multiple-wire counters and studied 
their detection of alpha particles and 
neutrons. A mechanism has been de- 
veloped that successfully explains the 
counting, quiescent-current character- 
istics, the number-vs-distance curve for 
an alpha source, and our observed tem- 
perature dependence. Some appar- 
ently contradictory observations of 
earlier workers are satisfactorily ac- 
counted for. 
Single-Wire Counter 
The form of our single-wire counter 
is shown in Fig. 1. A slightly oxidized 
brass plate and two rectangular Lucite 
studs rest on the base of a Lucite frame. 
A 0.l-mm tungsten wire is stretched 
over the studs and held tight by screws 
at both ends. The plate and the studs 
are accurately machined to ensure uni- 
form separation (~1.2 mm) between 
wire and plate. The plate can be 
moved by pushing the handle attached 
to it in and out. The handle can be 
fixed at any position by the spring-and- 
screw arrangement. 
A positive voltage of 2-5 kv is ap- 
plied to the wire. The cathode is 
grounded through a pulse-quenching 
circuit, and a small part of the positive 
pulse at the cathode is taken to a 
Marconi TF 922 scaler for final count- 
ing. Figure 2 shows the circuit ar- 
rangement used for registering the 
number of sparking events. 
Neutron Detection 
Savel (7) was the first to use a modi- 
fied form of the Rosenblum counter for 
measuring slow-neutron doses and 
alpha contaminations. Nearly simul- 
taneously Connor (6) developed an 
alpha monitor using a multiple-wire 
system in front of a metal plate. To 
— 
o 
st 
— 
= 
e 
= 
ca 
= 
oO 
CS) 
s 
S 
oC 
aS 
c 
oD 
oO 
DP 
te 
3 
oC 
(30. 1635: 3 40 
4§ 
Applied Voltage (kv). 
develop this technique for efficient neu- 
tron detection and study of n,a@ reac- 
tions, we undertook construction of a 
suitable spark counter. Our early ex- 
perience indicates that we can detect 
neutrons with some confidence and re- 
producibility even with as weak a 
source as 100 mg of radium-beryllium, 
that is, ~107 n/sec. 
We designed a system with 15 wires 
in front of a 4 X 4-cm flat brass 
plate. At first there was some diffi- 
culty getting all wires to work together, 
but by minor adjustments a narrow 
voltage region was found in which all 
the wires worked satisfactorily. The 
counting efficiency was greatly en- 
hanced by using many wires. 
Alpha particles released by the n,a 
reaction from a very thin layer of pow- 
dered boron placed close in front of the 
counter wires were used to detect neu- 
trons. The whole assembly including 
the neutron source was placed inside a 
paraffin block of 10-cm wall thickness. 
With the 100-mg Ra-Be source on a 
line perpendicular to the plate, 7 + 1 
cpm was observed. When the boron 
layer was removed, no background 
counting was observed over a period of 
about an hour, even in the presence of 
the intense gamma rays (~1 r/hr) 
from the source. In our geometry one 
observed pulse roughly corresponds to 
the passage of 20 n/cm2/sec in the 
counter area. It seems, however, that 
a further improvement in the efficiency 
of neutron counting is possible by im- 
proving the counter design and using 
a boron-coated cathode. Work is in 
progress in this direction. 
Counting Mechanism 
Our mechanism for the counting ac- 
tion can be described in terms of the 
Vette (Kv) 
5.0 
spark-counter arrangement of Fig. 2. 
A strong electric field exists between 
the wire anode and the plate cathode 
separated by a distance of the order of 
1 mm in atmospheric air. In the ab- 
sence of any strongly ionizing particle 
passing between the electrodes and for 
a sufficiently strong electric field, a 
quiescent corona current starts and 
with increased field is ~100 wa. This 
indicates the presence of a fairly dense 
space-charge sheath forming a small 
distance from the wire in the direction 
of the plate. Most of the properties 
of the spark counter seem to depend 
largely on the behavior of this space- 
charge sheath under various circum- 
stances. This is a situation that does 
not exist in the G-M counter, where 
there is practically no quiescent corona 
current. 
There is an external resistor R = 
R, + Rz in series with the electrodes, 
and a corresponding drop RI when a 
corona current J flows. The effective 
potential drop between the electrodes 
is therefore 
V ett = Va = RI 
where V, is the applied potential. It 
is therefore clear that the effective elec- 
tric field between the electrodes, which 
determines the counting behavior, is 
partly dependent on the external resis- 
tor. As the applied voltage is varied 
the corona current J also changes, and 
therefore Vey may not change in the 
same way as V,. 
Suppose an ionizing particle is now 
allowed to pass between wire and plate. 
On account of the strong electric field 
in the neighborhood of the wire an elec- 
tron avalanche starts at once in the 
same region. But this will succeed in 
producing a spark breakdown between 
R=n4.2 Meg — 
3.5 4.0 45 
Applied Voltage (ky) 
5.0 
FIG. 4. Plots of | vs Va for different values of external 
resistance R show that corona sheath obeys equation, 
Vett = Wa [1 — K(R)R] 
FIG. 5. When Veg: is computed from Fig. 4, result indicates 
that for a critical value of R, Vers should be constant with 
changing Va 
38 
