VAPOR AND GASEOUS POLLUTANT FUNDAMENTALS 1239
concentration level. Since the total flow is large enough
that multiple scrubbers would probably be necessary
anyway, the first scrubbers in the line take the dirtiest gas
and clean it to the level of the next stream. The two are
mixed and sent to the next scrubber, and so on. Only the
last scrubber need be large enough to accommodate the
entire flow. This should solve the air pollution problems to
a sufficient degree.
There are two problems which deserve mention: these
are water pollution from the dissolved fluorides and solid
waste sludges which slough off the baffles. Both of these are
highly contaminated and must ultimately lead to a disposal
problem.
Teller suggests the water problem be solved by the use
of a closed cycle holding pond. Byproduct recovery would
appear to be a possible way to solve this problem, but the
situation requires further study. There is mention of neutral-
ization of HF, but this will give a CaF 2 precipitate, which
would eventually fill even a large holding pond. Regardless
of the method chosen, the water must be cleaned of excess
fluoride before it can be recycled.
Efficiency of fume control in the aluminium smelting
industry has been the subject of a thorough study (Cook et al.,
1971). They note that the recovery process may be divided
into two categories: “interception of total particulate which
includes both fluorine compounds and relatively innocuous
aluminium oxide.
Hydrogen fluoride is by far the dominant gaseous fluo-
rine compound in pot gas. Recovery of HF is a function
of surface area of alumina, the HF concentration in pot
gas delivered to the reactors, quantity of alumina through
the system, and contacting efficiency between pot gas and
alumina. Reaction between HF and alumina is rapid and
with proper reactor design, contacting is no problem. With
a sufficient feed rate of alumina to the reactor, reaction
efficiency remains high even at very low bed levels. On the
other hand, insufficient feed rate or total cessation of feed
for a period of time can impair the efficient recovery of HF
as the bed becomes saturated. Similarly high concentra-
tion of HF in pot gas due to poor operating conditions also
can overload the sorptive capacity of available alumina.
Provided an adequate alumina supply, the recovery of HG
gas is invariably above 99% and frequently above 99.9%.
The interception of particulate fluoride compounds takes
place primarily in the fluid bed of the reactor. Fluoride con-
centration in the bed is found to be about 3½% nearly twice
the concentration of hydrogen fluoride sorbed in alumina.
The final interception of particulate, both alumina and
fluoride, is a mechanical filtration step which depends upon
bag condition. Old or worn bags or poor connection to the
tube sheet cause particulate loss. Bag fabric quality, a highly
variable factor in its own right, also weighs heavily on par-
ticulate loss. Bag fabric quality, a highly variable factor in
its own right, also weighs heavily on particulate recovery
performance.
Solids efficiency tends to be less than gas efficiency,
although this is not always the case. The two are not relatedbecause of factors previously discussed. As we are more
vulnerable to particulate losses than to gas phase fluoride
losses, so are we more vulnerable to fluoride loss in particu-
late than to alumina loss. About 10% of total particulate loss
is fluoride compounds.
The Alcoa 398 system, or any other treatment system,
can only capture fluorides from pot fume collected and
delivered to the system. Hooding and fume duct, therefore,
become important parts of the control system. In newer pot-
lines, hooding is designed to achieve 95% interception.
“The recovery of fluorine in a directly reusable form in
the process results in reduced fluoride additions to smelting
installations. In addition to the performance measurements
taken from operating reactors and pot-room monitors, pat-
terns of fluorine consumption by individual pots, potlines,
and entire plants were examined. Observation of the cells
served by the prototype installation or Alcoa 398 Process
revealed a reduction in aluminium fluoride consumption
from 101 lb/ton aluminium produced to 32lb/ton aluminium.
Since AlF3 is added to bath neutralize Na 2 O incoming with
alumina as well as to compensate for loss of fluorides to pot-
lining and atmosphere, the fluorine cycle in smelting cells
is complex, and part of this reduction in aluminium fluo-
ride consumption was compensated by additional cryolite
requirements. A more general measure of fluoride recovery
was derived by evaluation of calcium fluoride sludge col-
lected in ponds serving wet scrubber installations. All of the
HF and solids which end in a sludge pond, and a little more
besides, are returned directly to the pots with this process.
On a fluorine basis, 30 lb of fluorine are retrieved per ton of
aluminium produced.”
Packed towers have also been used successfully (Specht
and Calaceto, 1967) particularly in phosphate fertilizer man-
ufacture.
Use of the packed tower principle has been twofold:1) To handle high SiF 4 gas loadings in a single,
long vessel. When high SiF 4 gas loading are
scrubbed, the plugging effects can be disastrous
unless there are provisions for a spray chamber
to reduce the concentration of SiF 4 and for the
subsequent removal of silica. A working design
includes a cross-flow packed section for opti-
mum removal of SiF 4 and then a counterflow
packed section.
2) To handle very weak SiF 4 concentration (less than
0.5mg as F/std. cu. ft.) in a simplified counter-
flow tower or to serve as a positive entrainment
eliminator.“Generally, the cyclonic scrubber or packed tower
requires 7 to 10 gal for every 30mg of F reduction. Depending
on the severity of service, liquid pressure at the cyclonic
and spray chamber may be 50 to 1001lb/ sq in. gauge, while
the liquid pressure at the packing area, depending on the
method of distribution, may be 5 to 20lb/sq in. gauge.”C022_001_r03.indd 1239 11/18/2005 2:33:16 PM