358 ABSTRACTS OF PAPERS
between measured and calculated characteristics. Fig. 1 refers
to an experimental type of triode, with an electrode configuration
as shown; Fig. 2 represents the characteristics of a mass-produced
variable-mu pentode.
The theory is also capable of dealing accurately with certain
design problems where experimental work is either impracticable
THE ELECTRICAL PROPERTIES OF SALT-WATER SOLUTIONS
OVER THE FREQUENCY RANGE l-4 000Mc/s
By R. COOPER, M.SC, Graduate.*
(ABSTRACT of a Radio Section paper which was published in March, 1946, in Part HI of the Journal)
The paper is concerned with the electrical properties of JOIU-
tions of sodium choride in distilled water at frequencies within
the ranges 1-13 Mc/s and 690-4 320 Mc/s. In the lower of
these ranges, conductivity measurements were carried out on
solutions of concentrations up to 4% by weight, and showed
this parameter to be independent of frequency over the range in
question. Measurements in the higher-frequency range were
restricted to a 3% solution. These yielded values for the at-
tenuation coefficient and refractive index of the solution, from
which the conductivity and permittivity are determinate by use
of Maxwell's equations. The variation of conductivity with fre-
quency at a temperature of 21° C, derived from these measure-
ments, is shown in the Figure.
The conductivities obtained at the four lowest frequen-
cies of measurement in the 690-4 320 Mc/s range were 4-16,
3-72, 3-90 and 3-78 x 10~2 mho/cm. These compare closely
with the constant value of 3-92 x 10~2 mho/cm, curve (a),
obtained for the same solution in the 1-13 Mc/s band. Beyond
about f 000 Mc/s, however, the conductivity increased rapidly.
This effect is considered in terms of the Debye-Falkenhagen
theory of dilute solytions, which predicts a rise of conductivity
in this region of frequency. The increase to be expected on this
basis is shown in curve (b); evidently it is quite inadequate to
explain the large increase observed experimentally.
The cause of this large increase of conductivity has in fact
nothing to do with the presence of sodium chloride in the water.
It lies in the dipolar character of the individual water molecules,
and in the close proximity of a strong absorption band in water
to the upper frequency limit of the measurements. The effect of
this is shown in curve (c), which has been drawn from the ex-
perimental data for distilled water available in the literature.
Curve (d) is the resultant obtained by arithmetic addition of the
ordinates of curves (a) and (c), and it will be seen that it follows
the experimental points very closely. Confirmation of this addi-
tivity of the ionic and dipolar contributions to the conductivity
was afforded by a series of variable-temperature measurements at
a frequency of 2 710 Mc/s.
A review of the available measurements on sea water is given
in the paper. These do not extend beyond 10 Mc/s, and up to
• Manchester University.
or undesirable, for instance in the case of very small electrode
clearances, or of excessive anode loading. The following examples
of the application of the theory are discussed in more detail: the
influence of a change in control-grid diameter for fixed electrode
distances and given amplification factor; of a statistical variation
of grid pitch during manufacture; and of cathode misalignment
009
006
2 (3)
(2) (3)
001
40 30 20
Free-space wavelength,cm
600 750 1000 1500 3000
Frequency. Mc/s
Variation of conductivity with frequency at 21° ± 1°C, 3% NaCl
solution.
(a)—conductivity given by measurements in the 1-13 Mc/s band.
(b)—estimated rise of conductivity due to disappearance of the relaxation effect in
the Debye-Falkenhagen theory of electrolytes.
(c)—effective conductivity of pure water arising from dipolar nature of the individual
water molecules.
(rf)—arithmetic addition of curves (*) and (c).
The experimental points are shown 5
(2), (3) etc. denotes the number of independent measurements, and the vertical line
indicates the extent of the scatter among the results of those measurements.
this frequency there is no evidence of a frequency dependence of
the conductivity. The measurements recorded above suggest
that this frequency independence will extend up to the region of
600 Mc/s, and it seems reasonable to conclude that beyond this
frequency the behaviour can be deduced, to a first approximation
at any rate, from measurements of the low-frequency (ionic)
conductivity, and the available dipolar absorption data for dis-
tilled water.