D O C . 7 2 P O R E S I Z E O F F I LT E R S 7 1
the radius of the widest bottleneck, then is the necessary overpressure for pas-
sage by the air. Nothing of essence changes in this result if the channels form a net-
work, as can be easily seen. is the defining channel radius for the filtration limit;
for the sake of brevity we will call the “filter width.”
We determined experimentally the “filter width” of a
clay filter as is used for certain bacteriological purposes.
The figure shows the basics of our arrangement. The filter
had the shape of a hollow vessel. It was attached to a hose
for the compressed air and protruded into the ether filling
a glass container. Ether was chosen as the fluid because it
has a capillarity constant about four times smaller than
Through separate trials we found that the pores of the
clay vessel fill very rapidly when it is dipped into the liq-
uid. Submersion of the channel walls therefore certainly occurs.
The appearance of air bubbles in the ether was observed, which originated from
the air that had passed through the clay filter. They appeared at around one atmo-
sphere overpressure.
If we neglect the circumstance that the cross section of the channels (at their nar-
rowest spot) may deviate from the circular, we have to set
where approximately p =
a = 18 should be set. One obtains
The “filter width” thus came to about 0.7
If, on the other hand, the width of the filter channels is determined from the
experimentally ascertained viscous filter resistance, and the experimentally ascertained
total volume of the filter channels, assuming that the filter possesses an unknown number
of identical channels of constant cross section, with the aid of Poiseuille’s law, one thus
obtains for 2r0 a value about ten times larger. This cannot cause wonderment because in
reality the channel cross section is very inconstant. Nevertheless, measurement by means
of capillarity produces exactly the filter width that is essential for the filtering.

0 0
Previous Page Next Page