to be useful not only for directly cali- 
brating primary analytical standards 
but for general application as an 
umpire-grade method in cases where 
requirements of sensitivity or limita- 
tions in form or size of sample preclude 
other methods. 
BIBLIOGRAPHY 
1. W. A. Brooksbank, G. W. Leddicotte, H. A. 
Mahlman, J. Phys. Chem. 67, 815 (1953) 
2. A. A. Smales, Atomics 6, 2, 55 (1953) 
S$. W. Herr, Angew. Chem. 24, 679 (1952) 
4. G. W. Leddicotte, 8. A. Reynolds, NucLEon- 
Ics 8, 3, 62 (1951) 
6. H. V. Churchill, et. al., ‘Chemical Analysis of 
Aluminum,” (Aluminum Company of America, 
Pittsburgh, Pa., 1950) 
