University of Central Florida

STARS

Retrospective Theses and Dissertations

Spring 1979

Scrubber Design for Phosphoric Acid Production Facility

Edward E. Mayer
University of Central Florida

Part of the Engineering Commons

Find similar works at: https://stars.library.ucf.edu/rtd
University of Central Florida Libraries http://library.ucf.edu

This Masters Thesis (Open Access) is brought to you for free and open access by STARS. It has been accepted for
inclusion in Retrospective Theses and Dissertations by an authorized administrator of STARS. For more information,
please contact [email protected].

STARS Citation
Mayer, Edward E., "Scrubber Design for Phosphoric Acid Production Facility" (1979). Retrospective Theses
and Dissertations. 435.
https://stars.library.ucf.edu/rtd/435
 SCRUBBER DESIGN FOR
PHOSPHORIC ACID PRODUCTION FACILITY

BY
EDWARD E. NAYER
B.S., Florida Technological University~ 1974

RESEARCH REPORT
Submitted in partial fulfillment of the requireMents
for the degree of Master of Science
in the Graduate Studies Program of the College of Engineering
of University of Central Florida at Orlando, Florida

Spring Quarter
1979
 ABSTRACT

This report is an examination of the practical design of a

phosphoric acid production facility wet scrubber. It includes a
brief background of the Florida Phosphate Industry~ in adcition to
the rules and regulations affecting the emission of fluoride
contaminates~ The theoretical aspects of a packed absorption tower
is discussed prior to the actual design of a cross-flow wet scrubber
of the type typically found in the Florida Phosphate Industry.
 ACKNOWLEDGMENTS

I am indebted to the faculty of the Environmental

Engineering Department of the University of Central Florida for
providing me with the basic background necessary to perform as
an engineer.
I would also like to acknowledge the support and technical
advice of the Engineering Staff of Agrico Chemical Company, South
Pierce Chemical Works.
I include a special thanks to my wife, Diane, for her
patience and·many hours of typing.

iii
 TABLE OF CONTENTS

LIST OF ILLUSTRATIONS . v
INTRODUCTION . . . . . . ... . . . 1
Chapter
I. AIR POLLUTION IN THE PHOSPHATE INDUSTRY. 2
Phosphate Industry in Florida. 2
Fluoride Contaminates . . . 3
Rules and Regulations . . . . . . . 4
II. DESIGN OF WET SCRUBBERS. 7
Basic Concepts of Mass Transfer. 7
Theoretical Design of a Packed
Absorption Tower . . . . . . . . . . . 10
Practical Considerations for Fluoride
Wet Scrubbers . . . . . . . . . . .. . . 18
III. DESIGN OF A PHOSPHORIC ACID PLANT
FLUORIDE CONTROL SYSTEM. . . . . . . 24

Phosphoric Acid Production Process . 24

Sources of Air Pollution .. . 27
Design Procedures. . . . . . .... . 27
Design of Packed Section . . 40
Design Details . . 46

SUMMARY . .... . . . . . 49

FOOTNOTES ...... ... . . ... . • 50

iv
 LIST OF ILLUSTRATIONS

Figure
1. Gas to Liquid Transfer Phenomenon 8
2. The Two-Film Theory of Interphase Diffusion .. 9

3. General Schematic of Absorption Tower . . . 13

4. Typical Equilibrium Curve and Operating Line. . . 14

5. Generalized Pressure Drop Correlation . 17

6. Equilibrium Curve for Low Concentration of F. . . . . 20

7. Equilibrium Curve for High Concentration of F . . 21
8. Typical Cross-Flow Scrubber . . . . . . . . 22
9. Wet Process Phosphoric Acid Production. . 25
10. Bird-Prayon Tilting Pan Vacuum Filter . . . . 28
11. Typical Layout of Phos Acid Plant
Indicating Position of Scrubber . . . . 30
12. Canopy Hood for Filter. . . . . . 32
13. Known Inputs into Scrubber . . . 35

14. Effect of Liquid/Gas Ratio on NTU for

Spray Chamber Using Gypsum Pond Water . • • 39

15 Phos Acid Scrubber . . . . . . . . . . . 48

v
 INTRODUCTION

The production of fertilizer products from the Florida

Phosphate Industry is important not only to the economy of Florida,
with 30,000 jobs, but to the United States and the world. Florida
phosphate rock is relatively easy to mine and provides an inexpen-
sive source of a much needed agricultural nutrient. This gigantic
industry, which is primarily located in the backwoods of west-
central Florida, impacts not only the economy but the ecology as
well.
The purpose of this paper is to address one aspect of
pollution control within the industry--fluoride emissions. A brief
survey of the industry and the regulations governing fluoride
emiss1ons will be presented, followed by a discussion of the theo-
retical aspects of gaseous removal. This will hopefully provide
the reader with a background on the problem. The remainder of the
discussion will revolve around the practical design of a phos-
phoric acid production facility wet scrubber.
It is hoped that the result of this paper will be to
provide an easy to follow manual on wet scrubber design suitable
for use by interested students in environmental engineering.
 I. AIR POLLUTION IN THE PHOSPHATE INDUSTRY

Phosphate Industrx in Florida

Large deposits of pebble phosphate rock exist in a "shield-
shaped .. area, nearly 50 miles long and 40 miles wide") covering Polk,
Hillsborough and Hardee Counties. Discovery of phosphate rock in
the Peace River south of Fort Meade in the 1880's ushered in the
industry. Approximately 70 percent of the nation's output and 30
percent of the world's usage come from the central Florida mine-
fie 1ds .1
The deposits were created ten to fifteen million years ago
during the late Miocene Age. At this time water covered Florida,
parts of Georgia and Alabama. Shoreline and ocean dwelling animal
bones provided the calcium that, when leached and reacted with lime-
stone, under pressure, provided phosphate rock.
Florida phosphate deposits consist of three types: land
pebble, hard rock and soft rock. The land pebble, which accounts
for 95 percent of the total Florida production, is centered in Polk
and Hillsborough Counties.2 Shoreline activity produced these
dense deposits. Hard rock deposits are primarily located in north-
ern Florida. Soft rock is from waste ponds of former hard rock
operations.
The typical Florida phosphate company will have two types
of operations: mining and chemical processing. The mining oper-
 3

ation will consist of removing the rock from the ground and prepa-
ration for chemical processing. The preparation usually includes
washing, sizing, drying and benefication.
The chemical processing complexes convert the phosphate
rock into a form usable to plant life. The operation will usually
consist of facilities to produce sulfuric acid, phosphoric acid,
ammonium phosphates, and granular triple superphosphate. It is in
the manufacturing of phosphates that fluoride is released as a
pollutant.

Fluoride Contaminates
Phosphate rock in Florida has the approximate chemical
formulation 3Ca 3(P0 4) 2 • CaF .3 This is a complex of tricalcium
2
phosphate and calcium fluoride. The tricalcium phosphate is only
slightly soluble and when combined with calcium fluoride is nearly
insoluble in water. The compound has about 3.5 to 4.0 percent by
weight of fluoride. Removal of fluorine is necessary in order to
provide a non-toxic phosphate product.
The gaseous formation of fluorides, HF and SiF 4 , is pro-
duced when heat or acid is applied to phosphate rock. Particulates
of rock dust may also contain up to 4 percent fluoride. The
particulates will be released in processing facilities such as
drying, grinding, and material handling.
Fluoride is a cumulative poison and the degree of its
toxicity is a function of both ingestion level and length of ex-
posure. Fluoride ingestion causes a disturbed calcification of
 4
growing teeth. Fluorides are also a protoplasmic poison.
Over the years there have been an increasing number of
reports of injury to livestock and vegetation due to atmospheric
pollution by fluorides. The importance of fluoride as an atmos-
pheric pollutant was emphasized by a number of investigators at the
U.S. Technical Conference on Air Pollution in 1950.4
When present in sufficient concentrations, fluorides in
gaseous form are highly toxic to growing vegetation, humans, and
animals. The President•s Science Advisory Committee 5 in its report
to the President ranked investigations into the systemic effects on
humans, animals, plants, and materials in the highest priority
category along with sulfur dioxide, carbon monoxide, and carbon
dioxide. The minutes of the Florida Air Pollution Control Commis-
sion6 have innumerable pages of testimony regarding the damages to
vegetation, animals, and humans caused by fluoride emissions from
the phosphate industry.

Rules and Regulations

The emission of fluorides from the phosphate industry is
restricted by both the Environmental Protection Agency and the
Florida Department of Environmental Regulation. Both agencies
differentiate between existing and new point sources. In addition,
non point sources are cited.
The E.P.A. proposed standards of air pollution control
performance for five affected facilities within the phosphate
fertilizer industry on October 22, 1974, (39 FR 37602). The final
 5
version of the standards was published on August 6, 1975,
(40 FR 33152) with an effective date of August 4, 1975. These
standards are promulgated under the authority of Section III (a),
(b), and (c) of the Clean Air Act, and apply to new sources for
which construction or modification commenced after the publication
of proposed regulations; i.e., October 22, 1974.
Section III (d) of the Clean Air Act requires that the
E.P.A. establish procedures under which states must develop emission
standards for certain pollutants from which new source performance
standards (NSPS) have been promulgated.
As required by the procedure, E.P.A. first defined desig-
nated pollutants, such as fluoride as a weJfare-related or a health-
related pollutant, and then established a guideline for existing
sources which recommended a level of emission control for these
sources. This guideline determined fluoride to be a welfare-
related pollutant and was published in April, 1976.
Table 1 is a summary of the NSPS Standards promulgated by
the E.P.A. for the phosphate fertilizer industry. The allowable
emissions are an indication of the absolute amount of fluoride
produced in the processes. More fluoride is released in the pro-
duction of GTSP than the other processes. DAP manufacture is
second, phos acid third, and the storage of GTSP last. The emis-
sion limitations in this table will be adhered to in the following
pollution control equipment design.
 6

TABLE 1
FLUORIDE EMISSION FACTORS
Source NSPS Emission
Wet Process Phosphoric Acid . . 0.02 lb F/ton P205
Granular Triple Superphosphate . . . 0.15 lb F/ton P2o5
(GTSP)
Diamrnonium Phosphate . . . . . .
(DAP)
Granular Triple Superphosphate
Storage Facilities . . . . . . . 5.0 x 10- 4 lb F/hr/ton stored

Source: U.S. Environmental Protection Agency, Inspection

Manual for Enforcement of New Source Performance Standaras,
Phos hate Fertilizer Plants, Stationary Source Enforcement Series
Pu n. EPA 340/1-77-009 Washington, D.C.: Government Printing
Office, 1977), p. 7.

Florida's regulations for new sources follow the guidelines

of the E.P.A. Existing sources fall under the ".4 Rule." This
states that a facility cannot emit more than .4 pounds of fluoride
per ton of P20s input into the wet phosphoric acid production
unit. 7
Controls are usually accomplished by the use of wet scrub-
bers. The scrubbers utilize the phenomenon of absorption to
remove the fluoride that the processes will emit.
 II. DESIGN OF WET SCRUBBERS

Basic Concepts of Mass Transfer

The object of a wet scrubber is to remove the gaseous pol~

lutant from the air stream. This is accomplished through the

phenomenon of absorption, which is the gas-liquid contacting process
for gas separation, which utilizes the preferential solubility of
the liquid phase.
The mechanism of absorption is illustrated in Figure 1.
Molecular diffusion in a gas is rapid compared to molecular aif-
fusion in a liquid because of the greater space between molecules in
the gas. The more closely spaced molecules of a liquid inhibit
this free motion, and the movement is predominantly by mechanical
mixing which sweeps the absorbed molecules away from the gas-liquid
interface.
The phenomenon of absorption is further explained through
the two-film theory first proposed by Lewis and Whitman. This
model as shown in Figure 2 presumes a mass transfer zone between
phases comprised of two films, one gas and the other liquid. There
is a concentration gradient that exists in the gas and liquid films.
The gradient also exists outside the film because some mass trans-
port occurs in the main body of the gas or liquid phase due to
turbulence.
In the gas phase the concentration of the solute decreases
 8

GAS PHASE LIQUID PHASE

0 0 0 f)f) 0 ¢
0
_,/., 0 ~ I)
~
/'
0¢
0
~
/"' 9) 0~9)
0 .,/' (to 0 ¢ tJ
0
~
~ t1 otlO
~ 00 0
0 ,¢,
~
0 ¢ g8
0 CARRIER GAS f/1 SOLUTE 0 LIQUID

Fig. 1. Gas to Liquid Transfer Phenomenon

Source: William F. Todd, 11 Absorption in Control of
Gaseous Pollutants, Manual 415 for Air Pollution Training
Institute (Research Triangle Park, North Carolina:
Environmental Protection Agency, 1973), p. 1, section 5.
 9

INTERFACE
+
GAS PHASE LIQUID PHASE
g ___....... Z1

t1Pa

~Ca

Pa*

DIRECTION OF MASS TRANSFER

DISTANCE, Z

Fig. 2. The Two-Film Theory of Interphase Diffusion

Source: W.K. Lewis and W.G. Whitman, nAbsorption
Symposium, 11 Ind. Engr. Chern. 16 (1924): 1215-1236 .
 10

from Pag (mean concentration in the bulk of the gas) to Pai

(concentration of the gas at the interface). Resistance to dif-
fusion across the interface is disregarded in the Lewis-Whitman
model.
Mass transfer represents a flow of material from one phase
to another across the interface. This flow encounters resistance,
and a force is required to initiate and maintain the flow. This
11
dri vi ng force 11 is represented by mass flow rate (Na). An expres-
sion for mass transfer of solute A from the bulk of the moving gas
stream toward the liquid-gas interface is given by:B
Na = Kg(Pag- Pai) + Ky(Yag- Yai)

Where Kg is the gas phase mass transfer coefficient based

on partial pressure; Pag is the partial pressure of A in the bulk of
the gas phase; Pai is the partial pressure of A in the gas phase at
the interface; Ky is the gas phase mass transfer; Yag is the mole
fraction of A in the bulk of the gas phase; and Yai is the mole
fraction of A in the phase at the interface.
Resistance to mass transfer can either be in the gas phase,
liquid phase, or both. When the resistance is primarily in the gas
phase, which characterizes the majority of absorption problems in
air pollution,9 the situation is said to be ngas film controlling.~~

Theoretical Design of a Packed Absorption Tower

A packed tower is a tower that is filled with one of many
available packing materials. The packing is designed so as to
expose a large surface area. When this packing surface is wet by
 11
the solvent, it presents a large area of liquid film for contacting
the solute gas.
Usually the flow through a packed column is countercurrent,
with the liquid introduced at the top to trickle down through the
packing while gas is introduced at the bottom to pass upward
through the packing. This results in highest possible efficiency,
since, as the solute concentration in the gas stream decreases as
it rises through the tower, there is constantly fresh~r solvent
available for contact. This gives maximum driving force for the
diffusion process throughout the entire column.
The general design procedure for a packed tower consists of
a number of steps which one must examine. These include:
1. Solvent selection
2. Mass balance
3. Equilibrium data evaluation (liquid flow rate and number of
transfer units)
4. Calculation of column diameter
5. Estimation of column height
6. Determination of pressure drop through the column

Solvent Selection
The first design consideration is the determination of the
scrubbing liquid. The type of solvent depends primarily on the
pollutant gas. Water is the most common and usually the least
expensive. Some acid gases may require a base liquid such as
limewater or some base gases may require an acid wash such as
 12
weak phosphoric acid.
Mass transfer involves the transfer of pollutant gas to a
liquid. This contaminated liquid must be disposed of, and there-
fore, may present the controlling factor in solvent selection.

Mass Balance
Figure 3 is a general schematic of an absorption tower
indicating the gas and liquid flow. The input value of the pol-
lutant gas, v1 , is combined with the total gas flow, G, to produce
inlet concentration. The output pollutant, Y2 , mixed with the air
flow produces the tower exit concentrations. The liquid input sol-
vent flow will contain pollutant amount, x1. The liquid will ab-
sorb the pollutant species and exit the tower as x2.

Equilibrium Data Evaluation

For a given gas-liquid system, with constant temperature
and variable gas partial pressure, the gas concentration in the
liquid changes to an equilibrium concentration at each partial
pressure. If the system consists of a soluble gas to be removed,
an insoluble carrier gas, and a solvent, then as the amount of
soluble gas in the system increases, the equilibrium concentration
of the soluble gas in the liquid increases but not proportionally.
Figure 4 is a typical equilibrium curve. The Y axis is
the partial pressure of the gas while the X axis is a liquid mole
ratio of pollutant concentration in solvent.
The operating line of the packed tower must be above the
 13

AIR AND DECREASED

POLLUTANT OUT
-- G2, Y2

lf:" CROSS SECTIONAL

AREA
- -------
dZ
-- --- ---
PACKING

4~

LIQUID AND AIR AND

POLLUTANT OUT POL LUTANT IN
Ll, xl -
~

G1' Y1

Fig. 3. General Schematic of an Absorption Tower

Source: Kenneth Wark and Cecil F. Warner, Air
Pollution, Its Origin and Control (New York~ Harper and
Row, 1976), p. 265.
 14

~- _._._ ... ---- ..-.- .... -

~ ...... -- .. ---

OPERATING LINE
SOLUTE (F)
CONCENTRATION
IN GAS

Yz EQUILIBRIUM CURVE

Xz
SOLUTE (F) CONCENTRATION IN LIQUID
Fig. 4. Typical Equilibrium Curve and Operating line
Source: Kenneth Wark and Cecil F. Warner, Air
Pollution, Its Origin and Control (New York: Harper---
and Row, 1976), p. 268.
 15
equilibrium curve. The slope of this line may be determined by the
following equation:lO
Lm ~Y
"Gm=!;:[
If the slope of the operating line is known by the inlet-
outlet concentrations with a minimum operation assumed, then the
liquid flow rate can be estimated.
A transfer unit is a measure of the difference 1f the mass
transfer operation, and is a function of the solute gas in the gas
and liquid streams.
A number of transfer units needed in a system may be deter-
mined graphically or mathematically. The graphical method entails
drawing horizontal and vertical lines step wise between the oper-
11 11

ating line and the equilibrium line.

Mathematically the number of transfer units, NTU, is approxi-
mately equal to the total change in concentration divided by the
average driving force.
If an assumption is made that the equilibrium line is
straight, the following equation can be used:ll
yl - Y*
NTU = - - -
y2 - Y*
Y* = equilibrium concentration
The assumption that the equilibrium line is straight is not
valid in most situations and the curvature correction factor ucFu
is given by Buonicore and Theodore.12
 16
Column Diameter
The area of packed section is based upon the flooding cor-
relation as indicated in Figure 5. The value of the superficial
gas mass flow rate, G', may be determined by use of this graph.
The packing factor, F, may be found in most texts dealing with
tower design or from the packing manufacturer.
Once the superficial gas mass flow rate has been calculated,
a simple ratio of G over G' will produce the required column
diameter.
Column Height
The relationship for the height of (Z) of the column may be
expressed as:l3
Z = Hog x NTU
In this equation Hog is the height of a transfer unit. It
is generally a combination of the height for gas transfer, (Hg},
and liquid transfer units (Hl). These use experimentally derived
factors based on the type of packing and the gas and liquid flow
rates as indicated below:l4

and
L'
Hl = ~(----)n (Sc)0.5
~L

where
a, s, y, ~' and n are packing constant found in most
 17

0.4
PRESSURE DROP-
INCHES H20/FT
....__
r---........
1.5 I~ ~ .............
FLC OD NG LINE
c?l .....J
a.
f:::::: ~~
... - ~
N
. V,...;
. ............._ '~
0
- ~

- "' ~~
.....J .....J
:J. a.
(..!)
a.

~
u.. u

--
C)
N 0.1 ~

- ~
"""' '\
" '""' '\ ~
(.!3

;\~

.002

0.01
"\ \ I'\.

\
10
L 0. 5
G (p /P )
G L

Fig. 5. Generalized Pressure Drop Correlation

Source: Kenneth Wark and Cecil F. Warner) Air
Pollution, Its Origin and Control (New York~ Harper and
Row, 1976), p. 270.
 18
texts on tower design or from packing manufacturers
~L = viscosity of liquid, lb/hr ft
L' =superficial liquid flow rate, lb/hr/ft2
Sc = Schmidt number
The terms, Hg and Hl, are combined in the following
equation to produce the height of a transfer unit.lS
mGm
Hog = Hg + ( - ) Hl
Lm
where
m = slope of the equilibrium curve
Gm = gas rate, lb-moles/hr

Lm = liquid rate, lb-moles/hr

Pressure Drop Through Packing

The pressure drop through the packing may be determined
by the following empirical correlation:l6
t.i = m{lo-8) (lonl'/PL) (G')2
PG
where
AP = pressure drop per foot of packing
z
m and n = packing constants found in texts or obtained from
packing manufacturers
pl, PG = density of liquid and gas

Practical Considerations for

Fluoride Wet Scrubbers
The removal of gasious fluorides in wet scrubbers in the
 19
phosphate industry is an absorption process in which the gaseous
fluorides, (HF or SiF 4), in the vent air from the various processes
are dissolved in a liquid. In the typical situation the scrubbing
solvent is a water solution containing weak fluosilicic and phos-
phoric acids plus non-volatile dissolved inorganic compounds and
is commonly called contaminated pond water.
Fluoride gaseous concentrations in equilibrium with the
pond water are extremely low and on the order of 5 x lo-6 lb moles
F- per lb moles air. Figures 6 and 7 are typical equilibrium
curves for fluoride in fluosilicic acid solutions (H 2SiF 4). It
must be noted, however, that each phosphate processing plant will
have its own equilibrium curve for each individual pond. It should
also be mentioned that vapor-liquid equilibrium data may tend to
indicate that scrubbing below acceptable regulatory levels is
impossible. This is not true in actual practice. Based on stack
sampling data at the New Wales Processing Plant in Polk County, it
was concluded that actual scrubbing can lower emission levels below
the equilibrium point.17 This is most likely the result of the
fluorine in the contaminated water pond not existing in a completely
soluble state but may be largely suspended in the water phase. A
suspended solid will not contribute to vapor pressure.
Figure 8 is a typical wet scrubber used in the phosphate
industry. It is of the cross-flow design and is very versatile.
It can be used for tail-gas scrubbing of phosphoric acid, diammonium
phosphate, monoammonium phosphate, granular triple superphosphate,
 20

- 45°C

0.5 1.0 1.5

% F IN FLUOSILICIC ACID
Fig. 6. Equilibrium Curve for Low Concentration of F
Source: Gordon F. Palm, 11 Characteristics, Theory of Design,
and Expected Performance of Gaseous Fluoride Scrubbers Used in the
Phosphate Industry, .. letter written to the Florida Department of
Environmental Regulation, 20 February 1978. (Mimeographed.)
 21

10

- 0')
:I:
9

~- 8
-
L\J
C..'
........
a::
7
0
:::>
_J
u.
c( 6
,.....
0:::

UJ
1-
z 5
0
u
........
_J
........
V)
4
u.
0
UJ
ex: 3
:::>
V)
V)
UJ
0:::
a_ 2
_J
c:x:
.........
1--
0:::
ex: 1
a_

0 10 20 30 40

% FLUOSILICIC ACID
Fig. 7. Equilibrium Curve for High Concentration of F
Source: A.L. Whynes, "Some Aspects of the Absorption
of Silicon Tetrafluoride Gas in Water," Trans. Instn. Chern.
Engr. 34 (1956): 118.
 POND
WATER

FUMES

+ SPRAY
CHAMBER
~
CLEAN AIR
~ COOLING
SPRAYS

SEALED POND WATER DRAIN

fig. 8. Typical Cross-Flow Scrubber
Source: Craig A. Pflaum, 11 Practical Design of Cross-Flow Scrubbers in
the Phosphate Industry," paper presented at the Fertilizer Institute
Environmental Symposium~ 1978. (Mimeographed.)

N
N
 23
GTSP storage, and animal feed ingredients plant exhaust gases.
Tail-gas scrubbers in the phosphate industry are generally
designed to remove fluorine, traces of ammonia, particulate mist,
and water vapor. Non-dusty air enters the scrubber at one end and
discharges at the other end. The air is contacted stage-wise with
acidic pond water from the plant•s contaminated water pond. A
series of sprays is used to cool the fumes, condense the wat~r

vapor and begin the scrubbing operation. The final scrubbing takes
place in the packed section with pond water entering at the top and
on the face. The air leaving the wet packed section is generally
90°- 120° F, saturated. A final demisting pad or packed section
is used to reduce pond water mist entrainment to the atmosphere.
A variation of this design, used on phos acid plant scrubbers, is
to vent low fluorine gases into the scrubber after the spray
section.
Cross-flow scrubbers are generally used where gas absorption
is combined with particulate removal. In cross-flow scrubbers, the
liquid flows vertically down while the gas passes horizontally
through the irrigated packed bed. These scrubbers offer reduced
pressure drops and use lower pump recycle flows than the typical
packed tower under the same inlet conditions.l8
 III. DESIGN OF A PHOSPHORIC ACID PLANT
FLUORIDE CONTROL SYSTEM

Phosphoric Acid Production Process

Phosphoric acid is used in the production of superphosphate,
ammonium phosphate, and mixed grc.nular fertilizers.
The most common method of phosphoric acid production in the
phosphate industry is the dihydrate or "wet process." This is
basically the extraction of phosphoric acid from phosphate rock by
means of adding the rock to s u1fu ric acid, fi 1teri ng the "s 1urry ~"
and concentrating the phosphoric acid to desired strengths. It is
described as the dihydrate process because the gypsum byproduct
that is formed is substantially all in the hydrate form.
There are several variations of the dihydrate process in
current use by the phosphate fertilizer industry. The Dorr-Oliver,
St. Gohain, Prayon, and Chemica processes are all used. There is
little design difference among the different process types. They
consist of three major steps: reaction, filtration, and evapora-
tion. Figure 9 is a process flow diagram for the production of
wet phosphoric acid. The chemical formula for the reaction is
as fo 11 ows : 19
{M = minor miscellaneous elements present in small amounts)
 - - - - - - 1- - 'l - - - - - - - - - 1
I
I
EMISSIONS FILTER GYPSUM

,-- ~
I l SCRUBBER
PHOSPHATE I
ROCK SEAL TANKS
SULFURIC
ACID

j
HOTWELL
EVAPORATOR
REACTOR
F1g. 9. Wet Process Phosphoric Acid Production
~{)~ft~: A~r1&o Ch~m1ca1 Co., "Process Manual." Bartow, Florida. 1969.
 26
H2so 4 + Ca 10 (P0 4)6F2caco 3 + Si0 2 + M20 + FeP0 4 + AlP0 4 + H2o =
Sulfuric
Acid Phosphate Rock

Hydro- Phos Post Insol. Carbon Gypsum

flue- Acid Precipitates silica- Dioxide
silicic Fluorides
Acid
This reaction is the combination of sulfuric acid and phos-
phate rock. This takes place in the digestion system over a period
of approximately eight hours. The reaction itself is very rapid,
however, the proper formation of gypsum crystals takes time.
The desired products are the phosphoric acid and the hydro-
fluosilicic acid. The post precipitates and insoluble silica-
fluorides are undesirable but do not present much of a byproduct
problem. The main problem is to separate the calcium and the
sulfate.
The forming of gypsum crystals and filtrations is the most
practical method of removing the calcium and sulfate ions. The best
filtration is obtained with large gypsum crystals. These are pro-
duced if there is a slight excess of sulfate ions in the crystal~

lizing solution.
The operation to remove the gypsum crystals is carried out
on a series of filter surfaces. Belt or rotary type horizontal
tilting pan filters are superior to other types and are widely used
in almost all new plants.20
The fliter is a continuous circulation unit, horizontally
 27
circling in a counter-clockwise direction. There are usually 24
pans that provide different filtering functions. Dewatering of the
filter cake and washing out the phosphoric acid-rich liquor is the
function of the filter. See Figure 10.
The gypsum is removed and stored on a pile. The weak acid
which is about 30% P2o5 is concentrated to about 54% P2o5. Multi-
stage vacuum evaporators are used for this concentrating.

Sources of Air Pollution

The reactor in wet process phosphoric acid manufacture is
the primary source of emissions. This usually accounts for 90% of
the fluorides entering the control system. The digestion of phos-
phate rock in the reactor is the mechanism that releases both SiF 4
and HF.
The filter is the second most important source of fluoride
emission, since most of the fluorides are emitted where the feed
acid and wash liquor are introduced to the filter.
There are other minor sources of fluoride emissions that
could include vents from sumps, clarifiers, and acid tanks. The
gypsum pond may also evolve fluorides because of the vapor pressure
of the fluoride. The rate of evolution of fluoride will vary with
temperature, concentration, absolute pressure, and exposed area
of the liquid surface.

Design Procedure
The typical pollution control device for a phosphoric acid
 ACID WASH

NO. 2 FILTRATE

CELL DRYING
NO.

Fig. 10. Bird-Prayon Tilting Pan Vacuum Filter

11
Source: Agrico Chemical Co.~ Agrico Welcomes You, .. Bartow, Florida,
1971. {Mimeographed.)
N
(X)
 29

production facility is a cross-flow wet scrubber similar to the one

shown in Figure 8. The design of the scrubber will consist of:
1. Sizing the fume collection hoods
2. Sizing the ducting system
3. Material balance
4. Determining the amount of water needed for cooling in the
spray chamber
5. Determining the amount of mass transfer accomplished in the
spray chamber
6. Determining the number of transfer units needed in the packing
7. Calculating the area of the packing
8. Calculating the depth of the packing
9. Determining the total pressure drop in the entire system
10. Specifying the fan, stack and construction details
The emphasis will be primarily on practical design consid-
erations as practiced by the phosphate industry.

Fume Collection Hood

Figure 11 is a layout of a typical phos acid plant in-
dicating the position of the filter and the reactor.
The filter is approximately 60 feet in diameter with 8 foot
wide by 14 foot long tray sections. The venting of the filter will
occur at the point of acid wash. Two filter trays will be vented.
This constitutes a surface area of approximately 224 ft2.
A canopy hood will be used to enclose the vented area. The
hood will be enclosed on two sides and will be 4 feet above the top
 REACTOR
SCRUBBER

Fig. 11. Typical Layout of Phos Acid Plant Indicating

Position of Scrubber

w
0
 31
of the filter. Industrial Ventilation21 gives the following
relationships to determine flow for a canopy hood:
Q = (W + L)HV
where
Q = flow~ CFM
Wand L = sides, ft
H = height of hood above surface, ft
V = velocity, FPM
When a velocity of 200 FPM is assumed and the length of 2
trays is 28 feet, the width of a tray is 8 feet and the height is
4 feet, the flow will be 28,000 ACFM. This design should be for
30,000. Figure 12 indicates the approximate design of the hood.
The ductwork from the filter will have a flow of 30,000
ACFM. A duct diameter of 48 inches will be assumed. This will
produce a velocity of 2,388 FPM.
The reactor vessel will be fully enclosed. Assorted open-
ings in the top of the vessel will allow fumes to escape, there-
fore, it should be vented. The evolution of carbonate and fluoride
within the reactor could be predicted using chemical equilibrium
data. A survey of actual installations indicate that volumes of

between 15,000 and 20,000 ACFM will be sufficient to maintain a

negative pressure within the reactor.
A design velocity of 2,500 FPM within a circular 36 inch

duct will be assumed. This will produce a volume of 17,675 ACFM.

The system connecting the hood to the scrubber will have
 32

1
LU
0
.......
-co
(/)

s...
Cl)
+J
~ ~~ r--
...-
tJ....
s...
0
Cf-
"0
0
0
:I:

»
~
0
s::
co
u

I-
c:::
lLJ N
.
z t- co t-f
0
a:
LL-
--l
.....
u_
N
C')
.
......
u..
 33

four goo elbows and approximately 100 feet of ductwork. Table 2

is the calculations for the friction losses and the required
pressures.
The losses encountered in the short reactor duct will be
assumed to be minimal and, therefore, omitted.

Material Balance
Figure 13 is a diagram of the proposed scrubber indicating
the various known and unknown inputs into the system. The input
concentrations are typical of those found in the industry.
The first step in establishing a material balance will be
to convert all flows into molar concentrations. (Assume molecular
weight of F- to be 19 with a density of 0.062 lb ).
ft3
1. Reactor flow= 17,675 ACFM@ 180° F
17,675 ft3 x 60 minx 0.062 lb = 65,751 lb x 1 = 2,267 lb moles
m 1 hr w nr "29 hr
Input pollutant = 120 lb moles F- x 1 = 6.32 lb moles F-
hr 19 hr
2. Filter flow= 30,000 ACFM@ goo F (with density of 0.072 lb )
ft3
30,000 ft3 x 60 min x 0.072 lb = 129,600 lb x 1 = 4,470 lb moles
ii1Til 1 hr ff3 nr rg- hr
Input pollutant = 30 lb F- x 1 = 1.57 lb moles F-
nr I9 fir
3. Determine input pollutant concentrations
Reactor= 6.32 lb moles x 1 = 2.78 x lo- 3 molesF-
hr-F- 2,267 lb moles moles air
hr-ai r
 34
TABLE 2
PRESSURE DROPS IN FILTER DUCT

D' D

E
8

H G F A
H' Q = 30,000 CFM
----------------------~ F'
Point Restriction Loss, Inches
of Water

A Hood Entry .09

A- B 10' L .01

8' 90 EL .10

C- D 15' L .01

D' 90 EL . 10

E- F 15' L . 01

F' 90 EL .10

G- H 60' L .06

H' 90 EL .10

Sp = Loss + Vp = .58 11 + .36" = .94 11 of H o

2
REACTOR INPUT FILTER INPUT
17~675 ACFM 30,000 ACFM
180o F goo F
120 LB/HR F- 30 LB/HR F-

SPRAY INPUT 90° F

AIR + F-

Fig. 13. Known Inputs into Scrubber

w
01
 36
Fi 1ter = 1. 57 1b mo 1es x .,........,I"W"Z"._...lrT-_~ == 3. 51 x lo-4 rna 1es r
hr-F- 4,470 16 moles moles air
hr-ai r
4. Determine output pollutant concentration. Assume the plant
produces 26 tons of P2o5 per hour using the E.P.A. restriction of
o.o2 lb r
tons P2o5
0.02 lb F- x 26 tons x1 = 0.027 lb moles
tons P2o5 hr P2o5 ~ hr-F-

Design of Spray Chamber

The spray chamber serves the dual purpose of providing
some pollutant gaseous diffusion and cooling the gas stream.
The design of the spray chamber will consist of estimating
the amount of pond water needed to cool the reactor gas and deter-
mining the number of transfer units. The procedure is as follows:
1. Determine operating conditions.
Input Data (See Figure 13)
1) Volume of inlet gas- 17,675 ACFM
2) Moisture of inlet gas, %by volume- 45%
3) Temperature of inlet pond water- goo F

Output Data
1) Temperature of outlet pond water - 115° F

2) Temperature of outlet gas - 110° F

3} Moisture of outlet gas, by volume - 9%
2. Determine amount of input dry gas.
17,675 ft3 x .062 lb x 60 minx (1 - .45) = 36,163 lb dry air
min ft3 hr hr
 37
3. Determine amount of heat needed to be removed. All enthalpy
values are from Perry's Chemical Engineering Handbook.22
ha = enthalpy of dry air
has = enthalpy of saturated air
Inlet ratio of weight of water to weight of air at 180° F is
.45 x 17,675 ft3 ; 50.28 ft3 = 158 ~water vapor
min lb min
158 lb water vapor x 60 min -! 36,163 lb dry air= 26.3% w·=iter vapor
mTn nr hr dry air
Enthalpy of dry inlet gas at 180° F
ha = 43.3 BTU
1b dry air
Enthalpy of saturated inlet gas at 180° F
has = 748.5 x .263 = 298 BTU
.658 saturation ~1b---d~ry---a~i-r
humidity

Total enthalpy = 43.3 + 298 = 342 BTU

1b dry air
4. Determine outlet gas volume. Assume gas is saturated with 9%
moisture.
36,163 lb dry air x 14,359 ft3 = 519,264 ft 3 x 1 =
hr TO fir -:9T
570,620 ft 3 dry air
nr
5. Determine enthalpy of exit air stream.
hs = total enthalpy = 92.34 BTU
1b dry air

6. Difference in input and output enthalpy

342 BTU 92.34 BTU
16 dry air
= 249.66 BTU
.,6..-d-ry_a
... __i_r
1b dry air
 38
7. Enthalpy of pond water.
At outlet temperature 115° F hw = 83 BTU
Tb
At inlet temperature 90° F hw = 53 BTU
Th
~hw = 25 BTU
Tb
8. Amount of heat removed per hour.
249.66 BTU x 36,163 lb 1ry air = 9.028 x 106 BTU
lb dry air hr nr
9. Amount of water needed
9.028 x 106 BTU x 1 x 1 = 722 GPM
hr ~25~BT~U ~50~0--=1::-:-b-
Tb nr
GPM
The spray chamber will primarily be used for cooling of the
gases, however, some mass transfer will take place. The effi-
ciency of fluoride spray chambers was determined by John Craig at
the University of Florida in 1970. Figure 14 indicates the number
of transfer units that can be expected from a ratio of liquid to
gas.
With a design of 800 GPM and approximately 8.05 m standard
cubic feet of gas, a ratio of 100 exists. This would indicate that
the maximum of 3 transfer units would be possible within the spray
chamber. For design purposes only, one transfer unit will be
assumed. This will be done for conservative design practice. It
is common for the sprays to plug up with silica and not function
properly. Additionally, the temperature and F- concentration in
the pond water will vary with time, therefore, the efficiency of
 10
9
8
7
6
5
NTU
4

1~------~~--~--~~~~~~~~--------~--------~
1 2 3 4 5 6 7 8 9 10 20 40
LIQUID/GAS RATIO - GPM/M STD CFM
Fig. 14. Effect of Liquid/Gas Ratio on NTU for Spray Chamber Using Gypsum Pond Water
Source: John M. Craig, "Fluoride Removal from Wet Process Phosphoric Acid
Reactor Gases ... {Ph.D dissertation, University of Florida, 1970), p. 138.
 40
the chamber will change.

Design of Packed Section

Number of Transfer Units
In order to size the packed section, it will first be
necessary to determine the number of transfer units needed to pro-
vide required efficiency.
1. Determine output from spray chamber.
The previously defined relationship of
yl - Y*
NTU = (Ln ) CF
y2 - Y*
is utilized to determine the output from the spray chamber.
v1 = 2.78 x Io- 3 moles F- (input)
moles air
Y2 = Unknown (output)
Y* = 2.0 x 10-6 moles F- (equilibrium)
moles a1r
CF = 1.0 for low concentration values
NTU = 1 from spray chamber
therefore
2.78 X 10- 3 - 2.0 X 10-S
1 = (Ln ) 1.0
v2 - 2.0 x 1o-6
v2 = 1.1 x lo-3 moles F-
moles air
2. Determine reduced flow of air due to temperature change of
180° F to 110° F from reactor.
17,675 ft3 x 570°R x .069 lb x 60 min x 1 = 2,247 lb moles
m:rn 640oR ft3 rnr 29 hr
 41
3. Determine concentration of F- into packed section from reactor.
2,247 lb moles x 1.1 x lo-3 moles F- = 2.47 lb moles F-
hr moles air hr
2.47 lb moles F- from reactor+ 1.57 lb moles F- from filter=
hr hr
4.04 lb moles F- into packed section
hr
There is a combined air flow of 2,247 + 4,470 = 6,717 lb moles air,
hr
therefore, the inlet concentration of F- into the packed section is:
v1 = 4.04 lb moles F- x 1 lb moles air= 6.0 x lo-4 moles F-
hr 6,717 hr moles air
4. Determine NTU.
s.o x Io-4 - 2.0 x 1o-6
NTU = (Ln ) 1.0
4.Io x Io-6- 2.0 x 1o-s
NTU = 5.6
Design for 6 transfer units.

Amount of Scrubbing Liquid Needed

The slope of the equilibrium line as indicated in Figure 6
is approximately .04. The minimum amount of liquid needed can be
determined by:

~ = .04
where
Lm = liquid flow rate
Gm = gas flow rate (6,717 lb males air )
hr
therefore
 42
6,717 lb moles air x .04 = 269 lb moles liquid
hr hr
269 lb moles x 19 = 5,111 lb = 85 lb = approximately 10 ~
hr nr m mln
This figure of 10 GPM represents the theoretical amount of
water needed to perform the necessary mass transfer. The actual
amount of water needed should be a much greater volume due to
severa 1 factors:
1. The method assumes perfect mass transfer. The reaction of H2o
and SiF 4 will result in a silica shell forming on the water film.
This will impede the diffusion process.
2. The formation of silica in the packed section requires water
flushing for removal. Low volumes of water would result in the
plugging of the packing with solids.
3. The lower the amount of water used, the smaller the area of
the packing. A relatively small packed section coupled with high
volumes of gas would produce high velocities. This would result
in low water-gas contact time. Also, the high velocity would
strip water droplets off the packing, resulting in a demisting
problem.
A survey of several companies utilizing the cross-flow
packed section indicates an approximate water to gas flow ratio
of 15 to 20 GPM per 1000 ACFM.23 Assuming a ratio of 17.5, the
resulting amount of water usage for 40,000 ACFM would be 700 GPM.
High amounts of water usage is not a problem due to the fact that
water is recycled. In the Florida Phosphate Industry the only
 43
cost due to water use is that of pumping.

Area of Packing
The area of the packing is based on the flooding correla-
tion in Figure 5. The relationship of the gas flow rate, G,
divided by the superficial gas flow rate, G', will produce the area.
1. Determine the density of liquid and gas. Assume the density of
the pend water to be that of pure water (62.4 lb/ft3).
For the gas:
PM
PG = lfr
p = 14.7 lb
1ri2
R = 10.73 rsia - ft3
b mole oR
T= 90 + 460 = 550°R
M= 29.0 (Assume molar mass of air)
PG = 14.7~29.0) = .070 lb/ft3
10.7 (570)
2. Determine the ordinate for flooding.
352,800 lb .07 lb Q.5
L PL 0.5
- (-)
lir m .063
G PG = 188,181 lb (62.4 t6i =
nr ft3

From Figure 5, the ordinate for flooding is .16.

3. Calculate the superficial gas flow rate, G'.
1~
11
Assume that the packing material is intalox saddles. The
packing factor 11 F11 is 52. Assume the viscosity to be that of
water, 1 centipose.
 44
(G')2 F (ll L )0.2
• 063 =( ) -PW
gcpGpl
(G' )2 52 (1.0)2
.063 =( ) X 1
32.2 (.07) (62.4)

G' = .659 lb
.....2
-s--f~t

Assume the rate to be 60% flooding

.659 lb X .60 = .396 lb
s-ft2 --s-~ft~2

4. Determine area of packing.

188,181 1b
G nr b~- = 132 ft2
= '%"3~,6--o~o-s_e_c_x-.-==3~96-=---_...,
A= G'
~ sec ft2
This packing dimension of 8 x 16.5 is similar to that found in
actual use.

Width of Packing Bed

The width of the packing in a cross-flow scrubber may be
estimated by using the same procedure as that of a counter current
scrubber.
1. Evaluate the height of a transfer unit in terms of the gas
phase.
a(G')S 0.5
HrG = (Sc)
(L' )Y

Cl = 5.0
8 = .30
y = .50
 45
G = 1,426 1b/hr-ft2
1

L1 = 2,673 lb/hr-ft2
Sc = 1.24 (for SiF4)
5.0 (1,426)-30
HTG = (1.24)0.5
(2,673)·5

HTG = 1.1
2. Evaluate the height of a transfer unit in terms of liquid phase.
HrL may be considered to be zero due to the fact that the
absorption of gaseous fluorides has been shown by Whynes24 to be
gas film controlling. The rate of absorption is controlled by the
rate at which the gaseous fluorides reach the liquid-gas interface.
3. · Detenni ne the height of the packing.

mGm
HoG = HTG + ( ) HTL
Lm
HoG= 1.1 + 0
HoG= 1.1
Z = HoG + NTU
Z = 1.1 X6

z= 6.6 ft
Design for 6 feet which is a common width of packing used in actual
phos acid scrubbers.

Pressure Drop
The pressure drop can be calculated by utilizing the
following equation:
 46
~P- ( -8) (lonLI)'GI'2
! - m 10 - ~
PL PG
m = 5.66 for 1~ intalox saddles
11

n = .00225 for 1 11 intalox saddles

p = 62.4 lb
ft3
PG = .07 lb
1f.3
L1 = 2,673 lb/hr-ft 2
G' = 1,426 lb/hr-ft2
~ = 5.66 (lo-B) (lo·00225 2,673) (1,426)2
62.4 .07
~P = 2.05 lb per ft2 x .192 1n HzO = .39 11
HzO
~ ft lb per ft2 ft packing
For 6 feet of packing there will be a drop of 2.36 inches of H2o.
The normal drop under working conditions should be about 3 inches.
The formation of silica will tend to plug the packing and, there
fore, over design will be necessary.

Design Details
The action of gas passing through a wet packed section will
result in the formation of water droplets. A mist eliminator will
be needed to stop the particles from entering the fan and exiting
the stack. The screen should be capable of removing droplet sizes
greater than 0.5 microns. Several manufacturers produce units
that can remove mist droplets of that size. The pressure drop
should not be greater than 0.5 inches of water for the unit.
The fan selection for the scrubbing unit will consist of a
 47
centrifugal fan that will be required to produce a flow of
48,000 ACFM with a static pressure drop of 5 inches of water.
The materials of scrubber shell construction can be either
rubber-lined steel. FRP, or stainless steel. The most economical
at the present time is rubber-lined steel. All internal piping
and packing support grids should be made of stainless steel.
The length of the scrubber will be 31 feet. The scrubber
will be 3 feet high by 17 feet wide. The packing will have 6 inch
wide supports all around.
Figure 15 indicates the approximate layout of the scrubber.
Four rows of sprays will be used for the spray chamber. Each row
will have 36 sprays for a total of 144.
The packing will have 4 rows of sprays with 6 sprays for

a total of 24. In addition, the face of the packing will have 36

sprays.
The typical phos acid scrubber stack in Florida is approxi-
mately 100 feet high. The material of construction should be
rubber-lined steel. Sampling portholes should be provided at a
minimum distance of 3 diameters from the stack exit.
 FROM FROM DEMISTER
f REACTOR FILTER
g• PACKED TO FAN
SECTION

SPRAYS SIDE VIEW

1
L 31

11
48" DIA.
36 DIA.

17' 0
TOP VIEW

Fig. 15. Phos Acid Scrubber

 SUMMARY

The control of gaseous fluoride emissions from a phosphoric

acid production facility is most commonly controlled by the use of
a cross-flow wet scrubber. The typical fluoride inputs from both
the reactor and filter will be approximately 150 lb/hr F-. A
maximum output of .48 lb/hr F- is needed to meet E.P.A. New Source
Performance Standards. A 99.68% reduction in emissions will be
required.
The design of the scrubber necessary to meet the reduction
consists of the following steps:
1. Size of ductwork
2. Determine the material balance
3. Calculate the liquid flow
4. Determine the packing area and depth
5. Calculate the pressure drop
The scrubber as designed in the text will be 9 feet high,
17 feet wide and 31 feet long. It will utilize approximately 1,500
GPM of pond water for scrubbing and cooling. It will have 792 ft 3
of packing with a total pressure drop of 5 inches of water.
 FOOTNOTES

1Flor1da Phosphate Council3 "Phosphate-Florida's Hidden

Blessing," Lakeland, Florida, 1967.
2Agrico Chemical Company, "A Trip Through Phosphate
Center, U.S. A. , " Bartow, Florida 1974.
3 (Mimeographed. )
3rhe Sulfur Institute, "Phosphatic Ferti 1i zer~ Properties
and Processes," Technical Bulletin, No. 8 (1968).
4L.C. McCabe, ed., Air Pollution, 2nd ed. (New York:
McGraw-Hill Co., 1952), p. 193.
5Report of the Environmental Pollution Panel on Restoring
the ualit of Our Environment, the President's Science Advisory
omm ttee as 1ngton, . . :Government Printing Office, 1965).
6Florida Air Pollution Control Commission, Minutes of the
Florida Air Pollution Control Commission (Tallahassee, Florida:
n. p., 1957-1967).
?Florida, Department of Environmental Regulation, Air
Pollution {Tallahassee, Florida: n.p., 1972), Chapter 17-2.
BKenneth Wark and Cecil F. Warner, Air Pollution, Its
Origin and Control (New York: Harper and Row, 1976), p. 260.
glbid., p. 263.
10Ibid., p. 275.
11 I bi d. , p • 280 .
12Anthony J. Buonicore and Louis Theodore, Industrial
Control Egui~ment for Gaseous Pollutants, 2 vols. (Cleveland:
CRC Press, 1 75), 1:97.
13wark and Warner, p. 285.
14Ibid., p. 284.
l5Ibid.
16rbid., p. 212.
 51
17craig A. Pflaum, ,Practical Design of Cross-Flow
Scrubbers in the Phosphate Industry, .. paper presented at the
Fertilizer Institute Environmental Symposium~ 1978. (Mimeographed.)
18ceilcote Co., Guide to Selection of Air Pollution Control
Equipment, Bulletin 1201 (Berea, Ohio! by the author, 1977).
19 Agri co Chemical Co., 11 Agri co Welcomes You ., 11 Bartow,
Florida, 1971. (Mimeographed.)
20 u.s. Environmental Protection Agency, Inspection Manual
for Enforcement of New Source Performance Standards, Phosphate
Fertilizer Plants~ Stationary Source Enforcement Series Pubn.
EPA 340/1-77-009 (Washington, D.C.: Government Printing Office,
1977), p. 30.
21 committee on Industrial Ventilation of the American
Conference of Governmental Industrial Hygienists, Industrial
Ventilation 14th ed. (Ann Arbor, Michigan: Edwards Brothers,
3
1976), pp. 5-96.
22Robert H. Perry and Cecil H. Chilton, ed., Chemical
Engineers' Handbook, 5th ed. (New York: McGraw-Hill Co., 1918),
pp. 12-7 to 12-9.
23 A.J. Teller 3 "Control of Gaseous Fluoride Emissions,"
Chemical Engineering Progress 63 (1967) ~ 76.
24A.L. Whynes, Some Aspects of the Absorption of Silicon
11

Tetrafluoride Gas in Water, .. Transactions of the Institution of

Chemical Engineers 34 (1956): 117-126.