Rafter.* The purification procedure 
is basically similar to that of De Vries 
and Barendsen (5). It involves ap- 
propriate washings, absorption on lime 
at 700—750° C, and then re-evolution at 
a higher temperature (800-900° C). 
Electronegative impurities are also ab- 
sorbed on the lime but are notre-evolved 
even at 1,000° C. The purified CO» is 
then passed into the proportional 
counter. 
A low background counting rate is 
achieved by extensive gamma-ray 
shielding. Surrounding the  propor- 
tional counter with a ring of anticoin- 
cidence Geiger-Miiller counters (Fig. 2) 
identifies counts due to cosmic-ray 
mesons. The unshielded background 
of the proportional counter is approxi- 
mately 2,000 cpm while its background 
when operating in the equipment is 
only 10-14 cpm. This background 
counting rate is checked periodically 
by filling the counter with CO, pre- 
pared from coal (more than 50,000 yr 
old). 
Pulses from the counter are linearly 
amplified and then electronically sorted 
into four groups depending on their 
amplitude, as shown in Fig. 3. The 
pulses in each of these groups are then 
sorted into two groups, depending on 
whether they are or are not in coinci- 
dence with a pulse from the ring of 
G-M counters; thus the sorter has 
eight outputs. The energy distribu- 
tion of coincidence pulses is shown by 
the number of counts in the four coinci- 
dence channels, and the energy dis- 
tribution of anticoincidence pulses is 
displayed in the four anticoincidence 
channels. 
Counter Construction 
The proportional counter, con- 
structed from 5-in.-o.d. copper tubing, 
is 28in. long. Previous experience had 
shown that, of the readily available 
materials suitable for counter construc- 
tion, copper had the lowest proportion 
of radioactive impurities. 
The anode is 4-mil tungsten wire 
with a sensitive length of 2134 in. At 
one end of the counter the anode con- 
nection is brought out through a Pyrex 
insulator and is provided with a guard 
ring. At the other end, the anode has 
a long glass insulator, and no guard 
ring is provided. The ends of the 
counter are made approximately hemi- 
spherical to prevent them from flexing 
*T. A. Rafter, New Zealand J. Sci. Tech- 
nol. (in press). 
Ring 
Preamplifier. 
«Mercury 
~ Shield 
FIG. 2. Counter ring surrounds proportional counter, visible with cover removed. 
High voltage for proportional counter is supplied from banks of stable 671%-v bat- 
teries to anode and cathode in equal + and — parts respectively. 
Thus only two 
end insulators have to stand full counter voltage (up to 10,000 v); maximum voltage 
elsewhere is 14 that. 
cathode voltage. 
Small variations in voltage or gain are corrected by changing 
Preamplifiers shown use 4 6AK5's in a ring-of-3 feedback-stabilized 
amplifier with cathode-follower output; gain is 40. Heaters are d-c powered in 
series; plate supplies are heavily filtered. 
Locating preamplifier within y-ray shield 
results in short counter leads and good electrical shielding 
when the pressure is changed and to 
reduce dead space. The sensitive vol- 
ume is 80% of the 7.7-1 total. 
Filling the Counter 
The proportional counter remains 
permanently in position in the gamma- 
ray shield. A glass pipeline connects 
it to the filling system next door. 
The Pyrex filling system is evacuated 
by a mercury diffusion pump backed by 
a two-stage rotary pump. The pres- 
sure in the system can be reduced 
quickly to 10-4 mm Hg. Adequate 
‘condensation traps’’ are provided to 
allow for experimentation with final 
purification and filling procedures. 
The system is also arranged so that 
the most time-consuming operations, 
distillation of CO, and pumping the 
counter, can proceed simultaneously. 
Provision is also made for storing the 
sample being removed from the counter. 
After chemical purification the CO, 
Possible Dating Methods—Why CO:? 
This table indicates some of the comparisons involved in our choice of propor- 
tional counting of CO: as the basis of our radiocarbon dating system. It 
appeared to offer the best compromise of conflicting requirements. 
cus 
detection Possibility of 
efficiency Residual radioactive Man-hours per 
Method (%) background contamination sample 1 ea 
Solid-sample (1) 5 High, inorganic High, 3 X CO. Yes 
preps, solid 
CO, gas pro- Up to 100 Relatively high; Low, only gase- Lowest; convert No* 
portional one C per mole- ous contami- C to CO, first 
cule means large nants, 1.e., Rn in all methods 
counter 
Hydrocarbon Up to 100 Better than CO, Low, but more = High Yes 
gas propor- reagents than 
tional (13, 14) co, 
Liquid Up to 100 Best, small Medium; organic High Yes 
scintillation (15) volume 
t Possibility of isotopic fractionation. 
preps, liquid 
*100% transfer of COz. 
— —  ———— — .2.0€—C._— OO 
191 
