“Living” somple 
+background 
“Dead” sample 
(background ) 
- 
= 
a 
° 
2 
on 
o 
° 
rs 
© 
a 
° 
£ 
° 
° 
= 
= 
a 
O 50 100-— 50s 2.00 
Filling Pressure (cm CO, at 18°C) 
FIG. 7. Change in anticoincidence rate 
with filling pressure 
cause 0.12-cpm change in the antico- 
incidence counting rate. 
Thus very careful design and con- 
struction of counting equipment is 
essential to achieve the reliable count- 
ing required for this dating method. 
Checking Counter Performance 
The pulse amplitude distribution of 
the coincidence pulses provides a check 
on the CO, with respect to electro- 
negative impurities. The coincidence 
pulses are used because they are due to 
the cosmic-ray meson flux across the 
counter, which is essentially constant, 
apart from variations of a few per cent 
caused by barometric pressure changes. 
The anticoincidence pulse distribution 
is not very suitable because of the large 
change with different samples and also 
because of its lower counting rate. 
The number of coincidence pulses 
from the amplifier between 7.5 and 30 
volts and the number greater than 30 
volts (ie., channel 1 and the sum of 
channels 2, 3, and 4) are plotted versus 
counter voltage in Fig. 5. A small 
amount of electronegative gas dis- 
places these curves as shown. The 
counter voltage at which a pair of 
curves cross can be quickly checked, 
since both curves are steep there. A 
quick and accurate check of CO: purity 
is thus possible. 
The counter voltage at the cross point 
will repeat from one filling to the next 
within 5 volts in 5,500 for a 78-cm 
filling. If 1% of the electrons are being 
lost due to electron attachment, the 
cross-over point will occur at a gas gain 
1% greater than for no loss. A 1% 
change in gas-gain corresponds to a 
counter-voltage change of 23% volts. 
Thus from one filling to the next the 
change in electron loss due to electron 
attachment is less than 2%. 
194 
eee 
Distribution of Anticoincidence Events 
Channel 
Source 
(% of total counts in each channel) 
Residual counter background 14 61 22 3 
C14 disintegrations from CO, in counter 13 75 12 0 
CO*® source outside shield 9 62 29 0 
Radon in counter in equilibrium with daughters 6 32 18 44 
Observations of the shape of the 
curve of channel 1 with change of filling 
pressure suggests that not only is the 
change in electron loss less than 2%, 
but that the actual electron loss is less 
than 2%. Further work with other 
filling gases is needed to prove this. 
Total coincidence and _ anticoinci- 
dence counting rates plotted versus ap- 
plied voltage in Fig. 6 show flat alpha 
and beta plateaus. The flat beta pla- 
teau suggests that the counter is 100% 
efficient for ionization occurring in the 
sensitive volume. This conclusion is 
also supported by a comparison of the 
total meson rate and the cross sectional 
area for the proportional counter and 
for smaller G-M counters known to be 
better than 99.8 % efficient for detection 
of mesons. End effects must also be 
small because of the excellent plateau 
of the counter, but as a precaution, 
these have been eliminated by meas- 
uring the counting rate of living carbon 
using different anode lengths with the 
same style of anode terminations. 
The choice of energy range covered 
by each channel was made so that the 
lower limit of the lowest channel falls 
on the beta plateau while the highest 
channel includes only those pulses due 
to alpha rays. The boundaries of 
channel 2 were chosen so that approxi- 
mately 75% of the C'4 disintegrations 
fall in this band. 
The change in anticoincidence count- 
ing rate with change of filling pressure 
is shown in Fig. 7 for a ‘‘dead’’-carbon- 
sample filling and for a “living’’-car- 
bon-sample filling. 
Background 
Some radon is always present in each 
filling. This radon arises from a trace 
of radium in the lime employed in the 
final step of the chemical purification. 
The amount present is proportional to 
the time the CO; is left absorbed on the 
lime. The pulses due to alpha rays 
from radon and its daughter products 
are recorded in channel 4. However, 
beta rays from daughters of radon pro- 
duce pulses inchannels 1, 2and3. The 
number of these pulses bear a fixed re- 
lation to the number of pulses in chan- 
nel 4, and a correction can be applied. 
The table on this page shows the dis- 
tribution of anticoincidence events in 
the four channels for several sources of 
radioactivity, measured for normal 
operating conditions. 
The remaining background counting 
rate of the proportional counter is the 
sum of four components due to counter 
contamination, mesons not detected by 
the counter ring, gamma rays from 
radioactive contamination of the shield, 
and gamma rays associated with the 
cosmic rays. Previously, when G-M 
counters have been used, it has not 
been possible to investigate the relative 
contributions of these four components. 
The use of a proportional counter, com- 
bined with different filling pressures, 
now enables the following information 
to be obtained. 
1. Alpha and beta rays from radio- 
active contamination of counter mate- 
rials. An alpha counting rate of 0.33 
cpm is obtained when the counter is 
filled with radon-free COz. The source 
of this alpha rate is almost certainly 
traces of the elements of the U and Th 
series in the copper wall of the counter. 
Thus, beta rays will also be present, 
and owing to their greater range, their 
contribution to the background count- 
ing rate will be approximately five 
times the alpha counting rate. Thus 
the counting rate due to radioactive 
impurities in the copper of the counter 
will be approximately 2 cpm. The 
