was not necessary here and 4 gm/l 
p-terphenyl might have been substi- 
tuted for the PPO. 
After the sample was prepared by 
grinding and rinsing it into the bottle, 
stock scintillator solution was added 
to bring the total volume to 25 ml. 
Homogeneous internal standards. 
Benzenecarboxylic-C!* acid (C14-ben- 
zoic acid) standard (1) dissolved in 
toluene served as an internal-standard 
stock solution. Its specific activity 
was 96,860 dpm/ml. 
The internal tritium standard was a 
toluene solution of dihydrocholesterol- 
5,6-T2 (2), which was carefully cali- 
brated against standard tritium water 
by homogeneous liquid scintillation 
counting procedures. Its specific ac- 
tivity was 276,000 dpm/ml. 
Instrumentation. A Los Alamos 540 
coincidence system (3) was used in this 
study. The upper-level discriminators 
were disconnected to eliminate the 
possibility of encountering the some- 
times-useful phenomenon of loss of 
efficiency with increasing electronic 
gain. In its routine application, sus- 
pension counting will certainly make 
use of the desirable properties of 
upper-level discriminators for back- 
ground reduction and even for ‘“‘bal- 
ance point” operation (4). Photo- 
multipliers, shield, and sample were at 
room temperature. 
Two lead shields, with accompany- 
ing optical systems, were used. One 
required that the high voltage to the 
photomultipliers be turned off when- 
ever the sample was about to be 
removed from the counting region. 
A second shield (Fig. 1), which was 
designed for operation outside a re- 
frigeration unit, was constructed with a 
shutter system that allowed sample 
changing to be carried out with the 
high voltage always remaining on. No 
detectable light leak to the photo- 
multipliers occurred during sample 
changing. 
Settings for high voltage, amplifier 
gain, and lower discriminator level were 
such that an unquenched homogeneous 
C'-solution would count with about 
60% efficiency. 
Counting techniques. Two methods 
were available for obtaining repeated 
counts on a sample. In one, the sam- 
ple was shaken between each 1-min 
count and the results were averaged. 
In the other, a series of 1-min counts 
was obtained and the results plotted 
and extrapolated back to zero time. A 
12 
good fit of the plot to the data sub- 
stituted for the averaging used in the 
first method. Shaking the sample be- 
tween each count was the most popular 
method and became quite easy with 
the shield that allowed the high voltage 
to remain on continuously. In fact, 
the only occasion for using the second 
method was with studies of settling 
rates. 
Comparisons of the scatter in count- 
ing data with the two shields demon- 
strated that they were equivalent for 
counting rates >1,000 cpm. Below 
1,000 cpm both shields produced a scat- 
ter somewhat beyond that expected by 
statistics, but the shield with the spe- 
cial optical shutter system was signifi- 
cantly better than the other. It seems 
reasonable to conclude from this that 
improved counting stability at low 
rates results from never allowing the 
photomultipliers to see light and pre- 
serving the continuity of high voltage 
applied across them. 
Data on the efficiency with which the 
suspensions could be counted were ob- 
tained by counting standard suspen- 
sions prepared from assayed materials. 
Duplicate suspensions, to serve as 
backgrounds, were made from the cor- 
responding inactive samples. Averag- 
ing, subtraction of backgrounds from 
active-sample results, and division by 
the total activity present gave the 
desired self-calibrated counting effi- 
ciencies. Especially in the extensive 
C4 studies, the procedure was to add 
a known quantity of the homogeneous 
internal standard (C!*-benzoic acid) to 
both the background and the active 
sample and recount. The resulting 
increases in counts were divided by the 
known C"™ activity added, to yield 
internal-standard counting efficiencies. 
Counting Data 
It is a very useful practice in homo- 
geneous liquid scintillation counting to 
determine counting efficiency by add- 
ing an internal standard after making a 
count. This avoids the dangerous as- 
sumption that the standard counting 
solution has the same counting effi- 
ciency as the other samples. The in- 
ternal standard corrects for the vagaries 
of quenching. 
The ratio of suspension counting 
efficiency to homogeneous internal- 
standard counting efficiency is denoted 
by f. If f is independent of scattering 
and absorption of light by the suspen- 
sion and if it depends only on self- 
TABLE 1—Suspension f Values 
C14labeled Weight 
suspended suspended 
material (mg) f= 
BaCO; 19.4 0.91 
38.0 1.01 
41.5 0.93 
94.7 0.94 
101.4 0.92 
219.7 0.93 
234.9 0.96 
avg. = 0.94 + 0.02 
Phenylalanine 15.3 0.97 
37.3 1.04 
39.6 1.02 
101.7 0.98 
107.0 0.99 
242.5 0.90 
250.6 0.92 
avg. = 0.97 + 0.02 
Liver 20.0 1.05 
20.0 1.07 
60.0 1.01 
60.0 0.98 
avg. = 1.03 + 0.02 
Bacteria 20.2 1.02 
59.4 1.02 
120.0 1.02 
avg. = 1.02 
* Ratio of suspension-counting efficiency 
to homogeneous internal-standard counting 
efficiency, 
absorption of the beta particles, con- 
trollable by reproducibility of grinding, 
then knowledge of f will make the 
homogeneous internal standard very 
useful in suspension counting. 
C4, Table 1 gives f-values for vari- 
ous concentrations of BaCQs, phenylala- 
nine, liver, and bacteria when counts 
were made in the large bottles; f is rea- 
sonably constant in each case. The 
average values seem to be independ- 
ent of the color present in the liver 
samples and the high opacity in the 
larger concentrations of BaCO; and 
phenylalanine. 
These f-values are so close to 1 that 
they indicate almost no self-absorption. 
Because f-values of about 1 were ob- 
tained for the biological materials, liver 
and bacteria, a study utilizing this fact 
was made of the counting of a variety 
of tissues, both lyophilized and wet. 
Samples of about 50 mg were either 
