|
FLUORIDE REMOVAL FROM WET-PROCESS
PHOSPHORIC ACID REACTOR GASES
By
JOHN MUNRO CRAIG
A DISSERTATI1ON PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1970
UNIVERSITY OF FLORIDA\
3 1262 08552 4733
ACKN~OWLEDGMENT:N S
The author wishes to express his appreciation to
the many people and firms who made this project possible.
Gratitude is extended to the members of the
writer's supervisory committee, Dr. R. S. Sholeos,
chairman, Dr, R. 0. McCaldin, Dr. J. E. Singley, and
Special r-ecognlition is due to the many companies
who provided the items necessary for this work. The
following are singled out for m~ention~ing: Tampa El~ectric
Company, Tri-Coat, Agrico Chemical Corporation, Airetron
Engineering Corporation, Amecrican Air Filter Company,
Inc., Peabodyv Engineer-ing Corporation, and the Occidlental
Chemical Company. Special recognition is due Mlr. P.
Flanagan, Occidental Chemical Company, for his efforts
in coordinating thy author' s requirements with the mai~n-
tenance andi operating departments at the plant site.
The author is indebted to Mr. H. McGraw for his
efforts in erecting the research equipment and to Mrs, J.
Hunt for typing this manuscript.
The author wishes to express deep appreciation
to the Department of Healthl, Education, and Welfare,
National Air Polluti~on Control Administration for financing
his graduate studies and to the Agrico Chemical Company
for partial funding of this research project.
The author wishe~s to thanked the many- University
of Florida faculty an~d staff members, whose names are
too numerous to list, for their kind and patient assis-
tance.
In conclusion, the author would like to thank
his wife, Jo, and his children, John, Li~nda, and K~imi
for their patience and understanding during this period
in their lives.
TABLE OF CONTENTS
Pa~e
ACKNPUOWLEDGN~ENTS .................... ............... 11
LIST OF TABLES ...,. ....,,.. ..... .. ....,,.,..... vIii
LIST OF FIGURES ,................... .... ......... ...viin
KiEY TO EUIN ABBREVIATIONS .,,,............... ........ xit
ABSTRACT ..,,.......,.., .. .. ............. ...... xiii
CHAPTER
I, INTRODUCTION ................... ........ 1
II. FROSPHATE INDUSTRY IN FLORIDA .......... 6
Ph~osphpte Fertilizer Mlanufacturing
Processes ................... ....... 6
Geology ................... .... 6
Mining .............. .,. 9
Washing and Flotation ..... 9
Beneficiation ......... 12
Acid Manufacturing ..:............ 12
Fertilizer Manufactulrin ..... 18
Elemental Phosphorous Manufactur-
ing ............,.. .......... 18
Air Pollution Associated With the
Phosphate Industry ................. 21
Fluoride Toxicity ............... 24
Vegetation Damage ............... 25
Cattle Damage ................... 29
Human Health Damage ............ 31
Governmental Controls ........... 34
III, WET-PROCESS BROSPHORIC ACID ....,,...... 37
History ................ .. ........... 37
Prior to 19000 ...............,,,. 37
1900 to 1930 .................... 39
1930 to 1968 ................. ... 41
iv
Page
Fundamenta-l Prinrciples of the D~ihydirate
Process ........ .................. 43
Raw Materials .......,.......... 45
Process Chemistry .......,,..,... 49
Process Technlology ........,.... 52
Fluoride Evolutionl ,............ 57
IV, ABSORPTION THEORIES AND EMPIRICAL
RELATIONSHIPS ........................ 60
Absorption Theories ............. ... 60
General Principles ........,..... 61
Two-Film Theory ,,............... 66
Penetration Theory ............. 68
Surface Renewal Theory ,,......., 68
Film Penetration Theory ......... 69
Absorption with- Chemical Reaction 69
Fluoride Absorption ............... 72
General Design Concepts ......... 72
Absorption in the Cyclonic Spray
Chamber .....,............., ... 79
Absorption in the Venturti .,..... C1
Absorption in the Baffle Plate
Impingement Column ..,,,,. 82
Correlation of Scrubber E-fficien1-
cies ....,,,..,,................ 83
V. TH~E EXPERIMIENT ..,,...................,. 86
Experiment Design ................... 87
Operating Variabl~es ,........... 87
Preliminary Inv~estigation~s ,, 89
Response ,.............- . 90
Replication of Experiments .,,.., 90
Equipment Description ..,.............. 90
Wet-Process Phosphoric Acid Plant 93
Liquid Handling System ........ 95
Gas Handling System ,,.,.. 100
Cyclonic Spray Chamber Scrubber .103
Venturi -Cyclonic Scrubber ..... 105
Impingement Baffle Plate Column 105
Variable Throat Venturi ,,...... 108
Instrum~entation ....,... 109
Experimental and Analytical Procedures 110
Preparation for the Study ....... 110
Paee
Typical Expedi:;cntal Run ...:.... 111
Fluoride Samnp.ing and Anlalysis 114
VJ. RESULTS OF THE~ EXP1ERIMENT ............ 134;
Cy-clonic Spray Chamtber ..... 152
Venturi Cyclonice Scrubber ,. 156i
Baiflfl Plate Imnpingemrent Colulmn 58
Variable Throat Venturt .......16
VIT. CONCLUSIONS AND RECOMMENDATIONS ,.... .. 163
APPENDTX I ............. ..... ....... .,,,... ..... 170
APPENDIX II . . . . ., , .. . .. 18)7
LIST OF REFERENVCES .................. ..,........,,.,, 191:
BIOGRAPHICAL, SKIETCH ............... . ............. 200
LIST OF TABLES
Table Page
1 FLUORIDE EVOLUTION ACCORDING TO SPECHT
AND CALACETO ..... ...,,........,.,.......... 23
2 FLUORIDE EMISSIONS ACCORDING TO HUFFSTUTLER .... 24
3 TYPICAL COMPOSITION OF COMMERCIAL GRADES OF
FLORIDA PHOSPHATE ROCK ................... .... 46
4 TYPICAL COMPOSITION OF OCCIDENTAL BROSPHAITE
ROCK ................ ,,,.,......... ,,.,,...,, 46
5 FLUORIDE BALANCE IN SWIFT & CO. WET-PROCESS
PHOSPHIORIC ACID PLANT .. ................ ..,.. 56
6 LIQUID HANDLING SYSTEM ORIFICES ,...,........ .. 99
7 GAS HANDLING SYSTEM ORIFICES ,............ ... .102
8 OCCIDENTAL LABORATORY ANALYSIS .,..............125
9 STATISTICAL ANALYSIS OF CALIBRATION CURVE DATA .126
10 SPRAY CHAMBER EXPERIMEZNTAL DATA SCRUBBING
LIQUID WELL WATER ...,,........... .,......71
11 SPRA~Y CHAMBER EX'PERIMENTAL DATA SCRUBBING
LIQUID GYPSUM POND WATER ,,. ............ .,173
12 VENITURI EXPERIMENTAL DATA SCRUBBING LIQUID
WELL WATER ,................ ........ ......,175
13 VENTURI EXPERfIMENTAL DATA SCRUBBING LIQUID-
GYIPSUM POND WATER ....,,.,,...,..,... .........177
14 BAFFLE PLATE IMPINGEMENT COLUMN SCRUBBING
LIQUID WELL WATER ..............,....,.......179
15 BAFFLE PLATE IMLPINGEPMET COLUMN SCRUBBINiG
LIQUID GYPSUM! POND WATER, ............... ..181
16 VARIABLE THROAT VENTURI EXPERIMENTAL DATA
SCRUBB1N!G L1QUID WELL WATER ................183
17 VARIABLIE THROAT VENTURI EXPERIMENTAL DATA
SCRUBBING LIQUID -GYPSUM POND WATER .... ...185
V11
LIST OF FIGURES
.gure Page
1 TYPICAL PHIOSPHATE ROCK PROCESSING
FACILITY .. .... ... .... ...............,.,. ]0
2 WASHER AND FLOTATION PLANT FLOW DIAGRAM ...... 11
3 TYPICAL PHOSPHA~TE ROCK BENEFICATION FLOW
DIAGRAM ................... ................. 13
4 TYPICAL ACID-FERTILIZER MANUFACTURING COMPLEX
FLOW DIAGRAM ................... ........... 15
5 TYPICAL CONTACT SULFURIC ACID PROCESS FLOWV
DIAGRAM1 ................... ................. 16
6 TYPICAL WET-PROCESS PHOSPHIORIC ACID FLOW
DIAGRAM .................... ..............._ 17
7 TYP~ICAL TSRIPLE-SUPERP;OSPHIATE FACILITY FLOW
DIAGRAM ................... ................. 19
8 TYPICAL. DIAMMONIUM PHOSPHIATE FACILITY FLOW
DIAGRAMI ................... ................. 20
9 TYPICAL ELEMENTAL PHF~OSPH-ORUS FACILITY FLOWj
DIAGRAM .................................... 22
10 RELATION OF CONCENTRATION AND DURATION OF
EXPOSURE TO EFFECTS OF ATMOSPHERIC FLUORIDE
ONr TOMATO PLANTS .,,...... .,.,............. 28
11 PRECIPITATION AND STABILITY OF CALCIUM SULFATES
IN PH~OSPHORIC ACID ..............,,... ...... 51
12 EQUILIBRIUM DIAGRAM ................... ....... 62
13 DIFFUSION OF A\ THROUGH STAGNANT B ......... 64
14 INTERFACIAL CHARACTERISTICS IN PHYSICAL
ABSORPTION .............. ................. 70
15 INTERFACIAL CHAIRACTERISTICS INU ABSORPTION
WITH CHEMIICAL REACTION .,,,,,.,,..... ...,,.. 71
igure Page
16 VAPOR PRESSURE OF HYDROFLU0RIC ACID OVER
DILUTE AQUEOUS SOLUTIONS ...,,,.. ......,,,. 77
17 VAPOR PRESSURE OF SILICON TETRAFLU0RIDE OVER
AQIUEOUS SOLUTIONS OF FLUOSI.ICIC ACID ... .. 77
18 EQUIPMENT TRAILER FROM ROOF OF ACID PLANT
REACTOR TANK ...,,.......... ............ ... 91
19 SIDE VIEW OF EQUIPMENT TRAILER ............... 92
20 FLOW DIAGRAM OCCIDENTAL WET-PROCESS PHOSPHORIC
ACID PLANT ,....,........,.,,.,,,,,..... 94
21 OCCIDENTAL WET-PROCESS BROSPH1ORIC ACID PLANT
EFFLUENT CONTROL SYSTEM ..,,,.........,,,,... 96
22 SCHEMATIC DIAGRAM OF LIQUID SYSTEM ..,,,,,,... 97
23 LIQUID STORAGE TANK TRAILER ............,,...., 98
24 SCHE:MATIC DIAGRAM OF GA4S HANDLING SYSTEM ..... 101
25 DIAGRAMMATICVIEW OF CYCLONIC SPRAIY CHAMBER ... 104
26 DWIAGRMMATICVIEW OF VENTURI .................. 106
27 DLAGRAMMTICVIEW OF IMPINGEMENT BAFFLE PLATE
COLUMN ..,,,,............. ............ ...... 107
28 FRACTION O7F FREE FLUORIDE AS A FUNCTION OF
SOLUTION pH, WHERE HYDROGEN ION IS THE ONLY
COMPLETING SPECIES ..., ,.118
29 EFFECT OF pH ON ELECTRODE RESPONSE (SODIUM
FLUORIDEC` STANDARD) ....,.... ...118
30 CALIBRATION CURVE USING 2 PERCENT SODIUM
ACETATE SOLUT'ION ,,.........,,... .... 121
31 FLUjORIDE SAMPLING TRAIN ...., ..... 127
32 EFFECT OF LIQUID/GAS RATIO ON FLUORIDE REMOVAL
EFFICIENCY IN SPRAY CHAMBERS, .,,............. 135
33 EFFECT OF TOTAL CONTACTING POWER ON NUMBER OF
TRANSFER UNITS IN SPRAY CHAMIBER ........... 136
34 EFFECT OF LIQUID/GAS RATIO ONJ NUMBER OF
TRANSFER UNITS FOR SPRAY CHAMBER USING WELL
WA1TER ... . . . .- * * * * 37
igur-e Page
35 E;FFE:CT OF` L1QiUiI/GAS RATIO ON NiUNlUFR OF
TJRANSFERP UNIT'S FOR: SPRAY CHAZMBERI USINiG
CYIPSLUM POND WArtTERP .....,,.,...... ... ....... 138
361 EFFECT OF: LIQUID/GArS RATIO1 OlN F'LUORIDE REMOVAL
EFF~ICIEN~CY ;N VENTURI ,.,....... ....... ... 139
37 EFFECT OF TOTAL CONITACTING POWER ON N~UMBER OF'
TR'ANSFERI UNITS IN VE:NTURI .......,,......... 140
38 EFFECT OF: GAS CONTACTING POWER ON NUMBER OF
TRANDSFER UNITS IN VEN~TUiR1 .................. 141
39 EFFECT OF` L1QUID/GAS RATIO ON- NUMBERS OF TRAN\S-
FER UNITS FOR, VEN\TURtI U~SING WEI~l WATER ..,.. 142
40 EFFECT OF LIQUID/GAS RATIO ON NUMB~IER O3F TRANiS-
FER UNITS FOR VENTURI USING GYPSUM POND
WA4TER ............... ..... .................. 143
41 EFFECT OF LIQUID/G-AS RATIO ON FLUORIDE
REMOVAL:I EFFICIENCY IN BAFFLE PLATE: IMPINGE-
MENT COLU~MN ..,,,.,,....... ...... ........., 144
42 EFFECT OF TOTAL CONTRACTING FLOWER OiN NUMBERR OF
TRAINSFERZ UNI1rS IN\ BAFFLE PLATlE INFINUG;lEMENT
COLUMN? ....,,,............................... 145
43 EFFECT OF LIQUID/GAS RATIO ON NUMBlER~ OF TRANS-
FER UNITS FOR: BAFFLE PLATE IMPFINGEMENT COLUiMN
USING WELL WATER .,,,................ ...,.... 146
44 EFFECT OF LIQU1D/GAS PATIO ON NUMBER OF TRA~NS-
FER UNITS FOR, BAFFLE PLATE IMPINGEMENT
COLUMNU USING GYPSUM POND WATER ..........., 147
45 EFFECT OF LIQUID/GAS RATIO ON FLUORIDE REMG~VAL
EFFICIENCY IN VARIABLE ITHROAT VENTURI ...... 148
46 EFFECT OF TOTAL CONTACTING POWER. ON NUMBER OF
TRANSFER UNITS IN VARIABLE T;HROAT VENTURI .. 149
47 EFFECT OF LIQUID/GAS RATIO ON NrUMIBER OF
TRANSFER, UNITS FOR VA~IABL.E THRPOA\T VENTURI
USING WELL WA~TER .............,,. .. ..,,... 150
4;8 EF;E.CT OF LIQUID/GAS RATIO ON NUMBER OF TRANS-
FER UNITS FOCR VARFIABLE TRROAT VENTURI USING
GYPSUMH POND WXTIER ................... ....... 151
Figure P'age
49 COMPARISON OF: TtEl EFFECT OF TOTaL. CONTACT-
ING POW~iER ON NUMBER, OF TRANSFER UNITS IN
EQUIPMENT STUDIED .....,..................... 164
KEY 10 KllN ABDREVIATIONS
BPL bone phosphate of lime~, commonly uised to express
the content of tricalcium phosphate, Ca3(PO )2'
in phosphatre rockr, and used to express the
grade of ore.
ofm gas flow;, cubic feet per minute.
gpm liquid flow, gallons per minute.
HP horsepower.
mg/std cu m milligrams per standards cubic meter
(standard conditions of 250OC and 760 mlm mercury;).
my millivol~ts.
ppb parts per billion.
Abstract of Dissertation Presented to the Graduate Council
in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
FLUORIDE REMOVAL. FROM W:ET-PROCESS
PHOSPHORIC ACID REACTOR GASES
By
John Munro Craig
March, 1970
Chairman: Dr. R. S. Sholtes
Major Department: Environmental Engineering
Due, in part, to the growth of the phosphate
fertilizer industry, fluoride pollution of the air has
become an area of major concern. Since the state of
Florida is the center of this industry in the United
States, air pollution from the phosphate fertilizer
industry is a major problem in certain areas of the state.
Fluoride air pollution is a matter of great concern
since even at very low levels it is known to produce
detrimental effects on vegetation and animals.
Phosphate fertilizer manufacturing complexes
normally have acid plants associated with them which
produce the acid necessary to convert the highly in-
soluble fluorapatite in phosphate rock to a more readily
available form of a nutrient phosphate. One of the most
common methods of producing fertilizer grade phosphoric
acid is via thie wet-process phosphoric acid process.
X111
In~ this process, the reaction between the phosphate rock,
sulfuric acid, and recycled weak phosphoric acid releases
fluoride gases which may be harmful to plants and
animals if not closely controlled.
With this in mind, a research project was int-
tiated to study the effects of operating variables upon
certain types of air cleaning equipment and the efft-
ciency of fluoride removal from wet-process phosphoric
acid reactor gases.
For this study, a portable pilot plant was constyuct-
ed on two semi-trailers, taken to a 650 tons/day wet-
process phosphoric acid plant, and connected into the
center compartment of the reactor tank, The air clean.-
ing equipment used in this study included a 2600 cfm
capacity cyclonic spray chamber, a 2600 ofm capacity
venturi-cyclonic demister scrubber, a 350 ofm capacity
baffle plate impingement column, and a 1000 ofm capacity
variable throat venturi, The effects of liquid and gas
flow rates and type of scrubbing liquid uponl fluoride
removal efficiency were studied. The types of scrubbing
liquid used-were well water and gypsum pond water.
In the case of scrubbing with well water, fluoride
removal efficiencies of 63.2 to 96.1 percent, for less
than 1.5 inches of H20j, and for an energy input of less
than 0.7 horsepower/1000 std efm, are possible using the
cyclonic spray chamber. Efficiencies of 88.3 to 99.5
percent, for less than 15 inches of H20, and for an
xiv
energy input of less than 3.7 horsepower/1000 std efm,
are possible using the venturi-cyclonic demister scrub-
ber. Efficiencies of 90,1 to 99.0 percent, for less
than 8.5 inches of H20 pressure drop, and for an energy
input of less than 1.0 horsepower/1000 std ofm, are
possible using the baffle plate impingement column,
Efficiencies of 92.9 to 97.8 percent, for less than
40.0 inches of H20 pressure drop, and for an energy
input of less than 4.3 horsepower/1000 std efm, are
possible using the variable throat venturi.
In all instances, scrubbing with gypsum pond
water produced lower removal efficiencies than scrubbing
with well water at approximately the same pressure drops
and energy requirements.
The fluoride removal efficiencies developed by
the equipment studied were not sufficiently high
to meet the requirements necessary for emission control
from this process. Two-stage scrubber installations are
required if the fluorides from the reactor are to be
brought under control. Since two-stage scrubber installa-
tions are necessary, gypsum pond water would be the
preferable scrubbing liquid because the use of this
liquid will not disturb the existing water balance
established in the process.
CHAPTER I
INTRODUCTION
The emission of fluorides into the ambient air has
long been a concent of many people and with the growth
of certain industries fluoride pollution of the air has
become more of a problem. The accepted threshold limit
value for human exposure under industrial conditions is
3 ppm; however, atmospheric pollution would involve
concentrations far below this level, Criteria for ambient
fluoride levels are being prepared by the National Air
Pollution Control Administration under provisions of the
Clean Air Act of 1967, The American Industrial Hygiene
Association has recommended the following air quality
levels for gaseous fluorides, expressed as HF by volume:
4.5 ppb average for 12 hours
3.5 ppb average for 24 hours
2.0 ppb average for 1 week
1.0 ppb average for 1 month
There has been no demonstrated case of chronic fluorosis,
or acute poisoning, resulting from humans breathing air
containing these low concentrations of fluorine com-
pounds.1 Although some have suggested that fluorine
compounds may have been the prime agents in the M~euse
Valley and Donora air pollution disasters, no evidence
has been presented to substantiate these clainCs.23,4
Fluoride contaminants may be emitted to the ambie-nt
air in a wide vJariety of industrial processes in which
fluorine compounds are manufactured, utilized, or are
present as impurities in the process materials. The
processes include but are not limited to:
Major sources
1, hosphate fertilizer manufacturing
2. Aluminum manufacturing
3. Steel manufacturingn
Minor sources
1 Brick and tile manufacturlng
2. En~amel fric manufacturing
3. Manufacture of fluorin~e compoundss
4. Power generations;
5, Use of fluorine co~mpounds as catalysts
6, Manufacture of motor fuels
Although the previously mentioned industries have been
recognized as producing damage in some instances, numer-
ous other industries may emit fluoride compounids to some
extent in proce=ssing of mlinerals and ores. ~This is due
to the wide distribution, frequently in small conconl-
tratijons, of fluocrin~e throughout thie earnh's crust.
Recent estimates have- placed flu:orine as the thirteenth
most abundant eleme7nc in thep earth's crust.5
Somnrau and Specht6 consider the mianufacture of
phosphate fertilizers as the most important source of
fluoride contamination of the atmospheree. Since Florida
produces approximately 70 percent of the nation' s phos-
phate rock, air pollution from this industry is a major
problem in certain areas of the state. The objective of
this research was to study fluoride emissions and their
control from one of the processes involved in the phos-
phate fertilizer industry located in Florida and all
discussion from this point on will be limited to this
industry and its manufacturing processes.
The recovery of fluorine effluent, gaseous and
particulate, from the phosphate industrial operations
presents three equally demanding problems. First, level
of recovery; second, operational dependability; and third,
capital investment and operating costs. The recovery of
gaseous fluori~ne compounds is, at present, most econlomical-
ly achieved by absorption in water or 'In other aqueous
solutions. The design of a fluoride recovery system is
not without many problems, some of which are:
1. Multiple contaminants within an effluent
stream.
2. L~ow concentration of inlet. gas and the
extremely low concentration requirement
for effluentc.
3. Limitation in choice of scrubbing medium.
4. Low pressure drop requirements for the
system.
5. Disposal of scrubbing solution,.
In order to optimize the costs involved in con-
structing and mraintaininlg recovery systems, the following
steps are necessary:
1. Determination of optimum scrubbing solu-
tions.
2. DeterminationI of optimum scrubbing equiip-
ment configurations.
3, Determination of optimum liquid-gas ratios.
This research effort dealt with specific gas-
liquid absorption processes utilizing four types of
scrub'oers and two scrutbbing solutions. The types o~f
scrubbers studied were:
1 A~iretron Engineering Corpora-ion modifiedc
low pressure drop venturi.
2. Airetron Engineering Corporation modified
cyclonic spray chamber.
3. Peabody Engineering Company impin:gement
baffle plate column.
4. American Air Filter Company variable
throat venturi,
The scrubbing solutions studied were:
1. G~ypsumn pond water, pH 2, available at
all plant sites.
2. Well water, pH 5 to 7, available at all
planc sites and lowest cost buffer for
gy~psuml pondc %(aters.
This study sought to answerL tlhe following que~stions:
1. ihat is the relationship between the
operat~ing varianbles and removal effi-
ciency?
2. Wdhat are thie mrass transfer characteristics?
3, W~hat ar-e the optimumn Energy r~equijrements??
4. What is the m~axjimuml luoride removal?
The major independent variables studcied for each
systemr were:
1. Scrubbing solution.
2. Li~ui~d and gas flow rates,
CHAPTER II
PRIOSPHA~TE INDUSTRY IN FLORID.
Phos1ht Fer-tilizer Manufactu-ring Procese
Florida's phosphate industry had its start in the
1880's when phosphate pebbles and fossil bones were dis-
covered in the Peae River south of Fort Mede6.7 Early
mining efforts soon were shifted from thle river bed to
near Bartow an~d Mul~berry, wiher land-lock~ed phosphate of
much highor grade could be objtainred at far less cost.
From this smrall beginnings, the ph~osphate indulstrl in
Florida has grown to an industry which at tne: present
time exceeds 40 million long tonls a year or approxrirately
70 percent of the nations output.
The principal use of phosphate rock is in the
production of fertilizer. Other products are phosphorus,
phosphoric acid and ferrophosphate, By-products of the
p~hosphate indiustry- include g~psuml, fluorine compounds,
sulfur dioxide and carbon dioxide.
Geolcry
Tiwo general types of phosph!ate are deposits serve
as sour-ces of high~-grade material. Onep is of ignreous
origins ; ith other sedi;mentar-y, Both havie essentially the
same phosphate mlineral, calciumn phosphate of the apatite
class of minerals.
The igneous deposits are not so extensive in
number or as easily and economically processed as the
sedimentary deposits,8 This type of deposit accounts
for approximately 10 percent of the world's production
from mines located in Russia, South Africa and South
America.
The major phosphate deposits of the world are
sedimentary in origin and nearly all are calcium phlos-
phate. There are two types of sedimentary phosphate
deposits -- guano and pellet, H~utchinson 0has dis-
cussed the guano type in somne detail in his work. The
pellet deposits are the major deposits in the world.
They furnish over 80 per-cent o~f the world's phosphate
demands at the present time. According to Emigh9 they
exist in many parts of the world andi new deposits are
periodically being discovered. A recent discussion
of the geology of these deposits is given by Emigh.9
Recognizing that there is little phosphate in
sea water, past theories have tried to establish how
sufficient quantities of phosphate in solution could
be concentrated in any one place to furnish the large
amounts now present in the phosphate deposits, Some of
the theories that have been postulated are:
1. Accumulation of remains of organisms
due to local destruction of sea life.
2'. Chemijcal precipi~tation..
3. Phosphatizatlion of animal excrecment.
Emigh~ indicates th~e most accepted theory at this date
is the one postulating chemical precipitation.
The phosph~ate deposits in Florida are divided
into three types: land pebble, hard rock and soft rock,
The land pebble phosphlate deposits, centered in Polk
and H-illsborough counties, account for more than 95
percent of the total Florida production. These de-
posits are apparently ancient placer deposits concen-
trated by wave action along a former shoreline or shore-
lines.
The hard rock deposits occur in a narrow belt
extending southward from northern Florida through the
western half of the peninsula and parallel to its axis
for about 100 miles to a point about halfway down the
state. in this type of deposit the phosphate has ap-
parently replaced other types of rock and occurs as
fragmentary rock, boulders, plate rock, pebbles, and
soft phosphatic clay.
Soft rock phosphate production is from the waste
pounds of former hard rock oper~ations and from colloidal
clay deposited at the margins of the land-pebble phos-
phate area. The deposits currently beinge mined yield
phosphate ro-k containing from 65 to SO percent BP:..
BPL denotes bone phosphate of 11me; Ca3(PO4)2, and is used
as an expression of thec grade of deposit.
Minin
In the land pebble area of Florida, phiosphate
is covered over by deposits of quartz sand known as the
overburden ranging in thickness from 5 to 50 feet,
averaging about 15 feet. The phosphate ore-bearing
material known as "matrix" consists of approximately
30 to 60 percent phosphate rock, 15 to 40 percent clay
and colloidal phosphate, and 10 to 25 percent quartz
sand, is found beneath the overburden.
In mining, the overburden is removed by drag-
lines. After thie overburden is stripped off, the matrix
is removed and deposited into a pit where it is sluiced
into a sump. The slurry formed during the sluieinlg
operation is pumped from the sump to the flotation
washer plant, Figure 1 depicits the flow of the phos-
phater rock from the mine through shipment.
Washing and Flotation
At the washing and flotation plant, the slurry
of water, phosphate, quartz sand, and clay passes
through two processing stages, washing and flotation.
Figure 2 describes the flow of rock through the washer
and flotation plant.
Phosphatic material is scrubbed, washed, and
screened. The screening separates the material into
two sizes. less than and larger than 1 mm, Material
not passing a 14-mesh screen comprises pebble rock
10
w
a~L~
x
i-D rJ
S '""wuir ~
Y )P U
Y: Yp
I Lr,
vl
M
5, m
,, o
oI t: o
,P
vlo o
I o
o
"g ~L:
cr
ov,
u
,Y O
uy
8~ a
o X
E~ ~\ ~ u Ci
P ~= PZ
~" ,I
o~n
re ~.I 1~1 ri
wo PO
o
"X r,
L
L~ E~::
a w
i:e ~~
11
~8g
a E
:P or
LI
a eo
ep
a
1 ,:
rr
,, 1
a, cl
a" c
5: m
0-5 ? rr
a
-
S I ~n~-hl I a
~ Ifl f
: I ~B ~o
8 Y C
~ w~ rd
P
r,
P Y ~
1 :ci nl
e
PU~n ~I I
~ 'D~9tll i r,
which is more or less free from quartz sand and clay
when delivered to loadingp bins fo;- shipment. The smaller
material is sent to feed storage bins for further
processing in the flotation plant.
In the flotation plant reagents are used to
separate the quaertz sand from the phosphatric material.
Reagents used in the process are caustic soda, fuel
oil, and a mixture of fatty and resin acids known as
tall oil. The caustic soda cleans the phosphate parti-
cles and regulates the pH so that the tall cil can bJe
properly absorbed. The phosphate rock particles then
take on a film of oil and when agitst~ed float to the
top of the water while the sand sinks to the batcom,
Beneficiation
Various grades of rock are blended on a conveyor
belt to meet specific manufacturing and sales require-
ments. The rock is fed into either rotary or fluidized
bed dryers. Dried rock is then conveyed to grinders
for final sizing before shipment to fertilizer- plants
for the manufacturing of triple supe~rpousphat-e and
phosphatic fecrilizer mixtures, phosphoric acid, or to
produce elemrental phosphoru~s. Figure 3 illustrates
the rock flow in a typical beneficiation facility.
Acid Mlanuactugin
M~ost phospChate imanufacturing facilties have
sulfuric antd phosphocic a-id masnufacturing~ plants
ro~
uzo
~
w> r
,, uz u
~S
associated with them. Figure 4 illustrates the acid
flow in a typical acid-fertilizer manufacturing complex.
Sulfuric acid is manufactured. by the conventional
Contact Sulfuric Acid process.12 In this operation
elemental sulfur is burned in a combustion chamber to
form sulfur dioxide gas, The sulfur dioxide is then
passed through a series of converters, charged with
catalysts; there it interacts with air to form sulfur
trioxide gas, This gas passes on to the absorption
tower, where it interacts with water and weak sulfuric
acid to form the strong product sulfuric acid. Figure
5 is a typical flow diagram of a Contact Sulfuric Acid
process.
In the manufacture of phosphoric acid by the
Wet-Process Phosphoric Acid process,12ground phosphate
rock is mixed with the strong sulfuric acid in a reactor
to form weak phosphoric acid and gypsum. The weak
phosphoric acid is separated from the gypsum in the
filtering step, and the gypsum is pumped to ponds for
storage. The clean phosphoric acid is then concen-
trated to 54 percent P2C05 strength in vacuum evapora-
tors. The 54 percent strength acid is then stored for
use in fertilizer manufacture or for further concen-
tratirng to makte superphosphoric acid. Figure 6 is a
flow diagram depicitin~g a typical phosphoric acid
process.
E
Q
4
8 M
1 a
B "
x
a,
ct
a
o
cj
M
s
u
o
m
1
9
u
N
-
.-i
a,
3
i
O
ri
o
a
x
r
93% ACID COOLERS L _Y
98% D 93% ACID
99% ACID COOLERS L STORAGE TAINKS
COOLING WATER OUT c
Figure 5: Typical Contact Sulfuric Acid Process
Flow Diagram
SWATER FAN
FUFME HOOD~ /t WATER
uasHWITEWISHCA(E DRY CAKE DISCHARG I CEL
Figure 6: Typical Wet-Process Phosphoric Acid
Flow Diagram
Fertilizer Manufacturing
Triple superphosphate (TSP) is manufactured
by reacting phosphate rock with phosphoric acid in a
cone-type continuous mixer. The cone discharges to a
series of enclosed moving belts called the "Den,"
where the mixture completes its chemical reaction and
solidifies. From these belts the triple superphos-
phate is sent to the storage pile, where after further
"curing" it is ready for shipment, At this stage it is
called run-of pile (ROP) triple superphosphate. Figure
7 illustrates a typical triple superphosphate plant,
Granular TSP is made from ROP triple sup3erphos-
phate.,2 In this process ROP with a small amount of
water is rolled in a granulating drum to forml small
round pellets. These are dried in a rotary or fluid-
ized bed dryer, cooled, screened to size and sent to
finished produce' storage.
Diammonium phosphate (DAP) is made by reacting
phosphoric acid and anhydrous ammonia in a reactor,
to form a hot liquid DAP,91 This liquid is pumped to
a granulator where it mixes with recycled material
and solidifies, It is then dried, cooled, screened
to size and sent to storage. Figure 8 illustrates
a typical diamrmonium phosphate plant.
Elemnental Phosohorous M~a~nufacturino
At present there are only two elemental phos-
phorous manufacturing facilities in Florida,6 Elemental
19
= e~~ L~k~~
: .i 2 q~~Y J
m
u "g ,p
WY nj
~ el .i
g
CZ r :s a
54
u. 3 c
5 ~ Z
~ &t: ,~ Y o c,
~r wcr ~
~P "
P~ V) r(U
Or0 r(
B c
s
~
L"
~
? ~ o
or I ~ =
C.
5 % = w
I~
~ r
~n
u j
X
E a
w X
Wp ~U w
~"
oIO I
=~: s
;I
L. OOU t r-
+ cro
uo~ 9 o
F a,
e a d
g .~P
'" 8 = a
,g o
Y
u,
2,
w
E
~ rra pg
O0 rri
uo~ i
P
s
o
u
r,
a,
o,
m
E
1
c
o
9 E
LT LTI
i
D
3
m
o
.i
~u X
1'" "
vim ri
phosphorous is manufactured at those facilities by the
electrothecrmal process.12 In this process the phosphate
rock is mixed with cokec which acts as a reducing agent
and silica which acts as a flux and convoyed to electric
fur-naces. As thle furnace charge is melted the ele-
mental phosphor-us is released from the rock and passes
off as a vapor. The vapor is condensed and collected
as a liquid in pans under water below th-e condenser
pipes. The elemental phosphorus is then stored under
water because of its instability and is ready for uise
or sale. Figure 9 illustrates a typical elemental
phosphorous plant.
Air Pol~lution Associated Wi~jth the Phosrphate indutr
Florida phosphate rock is a non-crystalline
phosphorite consisting principally of fluorapatite
(a complex of tricalcium phosphate and calcium fluoride)
having the approximate formulation 3Ca (PO )2 aF2'13
The tricalcium phosphate is only slightly soluble and
when combined with calcium fluoride is nearly insoluble
in water. The fluorine represents 3.5 to 4.0 percent by
weight of the total compound. In order to make the
phosphate available to plants or animals and at the same
timle non-toxic, the fluoride must be removed.
The release of fluorides from the phosphate rock
is usually accomplished by treatment with heat or acid,
Unfor-tunately, this treatment produces fluorides in the
o
r-1
0
O
.0
o
for-m of H:F, SiF4, and H2;SiF6, In addition to the gases
andc vap-ors released fromT someC of the chem'incal processes,
insoluble dust containing upiwards of 6 percent~ fluoridet
may be dischiargod: from other processing facilities such
as drying, r-i~nding. and materiall handling,
Genierally in the presence of reactive silica
from the rock, fluoride is released as gaseous silicon
tetr~afluor-ide (SjF ), wjh~clh readily is hydrolyzed in
the presence of water to H2SiF6. Silicon tetrafluoride
evolution is not restricted to any one process in m~anu-
facture of fei-tilizer. It is normally given off in t~he
processes of (1) acidulation, (2) concentration by
evaporation, and (3) calcination when~ silica and wat-er
are present.
Specht and Calacetob have reported the fluoride
evolutions from fertilizer processes given in Table 1.
TABLE 1
FLUORIDE EVOLUTION'
ACCORDING; TO SPECHIT AND CALACETO
Pocess _F /_std cu m*;
Acidulation
Batch 10.6 20.6
Continuous 2.1 5.3
Triple superphlosphate
Granular 1.8 3.6
Phosphoric acid
Digestion (Acidul-tion) 0,3 !.2
Superphosphloric 7.0 17.5
a Standard conditions not- stated
Huffstutlerl41 reported fluoride emissions "that
are probably in the form of silicon tetrafluoride" as
noted in Table 2,
TABLE 2
FLUORIDE EMISSIONS
ACCORDING TO HUFFSTUTLER
Process fim F1/ht
Phosphoric acid plant 1 .8 79.5
Phosphoric acid reactor 1 .2 6.3
Run-of-pile triple superphosphate 23.4 331.5
Diammonium phosphate 8.5 39.0
Granular triolQe spephspat 4.2 102.6
Sherwinl6 has reported an evolution of 540 mg
SiF4/std cu ft from a 28 ton per hour superphosphate
manufacturing plant using phosphate rocki fr'om Morocco.
Fluoride Toxicity
Over the years there have been an increasing
number of reports of injury to livestock and vegetation
due to atmospheric pollution by fluorides, The impor-
tance of fluoride as an atmospheric pollutant was
emphasized by a number of investigators at the U. S.
Technical Conference on Air Pollution in 1950.
When present in sufficient concenltrations,
fluorides in gaseous forma are highly toxic to grow-ing
vegetation, humans anld aniimals. The President's Science
Advisory Comitticeel8 in its report to the President
ranked iLnvesti~gations into thle systemic effects oni
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 Conltroll Commission have innumer-
able pages of testimony regarding the damages to vege-
tation, animals and humans caused by fluoride emissions
from the phosphate industry. The International Clean
Air Conference2 meeting in London, Englanid discussed
the damage done to fruit, vegetation, gladiolus, and
cattle in the Florida phosphate belt due to the toxicity
of the fluoride emissions. The President's Science
Advisory Committee21 pointed out the severity of the
fluoride damage inl Southern Florida in its statement
regarding the effects of phosphate fertilizer effluents
on citrus, vegetables, flower crops and cattle. Largent22
has compiled a detailed reviiew of the reported effects
of fluorides, from all sources, on vegetation, animals
and humans.
Vegggtation Damaze
Florida's phosphate belt unfortunately is located
in the same general area that for years has been used for
raising citrus, truck crops and gladiolus among other
vegetation.
The leading agricultural industry in Florida
involves the growing and processing of citrus. Heavy
concentrations of citrus groves are found in the Polk-
Hlillsborough arca within a relatively short distance of
the ph~osphate area, Citrus damnage wh~ich has been at-
tributed to gaseous fluorides includes defoliation of
trees, reduction of yield, reduction of growth, leaf
burn and severe chlorosis.
Citrus trees have formantions of flushes of leaves
periodically through the year which rendersa the plants
continuously susceptible to atmospheric fluoride pollu-
tion. Studies at low concentrations and long periods
of exposure have beenl used to determine leaf injury,
growth, productivity, anld physiological effects in
honcsituations232,52 These studies indicate
that for a mean concentration of atmrospheric fluoride
of 10 mg/cu mn for 2 months to 2 reg/cu m for 12 months
significantr? injury occurs.
Relativelyr high concentrations and shorer exposure-
periods have been used to simulate a fumigation situation.
Acute effects of leaf injury and defoliation have occur-
red at 8,000 mg/cu m; for 15 minutes to 400 mg/cu m
for 2 hours.2 Leonard2 reported on controlled experi-
ments using HF for fumigation of citrus but Taylor28
in testimony before the Florida Air Pollution Control
Commission indicated that "data concerning yield and
leaf-fluoride content is sufficiently scarce and con-
tradictory that LL offers little scientific basis for
establishing a threshold limit of fluoride in foliage."
Weinstin ad Mcue" at the samre hearing indicated
that fluoride would affect plants differently at dif-
ferenit levels due to:
1. A critical stago in the development of the
plant occurs in which yield is most
likely to be affected. TheG stage is
usually during the flowering period.
2. The age of the foliage affects the
sensitivity. Younger foliage is con-
sistenltly more sensitive to fluoride
than older foliage.
3. Dilution by/ growth and losses to the
soil decrease the fluoride level in a
plant with time.
4. Some fluoride in the plant will be con-
vertedl to an inactive starre inI time.
5. Some investigations have shown that
intermittent fumigat-ions m~ay be less
effective than continuous exposures
since plants show a recovery process
during low-fluoride p~eriods.
Winter truck crops such as tomatoes, lettuce,
corn, endive, cabbage and other vegetables are also
produced in the Polk-Hillsboroughi County area. Although
vegetation of this type is fairly resistant to fluoride
damage, damage from flu~ori-de air pollution; has been
alleged more than once.l
A relati~vely large body of data i~s available on
the effects of atmospheric fluoride on tomato plants.
McCune3 has summarized the available~ data in trhe fornl
of Figure 10, Line a represents his criterion with
respect to the presence or absence of foliar markings.
Line b represents the times and concentrations that are
known to have produced economic damage, namely a re-
duction in the quantity or quality of the crop.
LI10 I 0 0
Fiur 10 eain fCnetrto n
Duaino xouet Efcso topei
Floieo omt lns
Beeic3 ha eotd hta on slae
floidues fumig nations of foronths ratio 10 m/cmo
HFshowed no sinficant ue t effect s on th gothof hesec
plants.nToat Pans
M~cCune0 in h;is summanryi of available data oni thle
effects of atmosp~heric fluoride onI corn inldicatfs thlat
the data are scarce on chronic exposuro-s but in the
available data; no significant reductions in growth or
yield were foundc in the limited r-ange of concentrations
and exposure times.
Gladiolus is one of the more susceptible plants
to atmospheric fluoridie damage. McCune30 reports on
the considerable data available on experiments concerning
both chronic and acute fluoride damage. In general,
atmospheric fluoride levels above 6 mg/cu m for 1 day to
1 ag/cu m for 10 days would cause permanent injury to
the gladiolus leaf.
Cattle Dmag
Cattle have been raised in the Polk-Hillsborough
County area for wany years. In recent years veter-
inarians have diagnosed fluorosis in some of the cattle
raised in this area. In most cases, the diagnosis
is chronic fluoride poisoning and not acute poisoning
indicating a long exposure to fluoride.
-Cattle develop fluorosis by feeding on pastures
contamin~ated cy fluorides hence the mode of entry into
the animals is through the digestive sy~stem. U~p to a
certain level fluorides are excreted, and when This level
is exceeded fluoride is then deposited in their bones
due o th afinit offluor-ine for calcium.
In order to combat fluorosis in domestic animals,
the Florida Air Pollution Control Commission3 see the
standard of a maximum of 40 ppm of soluble fluoride
(dry weight basis) for acceptable forage or grass that
is to be used for cattle, They stated that grasses
containing more than 40 ppm F will, "if consistently
used as feed or forage over a substantial period of time,
produce harmful effects." However, Hobbs and Merriman9
have reported that cattle grazing on pastures with up
to 44 ppm F and cattle consuming hay averaging up to
66 ppm F showed nlo significant damage during 10 years
of testing. Hendrickson3 has reported that reduced
milk production, a decrease in the reproduction process,
and a reduction in appetite will result from fluorosis,
Evidence published by the Natioanal Research Council3
indicated that the tolerance level for lactating dairy
cows lies between 30 and 50 ppm F. But Suttie and
Phillips3 have indicated that mature cattle could tole-
rate 50 ppm F for 3 years and show no adverse effects.
Largent22 is his text Fluorosis devotes an entire
chapter on the chronic effects of fluoride intoxication
on animals. In general, the typical effects of chronic
fluoride poisoning on cattle are:
1 Incipient dental changes
2, Staining, with detectable wearing of teeth
3, Decreased consumption of food
4, Loss of body weight
5. Docreanse in calf production
6. Decrease in milk and butter fat production
7. Overgr-owth at joints
Largent also discusses the effects of both
chronic and acute poisoning on rats, sheep, poultry,
swine,, rabbits, and h-ousehold pets.
The toxicity of fluoride appears to be greater
in forage contaminated with gaseous effluents such as
silicon tetrafluorlde rather than wjith phosphate dust.
Thirty to 50 ppm (by weight) of soluble fluoride compounlds
in the total diet of dairy cattle may cause deleterious
effects, while it takes 60 to 100 ppm of insoluble fluo-
rides to produce thel same effect,
38
Lewis, while taking the p~ospha~te industry to
task, brings out an inter-esting statistic regarding th-e
damage that fluorosis can cause. He states that in
1954 Polk County w~as the largest cattle producing
county in Florida with 120,000 head. In 1965, there
were 90,000 head of cattle and the number continues to
decrease and he states "it is not unusual to come
upon cattle crawling across pasturelanid on their knees,
starved from their inability to chew,"
Fluorides that are highly reactive chemically
will be irriLtating to exposed areas of the human body
whe~ln they are present in the ai~r in~ sufficient quantities.
Normally, exposures to such highly reactive fluorides
is limited to ocuainlepsrLess is known
about the effects on humans of irnhaled flucri~des that
are not surface irritants and whose action depends upon
their absorption from the lungs by the blood.
The fact the cattle grPzing in an area have had
health effects does not mean that people living in
the same area would have thle same effects, Cattle
obtain most of their fluoride fromn the forage and
water rather than from the air. The California Stale
Department of Public Health39 has estimated that 1 .0
to 1 .2 milligrams per day of fluoride is the optimurr
ingestion for the control of cavities in children.
The mean concentration of fluoride found in the
air of communities ranges fromr 0.0031 ppm in Charlestown
to 0.018 ppm in Baltimore.0 The maximum concentration
in most cities is about 0.025 ppm with a high of 0.08
ppm being reported in Baltimore, Thomas1 reported
a value of 0.29 pp~ i~n the vicinlity of a superphos-
phate plant.
Largent424 has observed the followiing effects
when hydrogen fluoride was inhaled by man under labora-
tory conditions:
1 At concentrations less than 5 ppm, no
local and immiediate systemic effects.
A sizable portion of t~he inhaled dose
was promptly ex-cr-eted.
2. At 10 ppm, many~ exposed per-sons com-
plained of discomfort.
3. At 30 ppmn, all exposed persons com-
plained about environment.
4. At 60 ppm, definite irritation of con-
junction anld nasal passages and diis-
comlfort of pharynx and trachea.
Derryborry44 reported on a study of a large group
of men exposed to a phosphate fertilizer plant for
periods of up to 25 years and the few suggestions of
toxicity "were not found to be formidable in light of
the fact that other factors may have been responsiblee"
A Florida State Board of H~ealth study of school
children in Polk County found no Evidence of fluorosis
in their teeth. The same conclusion wa~s reached by
the Bureau of Dental Health and the National Institute
of Dental Research after they toured schools in Polk
County,45
While individual cases of temporary illness have
been well documented by physicians and civic organiza-
!9
tions^ in the Polk-Hillsborough area, it appears that
to date there have been no permanent human defects
directly attributable to the phosphlate industry in this
ara34,46
Gover~nentaol Cnrl
Perhaps the occurance of illnesses alleged to be
caused by the emissions from the phosphate processing
industry gave the impetus to the formation of the
Polk-Hillsborough Air Pollution Control District.
Perhaps it was the economic loss to the area such as
$1 ,000,000 damage to new citrus groves, loss of 1 ,500,000
boxes of fruit per year, or a reduction of citrus
property value of $20,000,000 in the Polk-Hillsborough
Area.4 Perhaps, it was the decline in cattle produc-
tion in the two county area.
For whatever thle reason or combination of reasons,
the Polk Air Pollution Control District was formed in
March 1958 to handle complaints regarding air pollution,
mainly from the phosphate industry, in the county.
The district was formed under the 1957 Florida Air
Pollution Control Act which established an air pollu-
tion control commission in the State Board of Health.
The district was created with the understanding that
it would permit an "orderly study of air pollution in
the county, and that members of agricultural, live-
stock, industrial, and political interests would be
considered and the cooperation of those interests
encouraged."
Hillsborough~ County Air Pollution Control Dis-
trict was formed in July, 1959.8 The two control dis-
tricts were merged into the Polk-Hillsborough Air
Pollutioni Control District in June, 1960.9 It was
hoped that the merger would provide greater efficiencies
and better control for the areas involved.
The Polk-Hillsborougeh Control District did not
live up to its expectations due to a number of reasons
including: lack of proper legal counsel; delays in
taking lawful action due to the procedure of conference,
conciliation, and persuasion; inadequate funding; a
defeatist acttiude where the commission felt "hamstrung"50
due to not enough direction, responsibility, and authority
from the legislature regarding its desires; pressures
from! the industry; and lack of adequate anld properly
trained per~sonnlel
In 1967, the enabling legislation for the act
forming this district was repealed and the F'lorida Air
and Water Pollution Control Act was passed. One of the
purposes of this act was to strengthen the pollution
control agencies in the state. The act had been enacted
with the hopes of eliminating some of the major problems
associated with Florida Air Pollution Act. It gave the
control commission powers which if used properly would
provide the incentive for positive action in controlling
air polluting industries.
On June 30, 1969 the Florida Air Pollution Control
Commission was out due to governmental reorganization.
On July 1, 1969 the Florida Department of Air and Water
Pollution. Control. was created and it is under this
36
department that all present surveillance of the phas-
phate industry takes place. T'he requirements of the
department regarding fluoride emissions are "the unit
emissions of fluoride expressed as pounds of fluoride
per ton of P 05 or equivalent produced, shall not exceed
0.4 poundss" taking into consideration the latest
technology, existing pollution levels, the lowest value
attained by other plants, and location. This require-
ment necessitates that some method of fluoride recovery
be utilized.
CHAPTER III
WET-PROCESS PHOSPHORIC ACID
Mlost wet-processi phosphoric acid produced at the
present time is: used in fertilizer production since the
acid meets the critical requirements of cost and grade.
Until quite recently, the general availability of sul-
furic acid gave the wer-process acid a favorable economic
position and since the fertilizer industry can. use a
relatively impure phosphori- alcid, the wet-process
phosphoric acid process is used in Florida instead of
the furnace process. This chapter will be confined to
the process of producing fertilizer grade, orthophosphorcl
acid (H3O
Hi smy
Prior to 1900
The development of wet-process phosphoric acid
manufacture is closely related with the early production
of chemical fertilizers.
Chinese farmers are said to have used calcined
bones somo 2,000 years ago.5 and Fritish5 has stated
that bones have been applied for centuries to the vine-
yards In southern F'rance. Guan~o, a fertilizing material
consisting almost wholly of the extcreta of sea birds,
was used by the Incas as early as 200 B.C. Bird guano,
bones, and other organic substances formed the basis
for the early fertilizer industry and during thle 19th
century the use of these materials was limited only by
their availability.53
Brandt prepared elemental phosphorus from urine
in 1669, and Boyle made phosphoric acid from phosphorus
in 1698, Gahn in 1769 first associated these materials
with bones. Scheele in 1775 prepared elemental phos-
phorus from bones by treatment with mineral acids and
reduction with heated charcoal, In 1840, phosphorus
was recognized as the major component of bone manure.
Justus von Liebig,in 1840, proposed that bones be dis-
solved in sulfuric acid to make the phosphorus content
more available to crops." J. B. Lawes was issued a
patent in 1842 for making superphosphate from "bones
and other phosphoritic substances" and sulfuric acid by
"setting free such phosphoric acid as will hold in solu-
tion the undecomrposed phosphate of lime," Lawes modified
his patent, and confined his claim to apatite, phos-
phorite;, and other substances containing phlosphoric
acid. Laws essentially founded the mineral superphos-
phate industry.
In 1851 Albright and Wilsonl Ltd., at present
England's largest producer of phosphoric acid, w~as
formed at Oldbury, England.55 In 1870, a plant was
established in Germany to manufacture an improved grade
of supertphosphatet by reactinge low-grade rock with phos-
pho-ic acid in place of the sulfuric acid that had been
used to date. By 1900, at leaste 12 Europ~ean companies
were manufacturing phosphoric acidl for fertilizer
process use by batch oper-ations from low-grade ores.56
Commercial fertilizer operations in the United
States predate those of Europe. William Davidson
erected the first sulfuric acid plant in 1832 to acid-
ulate bones and oyster shells.57 The first triple
superphosphate plant was built by the American Phosphate
and Chemical Company in 1890 at Ba(ltimore, Maryland.
This plant was built using knowledge gained in 1867
when phosphate rock from South Carolina was used in the
acidulatio.: process in place of bones and oyster shells.58
1900 to 1930
In the United States wet-process phosphoric acid
production was insignificant prior to 1900, and the total
production, w-hich1 was mainly concentrated in the Blaltimore
area, did not exceed 2000 tons/year. The leaders in
the Industry in the period from 1900 to 1915 were Stauffer
Chemic-al Company, Virginia-Carolina Company (now a
division of M~obil Oil Company) and American Agricultural
Chemical Company (nowj a division of Continental Oil
Company). These companies and many other smaller con-
cerns were acidulating bones to mak~e the acid In! the
1900 to 1905 period, and then sw~itchedd over to the use
of phosphate rock inl the 1905 to 1915 period. The process
was a batch operation in which digestion of low-grade
rock took place in wooden tanks of 1 to 2 tons/batch
and the resulting slurry was filtered on lead-lined,
wooden filtering pans,59 The filtrate, containing 10
to 15 percent P205, was concentrated in rock or lead-
lined pans.
In 1915, the continuous process was introduced
into thie phosphoric acid industry by the Dorr Company.
By 1924, this process made possible an acid cont~aining
22 to 23 percent P 05, as compared with the 10 to 15
percent with the batch process, Briefly, the process
consisted of the reaction of a phosphate rock.-phosphoric
acid mixture in a series of primary digestion tanks
with sulfuric acid (66.6 or 77.7 percent H2SO4). Agita-
tion of the slurry produced was maintained and after
the slurry left the last tank it was washed counter-
currently in thickeners. Thle overflow from the primary
thickener normally analyzed 22 to 23 percent P 05'
The calcium sulfate dihydrate which settled in the
primary thickener was washed in the remaining thickeners
to remove th~e entrained acid.6
In the period 1916 to 1929, Dorr built strong
acid plants for some 31 companies in the United States
and Europe. The smallest of those plants processed 25
tons of rock per day. During this same period, the
Chemical Colnstruction Company (Chemico) built wet phos-
phoric acid plants for U. S. Export Chemical Corporation
(now U. S. Phosphloric Chemical Corporation) and American
Cyanamid.
1930 to 1968
During th~e 1930 to 1940 period, Nordengren and
associates in Sweden developed strong acid processes
that produced a calcium sulfate hemibydrate and a very
strong acid, 40 to 50 percent P205. However, after exper-
ience was gained with this process, the conclusion was
reached that the economies in producing the strong acid
did not justify the complications involved in the fil-
tration of the slurry.61 However, Nordengren and his
associates performed basic studies to determine the
conditions under which the anhydrite (CaSO ), hemnihy-
'drate (CaSO4 \HI20), and dehydrate (CaSO4 2H20) were
formed in phosphoric acid slurries and established the
relationship between crystal filterability and acid
concentrations for each of these crystal forms, Dorr
and other companies in the United States were also
working on producing an acid in the 40 to 50 percent
P205 rangei however, due to the same type of problems
encountered by Nordengren, they were not able to produce
a high strength acid economically.
During the same period, Dorr built the largest
wet phosphoric acid plant up to that time for Consolidated
Mining and Smelting Company at Trail. British Columbia.
The plant consisted of three trains, each having a
processing capacity of 150 tons/day of phosphate rock.
This plant was probably the major contribution of Dorr
Company during this period since it involved (1) the
recycle of gypsum slurry for better control of crystal
growth, thus giving larger crystals and improved filter-
ability, and (2) the separation and countercurrent
washing of the gypsum on continuous filters. The product
acid from the filters contained 30 to 32 percent P205'
This process gained the name of the Dorr Strong Acid
Process since it produced the highest strength commer-
cial acid to that date. The acid produced was already
strong enough for the production of ammonium or sodium
phosphates but it was not used in the manufacture of
triple superphosphate or food grade calcium phosphate
since a concentrated acid was required for these products.
This process produced a dibydrate gypsum and was the
one used for many other plants throughout the world,
In the 1940 to 1950 period, processes yielding
calcium sulfate anhydrite were studied by Davison
Chemical Company; however, none of these processes were
ever commercialized.6 Tennessee Valley Authority
(TVAC) also began work on a process producing an anhy-
drite at their Muscle Shoals, Alabama facility.
-In the 1950 to 1968 period, interest was renewed
in the hemihydrate process of Nordengen and commercial
installations have been built in Japan. Better recovery
of P205 has beenl obtained and the by-productgypssum is
of better quality in the hemihydr-ate process, A~n acid
containing 40 percent or more of P205 has been obtained
using this process in large-scale tests.63
During this same period, due to a shortage of sulfur,
interest was developed in substitution of hydrochloric
acid for sulfuric acid. The Israel Mining Company devel-
oped a process using solvent extraction to separate the
phosphoric acid from the calcium chloride formed. Small
commercial plants have been built in Israel and Japan
using this process, Other companies have used nitric
acid as a replacement for sulfuric acid.6
The major process used at the present time is: the
dibydrate process and most large commercial plants con-
structed today utilize this process with the latest im-
provements which include (1) slurry recycle, (2) single-
tank reactor, and (3) tilting-pan filter. The capacity
of the largest single-train plant in Florida is a nominal
1500 tons P205/day. According to Slack,6 as of 1968
this plant is also the largest single-tratn plant in the
world. The Occidental Chemical Corporation plant at
which all experimentation for this study took place had
a nominal 650 tons P205/day capacity.
Fundamental Princioles of the Dihydrate Process
The primary objective in the phosphate fertilizer
industry is to convert the fluorapatite in phosphate rock
to a form available to plants. The fluorapatite is
quite insoluble and is not normally used as a fertilizer
in its natural state because it has little value as a
supplier of nutrient phosphate. 'The most widely used
method for making the nutrient phosphate available is
treatment with a mineral acid such as sulfuric, phos-
phoric, nitric, or hydrochloric.
Since phosphoric acid can be readily manufactured
at fertilizer plants, it is normally used to treat
phosphate rock for production of triple superphosphate
and in the reaction with ammonia to make ammonium phos-
phates. The growing importance of both of these products
makes phosphoric acid a very important intermediate in
the fertilizer industry.
Production of phosphoric acid by the wet-process
involves the steps of (1) dissolving phosphate rock in
sulfuric acid, (2) holding the acidulate slurry until the
calcium sulfate crystals grow to adequate size, (3)
separating the acid and calcium sulfate by filtration,
and (4) concentrating the acid to the desired level.
The principal difference in processes for manufacturing
phosphoric acid is the degree of hydration of the cal-
cium sulfate, which can be changed by altering the tem-
perature and the P205 concentration of the acidulate
slurry. The dihydrate, CaSO4 2H20, is precipitated
under conditions of low temperature and concentration
(see Figure 11). Because of these two reasons, the
amount of wat:-prce)Css ci;d produ61cedL by methods other
thlan trec dibydrare process is not significa;nt.6
Raw M~aterints
The two major raw materials in wct-process phos-
phoric acid manufacture are phosph~ate rock and sulfuric
acid, The phosphate rock fed to the process is as high
in grade as economically feasible, usually ranging from
30 to 35 per-cent P205, The sulfur-ic acid normnall.y is
660 Be (93 to 98 percent).
The phosphate rock used exists principally in
the form of a fluorapatite, which is a broad term
given to a relatively complex mnl'reral whose properties
can vary widely, depenlding uponl origin. Slack656
normally uses the cheicf~al! formula Cal0(PO )6'2, in his
descriptions of the pr-ocess, while Wraggamnan61 and
Sauchelli use 3Ca (PO )2 CaF2 in their. descriptions
of, the process. Phosphate rock used for production
of acid by the wet-process varies fromt about 65 to 85
BPL, depending upon the origin and extent of beneficia-
tion. All rocks contain varying amounts of other com-
pounds, either physically mixed or substituted in the
mineral itself. Table 3 gives a typical analysis of
Florida phosphate rock. Table 4 gives a typical analysis
of the rock used by Occidental at their W~h~ie Springs
plant,
TABLE 3
TYPICAL COMPOSITION; OF COMMErRCIAL GRADES
OF FLORIDA PHOSPHATE ROCK62,a
S _pecified BP Ran
68L66A 2/70 77/76
BPL 68.15 72.14 77,12
P205 equiv. 31.18 33.00 35.28
H20 1 .3 1 .0 1 .0
Fe203 1.33 1,07 0,84
Al203 1.76 0.83 0,56
Organic 2,18 1,76 1.70
SiO2 9.48 6.46 2.02
CO2 3.48 2.87 2.98
F2 3.60 3.62 3,89
Ca0 45.05 48.10 51.53
SO3 1.05 1.11 0,66
a Dry basis
TABLE 4
TYPICAL COMPOSITION OF OCCIDENTAL
PHOSPHA.TE ROCK"
BPL 74.00
P205 33.80
Fe203 0.49
Al203 1.49
Si02 4.30
CO2 4.72
F2 3.93
Ca0 50.0
SO3 1.32
C 0.28
Mgo) 0.31
TABLE 4 CONTINUED)
Na20 0.69
K20 0.10
a Process Report, P. Flanagan,
9/27/68, Occidental Chemical
Company, White Springs Plant.
While the performance of a given rock cannot
be completely predicated from the chemical analysis,68
it is possible to draw certain general conclusions.
Normally, a preferred rock is the one that yields
phosphoric acid at the lowest cost; therefore, the
primary consideration is the price of the rock per unit
of P205, Other items which may affect production
include:
1. Ca0 to P205 ratio. This ratio eventually
determines the amount of H2SO4 used per
unit of P205 produced. A low Ca0 to P205
ratio is desired since the cost of 112S04
is a major factor in the production cost of
phosphoric acid.
2. Fluorine. Most companies do not consider
this an item of significance, however, high
values may increase the air pollution gene-
rated in the process and may increase plant
corrosion.
3. Iron and a~um~inum. They are major problems
in acid manufacture and are nlormally rem~oved
by beneficiation to a major extent.
Residual amounts dilute the phosphate content ,
may interfere with crystal growth, cause
sludge to form in the product acid, and
in general may cause a poor acid for use in
making fertilizer products.
4, Carbonates. A high carbonate content is
accompanied by a high Ca0 to P205 ratio
reducing the grade of rock (P205 content)
and causing foaming during acidulation.
5, Magnesium and potassium. Normally both
give rise to complexes that lead to sludge
problems when the acid is concentrated.
6, Silica. Silica is partially attacked by the
fluorine released during digestion. Normally,
it is not considered a problem.
7, Sulfur compounds. Sulfur, if present as
sulfates, can usually be credited against
H2SOq consumption.
8. Organic matter. High levels of organic
compounds will cause foaming problems and
tend to reduce filtration rates by blinding
the filter cake.
The rapidity of digestion is influenced by the
particle size of the rock since the attack takes place
primarily on the surface of the particles therefore,
the time required to dissolve a rock particle is pro-
portional to thie diameter so long as other factors are
held constant. Normally, the rock is ground to 60
to 70 percent minus 200 mesh; however, ungroundl phos-
phate rock can also be used in the process.
The sulfuric acid used in the wet-process is 660
Be (93 to 98 percent). The strong acid is employed so
that as much water as possible can be used in washing
the gypsumn filter cake. The amount of sulfuric acid
used in the wet-process is approximately stoichiometri-
cally equal to the CaO in the rock. The amount of
free sulfuric acid in the digestion system is an im-
portant operating variable, since it affects the size
and shape of the gypsum crystals formed.
Process Chemistry
Waggaman and Slack0 have reported the fol-
lowing reactions taking place throughout the process:
Ca3(PO )2 + 3H2SO4--2H3PO4 + 3CaSO (1
CaF2 + H2SO --- CaSO4 + 2HF (2)
CaCO3 + H2SO4--->CO2 + CaSO4 + H20 (3)
6.F + Si02--' H2SiF6 + 2H120 (4)
Na2(or K20) + H2SiF6 --Na2SiF6(or K2SiF6) i20 (5)
H2SiF6 + Heat and/or Acid----SiFq + 2HF (6)
3SiFq + 2H 0--->2H2SiF6 + SiO2 (7)
Fe203(or A12 3) + 2H3PC)4---*2FePO4(or A1PO ) + 3K20 (8)
Phosphiate rock is a very complex mineral; the
principal mineral constituent, fluorapatite, contains
calcium, phosphate, fluoride, carbonate, and other
elemetnts or groups bound together. When the rock is
treated with sulfuric acid, the apat-ite lattice is
destroyed and the phosphate is solubilized as phos-
phoric acid. Th!e principal reaction, Equation 1 is
the reactionI between tricalcium phosphate and sulfuric
acid to give soluble phosphoric acid and insoluble
calcium sulfate. The calcium fluoride reacts with
sulfuric acid to produce hydrogen fluoride and calcium
sulfate, Equation 2. The calcium carbonate is con-
verted to carbon dioxide and calcium sulfate, Equation
3. The entire reaction between the major constituents
and sulfuric acid has been described by Slack0 as:
Cal0(PO ) F2 CaCO3 + 11H2SO~-- -6H3PO4 + 11CaSO4 +
2HF + H20 +CO2 9
The hydrogen fluoride produced may react with
Bte silica present to form silicon tetrafluoride, which
then hydrolyizes to form fluqgsilicic acid as shown in
Equations 4 and 5.
The calcium sulfate formed inl the reactions can
be in three stages of hydration: anhydrite (CaSO4 '
semihydrate (CaSO4 HI20), or dibydrate (CaSO4 2HI20),
depending on the reaction temperature and the phosph~oric
acid concentration (see Figure 11), However, since
the dihydrate occurs with low reactor temperatures and
low P205 concentrations in the acidulate slurry, it has
Hemihydrate
00 Precipitated; Anhydrate
Stable
Dihydrate
80 Precipitated.
Anhydrite Stable
60
40~ Dihydrate
Precipitated and
Stable
20
S10 20 30 40 50
Acid Concentration, percent P205
Figure 11: Precipitation and Stability
of Calcium Sulfates in Phosphoric Acid
been antd still is the favored method of producing wet-
process phosphoric acid.
By increase in temperature or concentration,
conditions can be reached under which dihydrate is pre-
cipiltated but is converted in time to anhydrite (CaSO4)
In1 practice, however, the crystallizati~on is affected
by so many factors that p-rodiction of the result i~s dif-
ficult. Free sulfuric acid changes t-he stability regions
and impurities affect the rate of conversion of one
crystal form to another, Since there are so many factors
affectinlg crystallization, the dihydrate process is
favored due to the lower temperatures and the lower
concentrations. There are many variations in dthydrat~e
processes, mainly in regard to type of filter and in
design anld arrangement of digestion acidulationn)
equipment.
Process Techno~logy
There are many variations in dihydrate processes,
mainly in regard to the type of filter and in the de-
sign and arrangement of the digestion acidulationn)
equipment. The major areas in any phosphoric acid
process are raw material feeding, reactor system, filtra-
tion, and effluent control.
Raw material in the form of phosphate rock and
sulfuric acid is closely controlled, Gravimetric
feeders or automatic batch scales are normally used for
rock feeding. Sulfuric acid, recycled phosphoric acid,
and washwat~er are normally controlled by flowme~ters.
The reactor system varies from process to process.
The most recent type is the single tank reactor with
the necessary cooling, fum~e remo~val, and recycle facili-
ties. Older prucesses use a system of tanks. The
single tank reactor is normally equi~ppedl iith multiple
agitators that promote the reaction. One process has
a single tank with two com~partmentcs, an outer annulus
in which the reaction takes place, and an inner core
where the sIlurry formed in the reaction is held for
cryst-al growth and stabilization. Average retention time
in the reactor system is approximately eight hours.
Foaming in the reactor often is a problem when the rock
has a relatively high content of organic m~atter or car-
bonate. Various types of antifoaming agents are used
to control foaming.
The heat produced by acidulaltion of rock and
dilution of sulfuric acid is normally removed by one
of the three following methods (1) blowing air into
thle slurry, (2) blowing air. across the slurry, and (3)
flash: cooling under vacuum. Vacuum cooling is probably
the most commonly used method at present since the ef-
fluent control system is smaller and less expensive.
Filtration of the acid-gypsum slurry requires
that gypsum crystals of adequate size and shape be
produced in the reactor, Satisfactory crystals depend
upon:
1.Sulfate concentration. Insufficient sulfate
produces thin, platelike crystals that
are difficult to filter. Excessive sulfate
produces crystals that are easy to filter
but are difficult to wash free of phosphate.
The optimum sulfate concentration depends
upon many factors, but normally is approx;-
mately 1.5 to 3.0 percent H2So This
level produces rhownboid crystals agglome-
rated together which filter and wash well.
2. Slurry recirculation. Recirculation of
slurry is required in all systems in order
to (1) reduce the effects on the process of
surges, (2) reduce the effects of localized
high concentrations of rock and acid, and
(3) control the degree of supersaturation
necessary for good crystal growth.
3. Phosphoric acid concentration. Crystals
decrease rapidly in size as the acid con-
centration is increased above 32 percent
P205. This is one of the major limiting
factors on product acid concentration in
the dihydrate process.
4, Rock impurities, Slack7 reports that this
area is not understood at present but is
currently being studied.
Normally filtration occurs in three stages.
The slurry, containing 30 to 45 percent solids, is
filtered in the first stage to produce the product
acid. Wa~sh water Is added to the last stage to wash
traces of acid from the gypsum cakie. Filtrate from
this staeo is use~d as cash- water for the second stage?,
wlhere recycle acid of approximately 21 percent P205
is produced. The filter caklle formied ma~y vary from i Lo
4 inches. The type of filter varies widely depending
upon the process requirements. Early plants used
plate-and-frame presses and rotary drum vacuum filters.
At present, four types of horizontal vacuumr filters
are used:
1 Belt filter.
2. Horizontal rotary table filter.
3. In-line pan filter.
4, Horizontal rotary tilting-pan filter.
The horizontal rotary tilting-pan filter has become
thle standard! in ~the new large-capacity pl~ants since
very large units are feasible.
Approximately two-thirds of the weight of phos-
phrate rock and 98 percent of the sulfuric acid used in
the process consist of unsalable by-products.7 The
result of these unusable by-products is the production
of approximately 5 tons of by-product material per ton
of P205 produced. Effluent gases from the process
contain fluoride compounds which must be removed, Nor-
mally, these compounds are removed in somae type of
liquid scrubbers such as spray chambers, imrpingement
columns, packed towers, and venturi scrubbers, Fluorine
recovery is practiced? in somec planr;s where the resulting
fluosilicic acid liquor is recovered for sale. In other
plants, heliquor ispumped tothe gypsum settling
pond where it is recycled back to the process, Slack
reports that the fate of the fluoride in the pond is
not well u~nderstood at present, and that it "presumn-
ably remains with the gypsum in the pond or is evolved
from the pond surface as hydrogen fluoride." His
statement is based upon operating data which show
that the fluoride concentration in the recycled liquor
builds up to a certain level, depending upon operating
conditions, and stops there. Table 5 gives the fluoride
balance for one plant using Florida phosphate rock.
TABLE 5
FLUORIDE BALANCE IN SWIFT & CO.
WET-PROCESS BROSPHORIC ACID PLANT73ea
Material F content % F/ton P205, lb
Vapors from reactor 5.5 13
Gypsum cake 27.8 66
Vapors from vacuum evaporator 41.9 100
Concentrated Acid (54% P205) 24,8 60
Total 100.0 239
a Phosphate rock used contained 32.57% P205 nd3,8%
The quantity of gypsum produced in a wet-process
acid plant is normally 1.5 to 1.6 tons/ton of rock
digested or about 4.6 to 5.2 tons of gypsum per ton of
P205 produced. Approximately 1 acre-foot of gypsum
will be accumulated yearly as a result of the produc-
tion of 1 ton of P205 per day. In Florida, the gypsum
discharged from the filter is re~slurried with water and
pumrped to the settling pond where the solids settle
and the gypsumr~ pond water is recovered for reuse. Thle
settled solids are then used to build up the walls of
the gypFsum pond such that older plants have gypsum
piles exceeding 90 feet above grade.
The effluent water from coolers, scrubbers, and
the gypsum slurry passes through the pond, or ponds
depending upon the plant design, and is recycled back
into the process, N~ormally, little effluent water is
dumped into water courses unless it is absolutely
necessary due to high water conditions caused by heavy
rain and/or plant imlbalances. Before the water is
dumped into fresh water sources, it is treated with
lime to raise the pH.
Fluoride Evolut~ion.
The quantity of fluoride evolved as gaseous sili-
can tetrafluoride during the acidulation of phosphate
rock is usually only a fraction of the total fluorine
in the rock, as shown in Table 5. However, this arnlmoun
of fluoride can do considerable damage to animals and
vegetation if allowed to escape from the confines of
the plant.
The traditional view of fluoride evolution during
the mixinge of p~ohosphat rock and sulfuric acid gave
rise to the following reactions:
CaF2 2iSO4---*CaSOI + 2HF (2
6HF + SLO2---*H2SiF6 + H20 (II
4HiF +Si02-- =*SiF4 +- 2H20 (9
3SiF4 + 4H20---*SiO2 2H20 +; 2H2SiF6 (10
SiF + 2HF---rH2SiF6 (11
The formation of hydrofluoric acid by Equation 2
was generally considered to be an intermediate step since
further reaction within silica would give gaseous silicon
tetrafluoride (SiF ) and aqueous fluosilicic acid
(H2SiF6), Some of the silicon tetrafluoride produced in
Equlation 9 was released in the early stages of the proj-
cess while Equations 10 and 11 were liquid-gas reactions
that caused retention of the- remaining volatile silicon
tetrafluoridie. Th~e gases released in Equation 9 were
normally carried to the scrubbing system where water
or gypsum pond w~ater was used as the scrubbing media to
recover the fluoride as fluosilicate by Equation 10.
Recent studies ,776suggest that the tai
tional view wherein the formation of silicon te~rc-
fluoride is a secondary reaction between hydrofluoric
acid and silica in the rock, Equation 9, may not be
entirely correct, These studies indicated that fluo-
silicic acid is formed directly in acidulation of phos-
phate rock and that the reaction between moderately
strong hydrofluoric acid and either silica or- sili-
cates is too slow to account for the somewhat rapid
evolution of silicon Letrafluoride that is experienced
59
unlder normal operating conditions. The studies suggest
that fluoride behaves more as fluosilicic acid than
hydrofluoric acid in acidulated rock.
In a study of factors affecting the evolution of
silica tctmrafuoride during the acidulation of phas-
phate rock~, Fox and Hill71 found that the amount of gas
evolved on contact between the rock and acid increased
with increasing concentration and temperature of the
acid and with the degree of acidulation of the rock.
Other minor factors affecting fluoride evolution in-
cluded th~e order of mixing t~he rock, acid, and water
and the amount of water taken up in the formation of
the hydrates of the reaction products,
CHAPTER IV
ABSORPTION THEORIES AND
EMPIR1CAL RELATIONSHIPS
Treybal7 has defined gas absorption as "an
operation in which a gas mixture is contacted with a
liquid for the purposes of preferentially dissolving
one or more components of the gas and to provide a
solution of these in the liquid." A detailed discussion
of the absorption process is beyond the scope of this
dissertation which is concerned with the application
of this method to gas purification and so it deals
only briefly with the fundamental processes of absorp-
tion. A full treatment of gas absorption may be found
in specialized texts by Treybal,7 Kohl and Riesenfeld,7
Sherwood and Pigford, Morris and Jackson,80 and
Perry.8 Those aspects which are of particular interests
in the control of fluoride air pollution will be dealt
with in detail.
The theoretical principles involved in the ab-
sorption process are quite complex and controversial.
The newer theories attempt to resolve the apparent
discrepancies involved in using the older theories.
Absorption deals with the transfer of material
between two phases. In the case of gas absorption,
when a soluble gas comes in contact with a liquid,
molecules of the gas will enter the liquid to form a
solution, Gas molecules also will tend to leave the
solution to re-enter the gas phase at a rate which in-
creases as the liquid phase concentration increases,
until the gas pressure exerted by the solution is equal
to the pressure on the system. When this point is
reached, a dynamic equilibrium between the phases is
established and the gas concentration in solution will
no longer change. The cojncenltration of the gas in
solution at this point is termed the solubility of
the gas in the liquid and varies with the temperature,
pressure, and system components, At this equilibrium
point, the partial pressure of the soluble gas in the
main gas is related to the dissolved gas in th-e liquid.
With this idea In mind, it can be stated that a very
soluble gas will require a lower partial pressure than
a slightly soluble gas in order to obtain a given
concentration in the liquid. Figure 12 shows a typical
equilibrium curve illustrating this idea, Equilibrium
curves are normally determined by experimentation;
however, if the solution and gas are considered ideal,
certain gas laws may be applied.
lightly
Solbl
oa,
'nZ; Curquilibrium
Very Soluble
Concentration of gas in liquid
Figure 12: Equilibrium Diagram
Raoult's Law is normally applied to sys~tes that
approach equilibrium and states that
p = Ex
where:
p = partial pressure of solute gas
P = vapor pressure of solute gas at the same
temperature
x = mole fraction of solute gas in solution
When liquid solutions can not be considered
ideal, Hienr:'s Law is normally applied. Henry's Law
related the solubility of gases to the partial pressure
of the gas inl the gaseous phase
S =Hp(
where:
(12a)
:12b)
S = concentration of gas in liquid
H L- proportionality constant, sometimes called
Henry's solubility coefficient
p = partial pressure of gas in gaseous phase
The conlstant may have an~y units required; for example,
if S is in ppm, and p is in atmospheres, H will be
in ppm/atm. The constant is normally determined exper-
imentally for the particular system under study and is
applicable to that system. The law holds for the solu-
bility of nonreactive gases in wa~ter.
Since absorption of gases depends upon the trans-
fer of molecules from the bulk of the gas to the bulk
of the liquid, the case of a stagnant gas in contact
with a liquid surface has been widely studied. When
the liquid surface is brought into contact with the gas,
the diffusion of the gas molecules is by molecular
diffusion, and the rate of this depends on the pressure
and temperature of the gas and the molecular species
in the gas. The rate of molecular transfer is given
by Fick's Law
N =- (12c)
where:
N = rate of molecular transport, molar unit/
unit area/unit time
= concentration gradient in the direction of
diffusion,
D = molecular diffusivity, (length)2/time
Values for the diffusivity may be calculated using the
Stokces-Einstein equation but normally they~ are determined
experimentally and can be found in the Interatina
Critical Tables8 when available, Normally gas diffusi-
vity values are considerably higher, approximately 105
times as large, as liquid diffusivity values.
Equation 12chas been integrated for various cases
and the two most used cases are:
Case 1.
Steady-state equimolal counter diffusion
which frequently pertains to distillation opera-
tions and will not be described further.
Case 2.
Steady-state diffusion of A through non-
diffusing B. This occurs, for example, wheni
silicon tetrafluoride (A) is absorbed from air
(B) into water. In the gas phase if the evapora-
tion of water is neglected, only the silicon
tetrafluoride diffuses, since air does not dis-
solve appreciably in water, Figure 13 illustrates
this condition.
PB\PB
Pressure P1
PA2
Z1 Distance, Z Z2
Figure 13: Diffusion of A Through
Stagnant BI
Troybal8i3 reports the integrated form of Equation
12c for this case to be
In
where :
Na = rate of transport, [moles/(time)(area)]
D = diffusivity, (area/time)
P = absolute pressure
R = gas constant
T = absolute temperature
Pin = log mean pressure of component B
y = mole fraction
yi = mole fraction at interface
k = mass transfer film coefficient,
C(moles)/(time) (area) (mole fractiorn)]
The use of the mass transfer coefficient in Case
2 is generally accepted since most practical situations
involve turbulent flow in which it is generally not
possible to describe the flow conditions mathematically.
The mass transfer coefficient includes, in one quantity~,
effects which are the result of both molecular and
turbulent diffusion and which differ from situation to
situation, Mass transfer coefficients can, in a few
limited situations, be deduced from theoretical princi-
ples, but in the great majority of the cases, direct
measurement under known conditions is used for design
purposes. Many theories have been postulated In order
to attempt to interpre' or ex-plain the behavior of the
mass transfer coefficients. The most widely known are
the two-film, penetration, and surface-renewal.
TwofmThory
The oldest and most widely used model for inter-
pretation of mass transfer coefficients is the two-
film theory proposed by Lewis and Whbitman. This
model relies on a series of assumptions for its utili-
zation:
1 Steady-state conditions exist in both phases,
2. Rate of transfer is proportional to the
concentration gradient.
3. Equilibrium exists between liquid and vapor
at the interface with no interfacial resis-
tance to flow.
4. Hold-up at the interface is zero.
With these assumptions, when steady-state conditions
of transfer have been reached, the rate of transfer Na
from gas stream to the interface, and from the interface
to the liquid must be equal, and
Na = k (p p ) = k (Ci C) (14)
where:
p = partial pressure of the transferring com-
ponent in the gas stream
pi = partial pressure at the interface
C. = concentration at the interface
C = concentration in the liquid
kg = gas film mass transfer coefficient
kl = liquid film mass transfer coefficient
However, since the interface values pi and Ci are very
difficult to determine, the equation was redefined in
two equations:
Na = K (p p") = K (Cn C) (15)
where:
p* = equilibrium partial pressure of gas over a
solution having the same concentration
C as the liquid
C" = concentration of a solution which would be
in equilibrium with the gas partial
pressure existing in the main gas stream
K_ = over-all mass transfer coefficient for gas
phase
K1 = over-all mass transfer coefficient for
liquid phase
'and K_1
g m 1 m
kg 1l
where:
m = slope of solubility equilibrium line.
This model has been misapplied in highly soluble
gas systems where the slope m approaches zero. Equation
16 would then become Kg = ke (17)
indicating that the gas phase controls, and the liquid
phase offers no resistance to mass transfer. Teller
has pointed out that in absorption with chemical reaction
occurring a liquid is selected so that no equilibrium
partial pressure exists for the gas when dissolved and m
approaches zero, However, experimentation has shown
that Equation 17 does not hold and that the liquid
phase resistance can not be assumed to be negligible
when chemical reaction occurs.
Peneration TheorY
In 1935, Higbie8 emphasized that in many situa-
tions the time of exposure of a fluid to mass transfer
is short, so that the concentration gradient of the
film theory, characteristic of steady-state, would not
have time to develop. As an approximation to actual
conditions, he applied Fick's second law equation for
a one-dimensional case and with appropriate boundary
conditions found that even with a constant concentration
gradient, the rate of absorption decreases with time of
exposure. However, he did not apply his results to
absorption with chemical reaction which is the case in
m~any applications at present.
Surface Renewal Theory
In 1951 Danckwerts pointed out that the
Highte theory was a special case of a situation where
the liquid edd?- motion continuously brings fresh liquid
to the surface, increasing the rate of absorption. This
motions would tend to increase maEss transfer to a greater
degree than that predicted by Higbi~e's penetration
theory. While it expanded H~igbie's theory, it still
did not apply to absorption with chemical reaction and
it could not be applied directly to highly turbulent
flow conditions.
Film Penetration T~heory
Toor and M~arcello proposed their film penle-
tration model in 1958. They showed that the film and
penetration models were not separate concepts, but
only the limiting cases of their model. In doing this,
their theory seemed to better explainl some physical.
conditions than did previous theories.
Absorption with Chemical Reaction
When mass transfer occurs between phases across
an interface, the resistance to mass transfer in each
phase creates a concentration gradient in each phase.
In the case of purely physical absorption, the rate of
mass transfer is dependent upon concentration gradients
and the diffusivities of the gas in both phases as was
stated in the "film theory" and as is shown in Figure 14.
Ci
Liquid
Characteristics in
With the addition of a rapid chemical reaction to the
process, the distance through which the gas m~ust diffuse
in the liquid phase is generally decreased because
upon meeting the liquid,reaction occurs. Therefore,
the need for further diffusion no longer exists, in-
creasinge the over-all mass transfer rate. Figure 15
Lllustr~ates this concept.
Figue 14 Intrfaial
Physcal bsortio
8
m
lcl
L1
a,
LJ
C
CABL
Figure 15: Interfacial Characteristics in
Absorption with Chemical Reaction
When a slow chemical reaction occurs, the distance
through which the gas must traverse is only slightly
affected and the process approaches that of physical
absorption,
Teller indicates that as a result of the many
complexities involved in the mechanism of absorption
accompanied by chemical reaction, no rigorous analyti-
cal method of design has been developed; but for a
first order reaction, the rate of absorption is pro-
portional to the unreacted gas concentration at the
interface. For a second order reaction, an increase
in the concentration of unreacted gas at the interface
Reaction
Zone
tends to remove unavailable unreacted liquid from the
interfacial region and moves the reaction zone deeper
into the liquid. With this in mind, an economical pro-
cess dictates the selection of the proper liquid and
the occurrence of a first order reaction or of a pseudo-
first order reaction, an excess of liquid or low gas
concentrations in the gas stream.
Fluoride Absorction
Generally, the two-fi~lm theory is used for the
design of flu~oride absorption units and the interpre-
tation of their performance.
General D~esign Concepts
Simplified equations have been developed for
design calculations where low gas and liquid concentra-
tions are involved and where the equilibrium curve is
linear or can be assumed linear over the range of con-
centrations used in the design. Equation 18 is the
most widely used equation for design calculations and it
is based upon the two-film theory
h =H N (18)
og og
where:
h = height of tower
Ho height of transfer unit
N = numbSer of transfer units
og
This equation is arrived at using Equation 15 as the
starting point and a differential section of an absorp-
:ion column dv, in which the rate of absorption is
Nadv = K (P p") dv = K (C- C) dv (19)
If a is the interfacial area per unit volume of absorber,
and S is the tower cross-section, then dv = aSdh (20)
However, the value of a is usually not known and diffi-
cult to determine, therefore, it is usual to combine
the mass transfer coefficient with the interfacial area
to form a composite coefficient Kga and K a, the Equa-
tion 19 becomes:
NaaSdh = K a(p, p*) Sdh = K a(CA C) Sdh (21)
For gas absorption, where the molar rates of flow
of gas and liquid phases are not constant, due to the
changes in composition which occur over the length of the
absorber, it has been shown that
d(Vy) = V--dy (22)
where:
V = molar rate of flow of gas phase
y = mole fraction solute in gas stream.
Combining Equations 21 and 22 and utilizing simplifing
assumptions yields the equation for the height of the
absorption tower with gas film coefficients
b = d (23)
where:
G = molar mass velocity of gas, lb moles/(hr)
(ft ) ata
P = total pressure of system, atm
Y = mole fraction of solute in gas stream
Y, = mole fraction of solute in gas in equili-
brium with liquid
Y1 = solute concentration of gas entering the
column, lb mole solute/1b moles inert gas
Y2 = salute concentration of gas leaving the
column, Ib moles solute/1b mole inert gas
This equation assumes that:
1. The couilibrium curve is linear over the
range of concentrations encountered, there-
fore, over-all ccefficients may be used.
2. The partial pressure of the inert gas is
essentially constant over the length of the
column ,
3. The solute contents of gaseous and liquid
phases are sufficiently low that the partial
pressure and liquid concentration values
may be assumed proportional to the correspond-
ing values when expressed in terms of moles
of solute per mole of inert gas,
Chilton and Colburn8 introduced the concept that
the calculation of column height invariably requires the
integration of a relationship such as (fromn Equation 23)
yr2 'YeY
The dimensionless value obtained f-rom the integration
was called the number of transfer units based on an over-
all gas dr-iving force, Nog, and is a measure of the
difficulty of the gas absorption operation. By substitu-
tion~ Equation 23 becomes
h $ E --- (25)
and the height of the transfer unit, Hog, can then
be defined as
Hog h K af (26)
og g
Values of N~og are" particularly uIseful for expressing
the performance of absorption equipment in which the
volume is not of fundamental importance.
In fluoride absorption uni-cs, where a large
excess of water is used, the concentration of the acid
formed in the solution is low and Ye may be neglected
and Equation 24 is simplified to
N, = S1 = In (27)
y2
Since the absorption efficiency, E, is directly related
To y1 and y2 by
P1 7
E = ( ) x 100 (28)
'1
then Ecuation 27 can be rewritten asi
Nog =n 1- E1/100 \ (29)
The necessity for fluoride removal from effluent
gases has been stated in previous chapters. Fortunately,
both hydrogen fluoride and silicon tetrafluoride are very
soluble in water and most commercial installations for
control of fluoride emissions make use of this fact.
The vapor pressure of hydrogen fluoride over aqueous
solutions is shown in Figure 16, which is based on the
data of Brosheer,90 Whynes91 has presented the data given
in Figure i7 on the concentration of fluosilicic acid in
equilibrium with silicon tetrafluoride vapors.
When silicon tetr~afluoride is absorbed by water,
it reacts to form fluosilicic acid, The mechanism of
the absorption process has been studied by Whynes,9
who suggested that the reaction probably occurs in steps,
as represented by Equations 30 and 31.
SiF4 + 2H20:a-d Sto2 + 4HIF (30)
2HF +SiF4mr=EH2SiF 6 (31)
The simple fluosilicic acid probably reacts with addition-
al SiF4 or S102 to form a more complex form of this
compound, The reaction given by: Equation 31 is reversi-
ble to the point that solutions of fluosilicic acid
exert a definite vapor pressure of HF and SiF ,
As Figures 16 and 17 indicate, both hydrogen
fluoride and silicon tetrafluoride are very soluble in
1.5
1 .0 0
L~0.5
0 L -
0 5 ~10 15
Solution Concentration, WrT Percent HF
Figure 16: Vapor Pressure of H drofluoric Acid
Over Dilute Aqueous Solutionsv
O C
20 30 40
Solution Concentration, WT Percent H2SiF6
Figure 17: Vapor Pressure of Silicon Tetrafluyoride
Over Aqueous Sol-tions of Fluosilicic Acid.9
water and the gas film resistance would be expected
to be the controlling factor in their absorption. This
has generally been verified by experimental evidence
although Whynes9 found silicon tetrafluoride absorption
to be complicated by a tendency to form miist droplets
in the gas and the tendency of the silica, produced in
Equation 30, to form a solid film on the outside of the
water droplets, thus hindering absorption.
Whynes91 reported that even though the absorption
of silicon tetraflooride in water is complex and involves
a chemical reaction, the precipitation of silica, the
results could be explained in terms of the two-film
theory.
Types of absorption equipment which have been
used or proposed for the removal of fluoride vapors
from gas streams with water include the following general
types:
1 Spray chambers (vertical and horizontal)
2, Packed towers (10w-pressure grid packing)
3, Venturi scrubbers
4. Wet cells (beds of wetted fiber)
5. Plate columns
6, Ejectors
Since the scope of this research was limited to three
types of absorption equipmnat; namely, spray ch-ambers,
venturi scrubbers. and plate columns, the discussion
from this point on will be limited to the types studied.
Absoration in thec Cyclonic Snray Chamber
A spray chamber is thte simpliest type of absorp-
tion equipment. In general, it consists of a cylindri-
cal tank in which spray nozzles are mounted in the upper
section and the droplets generated fall domwnwrds with
some liquid flowing downward on the interior walls.
Although the liquid and gas streams may be concurrent,
the most common type is a count~ercurrent flow where
the gas stream is introduced at the lower section of
the tower tangentially so that a spiral motion is im-
parted to the gas and the liquid sprays are directed
radially. True, countercurrent flow is difficult to
maintain due to flow characteristics and spray entrain-
ment, which tend to limit the mass transfer available.
Some spray towers have been successful in the absorp-
tion of slightly soluble gases, but since the mixing
of spray and gas is not as vigorous as in other equip-
ment, the application of spray towers is best suited
to operations involving easily absorbed gases.
In spray chambers, liquid dispersion is created
by liquid phase mechanical power consumption. Due to
scrubber design, gas phase power consumption is usually
quite low, and Lunde9 found that removal efficiency
depends largely upon the power expended in the liquid
phase. Hie stated that the number of transfer units
would be related to power introduced in the liquid anld
gas phase by
N : - -(32)
og p 0.1
where:
Pi = power introduced in liquid
Pg = power introduced in the gas
Per, Sherwood and Pigford,7anTryl
have summarized various studies mainly pertaining to sul-
fur dioxide removal and have presented curves for the
number of transfer units versus liquid and gas mass
flow rates.
In any attempt to compare the performance of
different types of absorption equipment, fixed operating
conditions for each installation present a problem, and
an uncertainty exists as to the controlling factors in
the performance of the equipment, unless these factors
have been explored over a wide range on a pilot plant
scale. Unfortunately, laboratory or pilot plant data
on the absorption of fluorides are practically non-existant.
Sherwrin9 reported on 14 full-size widely differing
installations ranging from continuous spray towers to
continuous ejection systems. Efficiencies up to 99.3
percent were obtained and values for Kga were reported.
However,? the volume coefficient K~a has little value
as a means of correlating spray equipment because the
area for mass transfer a varies with liquid rate, noz-
zle design, liquid pressure, distance from nozzle, and
other factors. Performance of different fluoride
vapor absorbers may be compared on the basis of the
number of transfer units in a given piece of equipment,
or, more simply, on the basis of percentage removal
efficiency.
Absorption in the Venturi
In the venturi scrubber, the velocity of the gases
causes the atomization of the scrubbing liquid. A typi-
cal venturi scrubber has the gas stream pass at a high
velocity thr-ough the venturi with the scrubbing liquid
being introduced at relatively low pressure at the venturi
throat where shearing of the liquid and droplet formation
take place, Large areas of liquid are made available
to contact the gas as a result of a high degree of dis-
persion which takes place at the throat, due to an in-
crease in the throat velocity. The energy required for
forming droplets and intimate mixing with the gas is
furnished by the gas stream. The resulting spray of
liquid mixes with the gas and absorption takes place.
It is necessary that an entrainment separator be pro-
vided at the discharge of the venturi to remove the en-
trained liquid from the gas. The entrainment separator
is usually of cyclonic design and normally the venturi -
cyclonic spray chamber is installed as a single unit.
Venturi scrubbers were designed originally for
dust and mist removal, and considerable research949
has been done in this area. Lunde has reported on
venturi installations for sulfur dioxide removal and has
suggested that their performance on fluoride vapors
would be controlled by the same factors as SO2. The
literature contains isolated cases reporting on the per-
formance of various venturi scrubber installations in
removing gaseous fluoride from process gases.91,29
Lunde9 reported on venturi units using water as the
scrubbing liquid with removal efficiencies generally
in the 95 to 98 percent range when employed for hydrogen
fluoride absorption,
In general, the average pressure drop across a
venturi varies from 10 to 50 inches of water depending
upon scrubber design. The liquid requirement has been:
found to be between 3 to 15 gpm per 1000 ofm with power
requirements in the range from 1 to 10 horsepower per
1000 cfm, Lundle2 also reports on attempts to correlate
the number of transfer units with power requirements and
he proposed that
N : : (P 0.3) 7 0.8) (
og 1 g
This indicates the dependence of the efficiency of the
venturi on the energy introduced to the gas phase.
Absorption in the Baffle Plate Imnineement Column
In plate towers, the gas-liquid contacting takes
place on a series of plates installed in a cylindrical
column. The plates are perforated by small holes through
which the gas passes while the liquid flows across the
plate to a dow~ncomner to the next plate in the column.
Thle gas stream palsses up thirough the perforated holes
in the formr of irregular bubbles and through the liquid
layor on top of the plate and mixes with the liquid.
Rapid absorp~tion occurs by virtue of the large interphase
surface produced by the small bubbles generated.
The literature abounds with studies on this type
of scrubber and the American Institute of Chemical
Engineers has published a design manual relating many
of the variables of these studies. Studies have related
mass transfer efficiency to plate characteristics such
as weir height, to gas flow rates and velocities, and
to liquid flow rates. However, as Perry indicates,
studies must be performed under similar conditions and
this has generally not been the case. Since this type
of scrubber is not generally employed in fluoride re-
moval processes, no information was available in the
litera'-ure regarding fluoride scrubbing. However, if
a parallel may be drawnm with sulfur dioxide scrubbing,
efficiencies above 90 percent have been reported with
high liquid to gas ratios (50 to 100 gpm per 1000 cfm).
Correlation of Scrubber Efficiencies
Absorption data are usually presented in terms
of the effect of liquid or gas mass flow rates on the
mass transfer coefficient, or on, the height of transfer
unit. The economic evaluation of equipment requires
translating these data into over-all performance as
measured by eificiency or number of transfer units.
Perry anid Lunde2 have indicated that it is difficult
to compare two basically different types of equipment
inl terms of mass transfer coefficients, mass flow rates,
or eight of transfer unit.
Semrau, Lunde,9 and others have correlated
the efficiency of a scrubber expressed in number of
transfer units, with total power consumption, the pres-
sure loss of the gas passing through the scrubber and
the pressure loss required for atomization of the scrub-
bing liquid. To date this seems to be the moost accepted
method of making comparisons between different ty-pes of
equipment, Semrau nas plotted the number of transfer
units against total power consumption on a log-10g plot
and expressed the results in the form
Nt =GPB (34)
where a and F are characteristic parameters of the
material being removed, He applied this equation to
dusts and mists and determined that the efficiency of
collection could be expressed in terms of total power
used and the characteristics for the dust being collected,
and independently of the actual type of scrubber being used.
For this study, since the prediction of exact mass
transfer characteristics would be very difficult due to
the unavailability of data, the procedure described by
|
