AEROSOL SPECTROMETER AND ITS APPLICATION 
For the last four years the efforts of this lab- 
oratory have been directed toward devising a 
suitable method and realizing it with a practica- 
ble instrument—the aerosol spectrometer. The 
principle of its operation and the methods of 
analysis of the resulting size spectra are briefly 
described in the following because an up-to-date 
account has not yet been published in a con- 
veniently accessible place. This account is fol- 
lowed by the application of this method to nat- 
ural aerosol spectra in comparison with those 
obtained by laboratory methods. 
PRINCIPLE AND D&SIGN OF THE 
ABROSOL SPECTROMETER 
The solution of this task is based on the prin- 
ciple of applying high centrifugal acceleration to 
a continuous laminar flow of the aerosol under 
conditions which cause the fallout of the particles 
to follow Stokes law [Goetz, 1956, 1957a] at so 
small a velocity relative to the air flow as to 
render the surface shear negligible. The conflict- 
ing conditions of a low Reynold’s number (Vz < 
800) (to assure laminar flow) and a high absolute 
flow velocity were met by leading the aerosol 
through a helical channel, the latter being part 
of a rapidly spinning rotor. Consequently the flow 
rate relative to the channel walls can be small 
and is independent of the velocity of the rotor. 
The basic element of the aerosol spectrometer 
(A.S.) is thus a rotor, the periphery of which 
forms a groove covered by a flexible foil which 
seals the groove to an airtight channel. After 
operation the foil is removed from the rotor for 
the investigation of the particle deposit on it. 
If the time of residence in the channel is suffi- 
cient for all suspended particles to reach the outer 
wall, that is, the inner foil surface, the deposit 
thereon represents a complete ‘size-spectrum’ of 
the aerosol tested, effected by the difference of 
radial velocity between the large and the small 
particles: The former are all deposited near the 
entrance, the latter toward the end of the chan- 
nel. 
The locus of the deposit of each size class is 
thus determined by the geometry of the channel, 
the flow rate and the angular velocity of the 
cone, that is, by factors independently control- 
lable. Therefore the measurement of the deposit 
density along the channel yields the size (and 
mass) distribution of the aerosol tested, inde- 
pendent from its fate after deposition, because 
the locus of each particle was determined by its 
165 
‘Stokes’ diameter,’ while still in the airborne 
state. 
A considerable advantage of a spinning helix 
is its impeller action caused by momentum trans- 
fer from the channel walls upon the gas which 
causes a flow in downward direction along the 
channels. An auxiliary pump for moving the gas 
through the channels is thus unnecessary. 
The flow speed through the channel is con- 
trolled independently by inserting flow restric- 
tions (jets) of various sizes at the channel exit. 
As the instrument in its present form is de- 
seribed in detail elsewhere [Goetz, 1957b, 1958, 
1959], only a brief account of principal construc- 
tion and performance is given: 
It consists of two separate units (Fig. 1), one 
is the centrifuge, shown in Figure 2, the other, a 
chassis with the controls, connected with the in- 
strument by cable. 
The conical rotor R (Fig. 2) carries on its sur- 
face 219 turns of a double-helix and represents 
three of the four walls of two independent, iden- 
tical channels of uniform (parallelogram) cross 
section. The channels are sealed by an exactly 
fitting conical cup C which is held airtight against 
R by the ring nut N. 
At the upper end the rotor is guided by the air- 
lubricated ‘frictionless’ bearing assembly, con- 
sisting of the constituent parts K, L, U. 
The entire rotor assembly is contained in a 
heavy, removable shell P, which is sealed on top 
by the window Q. 
The electric motor M (14 hp, 110 V a.c.), 
mounted in the base B, supports and drives the 
rotor by the shaft S, also the armature T of an 
electric tachometer with a maximal speed of 
(26,000 rpm). 
The chassis contains a variable auto-trans- 
former unit for the speed regulation, voltmeter 
and ammeter (for checking the operating condi- 
tions of the motor), the tachometer, switch, fuses 
etc. 
To minimize turbidity prior to entry, the air 
is sucked through the stationary inlet tube I into 
the channels at a flow rate F (lit/min) and passes 
the channels with a relative velocity v, (em/see), 
and a residence time 7» (sec), determined by the 
orifice diameter O (mm) and the rotor speed 
N (rpm). Because of the rotor geometry the flow 
is subjected to a gradually increasing centrifugal 
acceleration starting at Ymin (gravity units) and 
ending at Ymax in the outlet. The flow rates in 
the channels are mutually independent, hence two 
different orifices can be used in one test. The 
