sponsored by florida institute of phosphate research...
TRANSCRIPT
The Florida Institute of Phosphate Research was created in 1978 by the Florida Legislature(Chapter 378.101, Florida Statutes) and empowered to conduct research supportive to theresponsible development of the state’s phosphate resources. The institute has targeted areas ofresearch responsibility. These are: reclamation alternatives in mining and processing, includingwetlands reclamation, phosphogypsum storage areas and phosphatic clay containment areas;methods for more efficient, economical and environmentally balanced phosphate recovery andprocessing; disposal and utilization of phosphatic clay; and environmental effects involving thehealth and welfare of the people, including those effects related to radiation and waterconsumption.
FIPR is located in Polk County, in the heart of the central Florida phosphate district. The Instituteseeks to serve as an information center on phosphate-related topics and welcomes informationrequests made in person, by mail, or telephone.
Research Staff
Executive Director
David P. Borris
Research Directors
G. Michael Lloyd, Jr.Gordon D. NifongDavid J. RobertsonHassan El-ShallRobert S. Akins
- Chemical Processing- Environmental Services- Reclamation- Beneficiation- Mining
Florida Institute of Phosphate Research1855 West Main StreetBartow, Florida 33830(863)534-7160
PROCEEDINGS OF THE SECOND INTERNATIONAL SYMPOSIUM
ON PHOSPHOGYPSUM
Sponsored by
FLORIDA INSTITUTE OF PHOSPHATE RESEARCH
Organized byUniversity of Miami
and Held atUniversity of Miami/James L.Knight
International Conference CenterMiami, FloridaDecember, 1986
Contract Manager: G. Michael Lloyd, Jr.
DISCLAIMER
The contents of this report are reproduced herein as receivedfrom the contractor.
The opinions, findings and conclusions expressed herein are notnecessarily those of the Florida Institute of Phoshate Research,nor does mention of company names or products constituteendorsement by the Florida Institute of Phosphate Research.
i
ii
iii
iv
V
CHEMICAL RECOVERY
UTILIZATION OF PHOSPHOGYPSUM FOR THE PRODUCTION OF SULFURIC ACID
T. D. Wheelock (Professor), C.-W. Fan (Research Associate),
and K. R. Floy (Research Assistant)
Chemical Engineering Department and
Engineering Research Institute
Iowa State University
Ames, Iowa 50011
ABSTRACT
A method for producing sulfur dioxide and quicklime from phospho-
gypsum and another calcium-containing waste was demonstrated with bench-
scale and pilot plant-scale fluidized bed reactors. The sulfur dioxide
produced by this method is suitable for conversion into sulfuric acid by
conventional methods and the by-product quicklime shows promise for a
number of applications. Similar results were achieved with two different
versions of the process. In one version a solid reductant (coal) was
employed, whereas in another version gaseous reductants derived from coal
or natural gas were utilized. A bench-scale reactor was used to
demonstrate basic concepts and to study important operating conditions,
and a pilot plant reactor was used to confirm results achieved with the
smaller unit. The pilot plant was operated with feed rates up to 70
lb./hr. and for periods ranging up to 30 hr.; few problems were
encountered. The results provided a basis for a preliminary design study
for a large-scale reaction and heat recovery system which showed that a
product gas stream containing 11-12% sulfur dioxide can be produced in a
practical system.
3
INTRODUCTION
The development of a practical and economical method for producing
sulfur dioxide and quicklime from phosphogypsum or other calcium sulfate
containing wastes and minerals has been the principal objective of a major
research effort conducted over a number of years at Iowa State University.
Such a method is intended to produce sulfur dioxide for conversion into
sulfuric acid and quicklime for use in a number of applications. It has
been shown that calcium sulfate contained in various materials can be
decomposed at high temperatures in the presence of either gaseous or solid
reductants to produce the desired products. Two variations of this basic
method have been demonstrated by employing bench-scale and larger
fluidized bed reactors and using either natural gas or coal to supply the
reductants and energy requirements (Swift and Wheelock, 1975; Wheelock,
1978; Smith et al., 1984; Morris et al., 1985; Wheelock and Morris, 1986).
In one version, coal (a solid reductant) is incorporated in the feed
particles whereas in the other version gaseous reductants (CO and H2) are
generated by partial combustion of either powdered coal or natural gas
within the fluidized bed. With either approach additional fuel is burned
within the fluidized bed to satisfy the heat of reaction.
In the first approach the additional fuel is mixed with close to a
stoichiometric amount of air to create a neutral atmosphere, whereas in
the second approach the fuel and air are supplied in such a way that a
reducing atmosphere is created in the lower part of the fluidized bed of
reacting particles and an oxidizing zone in the upper part. Hence, the
first approach makes use of a single reaction zone while the second
approach utilizes two reaction zones. Both methods have been demonstrated
4
in a bench-scale reaction system, but only the second method has been
demonstrated in a small pilot plant.
A summary is presented below of the results of applying the calcium
sulfate decomposition process to phosphogypsum and to waste gypsum
produced by neutralizing spent acid from the Radford Army Ammunition
Plant. In addition, consideration is given to the properties and
potential uses of the by-product lime and to the results of a preliminary
design study for a practical reaction and heat recovery system for the
process.
EXPERIMENTAL METHOD
REACTION SYSTEMS
The results described in this paper were obtained with two different
fluidized bed reaction systems. Although the reactors were of similar
design, one reactor had an inside diameter of 6.0 in. and overall height
of 7 ft. while the other reactor had an inside diameter of 10.0 in. and
overall height of 10 ft. Both reactors were constructed of refractory
materials. The tubular body of each reactor was made of silicon carbide,
and each reactor was fitted with a ceramic gas distributor and refractory-
lined cap. Four cone-shaped openings were cast into each gas distributor
for admitting most of the fluidized gas. Each gas distributor also had
openings for an overflow tube and a thermowell. A ceramic overflow tube
was used for removing solids and for maintaining a constant fluidized bed
depth. The cap on each reactor had a gas outlet, and the cap supported an
axial , ceramic dip tube which introduced feed particles and secondary air
near the middle of the fluidized bed. The main body of each reactor was
5
Figure 1. Pilot plant used for demonstrating CaSO4 decomposition process.
surrounded by an electric furnace for preheating the reactor and for
making minor adjustments in the operating temperature.
Figure 1 is a schematic diagram of the larger reaction system or
pilot plant. The reactor was supplied with granular feed by an L-valve
feeder and secondary air was used to transport the solids. The reactor
was also supplied with primary air and natural gas which were admitted
through the openings in the distributor plate. Reacted solids which
passed through the overflow tube were collected in a canister which was
emptied periodically. The reactor off-gas passed through two cyclone
separators in series to remove most of the entrained dust. The gas was
subsequently vented through an ejector which kept the system under
slightly reduced pressure.
6
The smaller, bench-scale reaction system was similar in design except
that it also included a coal feeder and utilized a different method for
removing dust from the off-gas. Thus, instead of cyclone separators the
smaller system employed a large settling chamber and a fiberglas filter in
series .
Equipment was also provided for analyzing the off-gas from either
reactor. A gas chromatograph was used to measure sulfur dioxide
concentration and a standard Orsat apparatus to measure relative
concentrations of acidic gases, oxygen, and carbon monoxide.
MATERIALS AND FEED PREPARATION
Phosphogypsum was supplied by Zellars-Williams, Inc., of Lakeland,
Florida, which collected the material from various sources and washed it.
Waste gypsum from the Radford Army Ammunition Plant was supplied by
Hercules Inc. Both materials were shipped to various manufacturers of
drying equipment for drying. The dried materials retained some water of
crystallization so that their moisture content ranged from 5 to 20%. The
impurity content of the dried phosphogypsum ranged from 5 to 8% and
consisted mainly of silica and unreacted phosphate rock. The impurity
content of the dried waste ranged from 1 to 3% and seemed to consist
mainly of calcium hydroxide.
High volatile bituminous coals were employed in some runs. One coal
was from the Illinois No. 6 seam near Middle Grove, Illinois, and another
from the Pittsburgh No. 8 seam near Fairview, West Virginia. The dried
coals had an ash content of about 13% and total sulfur content of about
3%.
7
The reactor feed was prepared by briquetting or compacting the dried
gypsum with a double roll press, crushing, and screening. Screening
separated the particles into three size fractions. The undersize
particles were returned to the compaction rolls, the oversize particles
were recrushed, and the intermediate size particles were used as feed.
When coal was incorporated in the feed, the ground coal and gypsum were
mixed before compaction.
REACTOR OPERATION
A similar procedure was used for making a run with either reactor.
The reactor was first preheated overnight with the electric furnace.
After starting the fluidizing air, the reactor was filled with reacted
solids from a previous run. Natural gas was introduced, and when the
desired operating temperature was achieved , unreacted solids were supplied
at a predetermined rate. For some runs the fuel was switched from natural
gas to powdered coal after the operation appeared stabilized. Operating
conditions were then kept constant for several hours to allow the system
to approach steady-state. During the operation, the canisters which
received solids from the reactor overflow tube and from the dust
collection system were emptied periodically. The recovered solids were
weighed and later analyzed. The reactor off-gas was sampled
intermittently and analyzed with the Orsat apparatus. Occasionally the
sulfur dioxide content was determined by gas chromatography.
EVALUATION OF RESULTS
To evaluate the results of operating the fluidized bed reactors, the
following quantities were determined for the steady-state portion of each
8
run : the sulfur dioxide content of the off-gas and the desulfurization,
sulfide content, and recovery of the reacted solids. A sulfur balance was
used to estimate the sulfur dioxide content of the off-gas, and the
validity of this method was verified on several occasions by direct
measurement with a gas chromatograph. After each run the reacted solids
were chemically analyzed. for sulfate, sulfide, and calcium by well known
procedures, and the results were used to calculate the desulfurization by
means of the following equation:
A calcium material balance was used to determine the distribution of
calcium between the reactor overflow product, fines entrained in the
off-gas and subsequently recovered, and various losses. The principal
loss seemed due to entrained dust in the off-gas which was not recovered.
EXPERIMENTAL RESULTS
BENCH-SCALE REACTOR RUNS
Approximately 100 runs have been made with the 6-in. diameter bench-
scale reactor to demonstrate the two versions of the calcium sulfate
decomposition process, to investigate the use of different fuels and
process conditions, and to apply the process to both phosphogypsum from
Florida and waste gypsum from the Radford Army Ammunition Plant.
To demonstrate the first version of the process, several runs were
made in which Illinois No. 5 coal was incorporated in the granulated feed
prepared from phosphogypsum; the amount of coal corresponding to
9
0.075-0.077 lb./lb. CaSO4. Additional powdered coal or natural gas was
mixed with the fluidizing air and burned in the fluidized bed to supply
the heat of reaction. The best results were achieved when the air to fuel
ratio was adjusted to provide a neutral off-gas, that is, an off-gas
containing neither excess oxygen nor excess carbon monoxide. The results
of two such runs are listed in Table 1 under run A and run B. These runs
were made using different feed rates. The indicated fuel rate corresponds
to the heating value of both the coal incorporated in the feed and the
additional coal which was burned in the fluidized bed. For both of these
runs the solids discharged through the reactor overflow tube were 97-98%
desulfurized, and the fines recovered from the off-gas were 82-83%
desulfurized. The sulfide content of the overflow product was 0.2% or
less and that of the recovered fines 0.6-1.6%. In run A the overflow
product accounted for 70% of the calcium fed to the reactor while in run B
it accounted for 63%. The remainder of the calcium was largely accounted
for in the recovered fines. In both runs the estimated sulfur dioxide
content of the off-gas was over 6%. The main difference in the results of
the two runs was the difference in distribution of the solid product
between overflow solids and entrained fines. At the higher throughput of
run B, a significantly higher percentage of material was entrained in the
off-gas with a corresponding reduction in percentage of material recovered
through the overflow tube. Unfortunately, the entrained fines were not
desulfurized as well as the overflow product. Particles smaller than 40-
mesh size seemed to be elutriated from the fluidized beds before they were
completely reacted.
10
Table 1. Results of selected runs mdewithbench-scale reactorusingeither phospho-
gypsum(PG)orwastegpsun. Average steady-state conditions are listed.
E?srameter RunA RunB RunC RunD RunE FunF
No. reactionzones 1 1 2 2 2 2
lypeoffeed EG FG FG FG Waste Waste
Feed size, mesh -12/+1cO -l2/+100 -12/+50 -12/+x) -12/&O -12/+40
Feed rate (dry), lb./(hr.)(ft.)2 90.9
Qpeoffuel CO&l
Fkl rate, Btu/lb. casO4 4300
Mole prim. air/mole Cl-I4
Total mole air/mole Casoq 11.1
Bed teqxrature, OF 2100
Estimated SQ2 cone, % 6.3
Overflow solids
iksulfurization, % 97.8
Sulfide as S, % 0.2
Recoveredfines
Desulfurization, % 83.4
Sulfide as S, % 1.6
cdlcium recoveq
In overflow, % 69.7
In fines, % 21.7
133.5
Cd
4300
11.7
2100
6.2
lM.4 104.4
coal coal
4210 4110
- -
12.6 12.6
2100 2100
5.9 5.7
80.4
gas
4250
5.0
12.6
2100
5.7
Ew
4170
6.0
12.4
2100
6.1
%.9 %.l 94.8 98.5 %.4
(0.1 (0.1 (0.1 (0.1 X0.1
82.1 84.7 83.7 89.0 88.9
0.6 0.6 1.1 0.3 0.5
62.5
36.0
68.8 82.4
26.5 10.0
82.0 78.0
18.0 20.7
120.4
11
To demonstrate the second version of the process, a number of runs
was made in which either phosphogypsum or waste gypsum was treated in the
bench-scale fluidized bed reactor operated with two reaction zones. A
limited amount of primary air was introduced with the fuel through the gas
distributor resulting in a reducing zone in the lower part of the
fluidized bed while excess secondary air was introduced near the middle of
the bed resulting in an oxidizing zone in the upper part of the bed.
Consequently as the solids circulated within the bed, they were
alternately exposed to both oxidizing and reducing conditions. The
results of several runs based on this approach are listed in Table 1. In
runs C and D phosphogypsum was treated using powdered Illinois No. 5 coal
as a fuel, and in runs E and F waste gypsum was treated using natural gas.
With minor exceptions, the results of these runs were rather similar and
not much different from the results of runs A and B. The runs made with
waste gypsum and natural gas resulted in somewhat greater desulfurization
of the entrained fines, and the sulfide content of the fines was lower
than for the other runs. Also it was apparent that runs made with a
coarser feed and/or lower feed rate resulted in a higher recovery of
overflow product.
To further investigate the second version of the process and the
effects of various operating conditions, additional runs were conducted
with the bench-scale reactor using waste gypsum and natural gas. From the
results presented in Figure 2 it can be seen that the specific feed rate
based on reactor cross sectional area had relatively little effect on
desulfurization of the solids even though the rate affected the average
particle residence time markedly. The principal effect of increasing the
12
Figure 2. Effect of specific feed rate on desulfurization of waste
gypsum.
feed rate was to increase the rate of particle entrainment in the off-gas
and reduce the rate of solids removal through the reactor overflow tube.
Thus as the feed rate was increased from 60 to 120 lb./(hr.)(ft.)2, the
calcium recovered through the overflow tube declined from 90% of the
calcium supplied in the feed to 70%.
The effect of temperature on particle desulfurization is illustrated
by Figure 3. As these results indicate when the temperature of the
fluidized bed was reduced from 2100 to 1920°F, desulfurization of the
overflow product declined from 98 to 87% and desulfurization of the
entrained fines from 89 to 48%. Thus the entrained fines were affected
13
Figure 3. Effect of fluidized bed temperature on desulfurization of
waste gypsum.
more by a change in fluidized bed temperature than the overflow product.
The effect of the primary air to fuel ratio is indicated by Figure 4.
Over the indicated range of this ratio (4 to 7) the combustion of natural
gas would generate large amounts of carbon monoxide and hydrogen since a
ratio of 10 is required for stoichiometric combustion of methane. As the
ratio was reduced, the concentration of carbon monoxide and hydrogen in
the lower part of the fluidized bed would have increased. The results
indicate that a ratio of 4.5 resulted in maximum desulfurization of the
14
Figure 4. Effect of primary air to fuel ratio on desulfurization of
waste gypsum.
overflow product and a ratio of 5.7 resulted in maximum desulfurization of
the entrained fines.
PILOT PLANT RUNS
Twenty-five runs have been made with the 10-in. diameter fluidized
bed reactor to demonstrate the second version of the calcium sulfate
decomposition process on a larger scale with phosphogypsum and with waste
gypsum using natural gas for both heating and as a source of reductants.
The results were similar to those achieved with the bench-scale reactor
when the two reactors were operated under comparable conditions including
the same fluidized bed depth and same specific feed rate per unit area of
reactor cross section. The results of several runs are listed in Table 2.
While most of the individual runs made with either the bench-scale
reactor or pilot plant were 5-6 hr. in duration, several series of runs
were made in which the bench-scale reactor was operated continuously for
up to 20 hr. and several extended runs were made with the pilot plant.
15
Table 2. Results of selected pilot plant runs made with either phospho-
gypsum (PG) or waste gypsum and natural gas. Average steady
state conditions are listed for two-zone operation.
Parameter Run I Run II Run III
Type of feed PG Waste
Feed size, mesh
Feed rate (dry), lb./(hr.)(ft.j2
Fuel rate, Btu/lb. CaS04
Mole prim. air/mole CR4
Total m’ole air/mole CaS04
Bed, temperature, OF
Estimated SO2 cont., %
Overflow solids
Desulfurization, %
Sulfide as S, %
Recovered fines
Desulfurization, %
Sulfide as S, %
Calcium recovery
In overflow, %
In fines, %
-12/+100 -8/+20
61.6 86.5
4650 4530
5.9 5.0
13.5 13.3
2100 2010
5.5 5.8
97.6 98.0 95.2
(0.1 (0.1 0.1
86.5 71.6 66.8
0.6 0.8 1.4
80.0 88.1
20.0 11.5
Waste
-8/+30
85.3
4170
5.0
12.2
2010
5.7
84.4
14.3
16
Run III listed in Table 2 was an extended pilot plant run which lasted 22
hr. The only significant problem encountered during this run was a build
up of scale in the off-gas piping which caused excessive pressure drop.
After taking remedial action, a subsequent run was continued for 30 hr.
Although the build up of scale was reduced considerably, the problem was
not completely eliminated. Various methods of dealing with the problem
are under consideration.
QUICKLIME CHARACTERISTICS
The by-product quicklime produced from calcium sulfate has a number
of potential uses including the treatment of acid soils, stabilization of
soils for roads and earthwork, neutralization of acid wastes, control of
SOx emissions from heat and power generation, and conversion into Portland
cement . Samples of the quicklime were subjected to two different tests to
determine the suitability of the material for some of these applications.
In one test, quicklime produced from phosphogypsum was found to be an
excellent soil stabilizing material. A standard method (Chu and Davidson,
1960) was used for conducting the test. Loess soil from Western Iowa was
mixed with 2% quicklime and an optimum amount of water and then molded
into small cylinders. After curing for seven days in a humid room, the
unconfined cylinders were loaded to failure in a testing machine. The
results achieved with the by-product lime are compared below with those
obtained with a standard commercial lime and with no lime. The by-product
lime derived from phosphogypsum increased the unconfined compressive
strength of the soil by 245% whereas the commercial lime increased the
compressive strength by only 10%.
17
MaterialCompressive
s t rength , l b . / in . 2
Loess soil alone 40
Loess + commercial lime 44
Loess + by-product lime 98
In another test, quicklime made from waste gypsum was subjected to a
standard slaking rate test (American Society for Testing and Materials,
1986) which shows how rapidly quicklime is hydrated under controlled
conditions. A sample of quicklime is mixed with water in a modified Dewar
flask and the resulting rate of temperature rise of the mixture is an
indication of lime reactivity. Furthermore the total temperature rise is
an indication of the calcium oxide content of the lime. The results of
such a test are shown in Figure 5 for by-product lime produced from waste
gypsum in run II (see Table 2) and for a commercial lime made from
limestone. Although the initial rate of reaction of the commercial lime
was somewhat faster than that of the by-product lime, the total slaking
time was the same for both materials. Also the calcium oxide content of
the by-product lime was greater than that of the commercial lime.
PRELIMINARY-DESIGN STUDY
The preceding experimental results were used as a basis for a
preliminary design study involving the two-zone fluidized bed reactor
concept coupled with an appropriate means of recovering heat from the
off-gas for use in preheating the granular feed and air supplied to the
reactor. One of the more practical methods of recovering heat from the
off-gas is illustrated by Figure 6. In this method some of the sensible
heat of the off-gas is transferred to the incoming feed in the upper stage
18
Figure 5. Results of slaking rate test applied to different types of
quicklime.
of a two-stage fluidized bed system. The calcium sulfate decomposition
and fuel combustion reactions are confined to the lower stage. Much of
the remaining sensible heat in the off-gas is subsequently transferred to
the incoming air in a conventional type of heat exchanger.
To calculate the fuel and air requirements for the process
represented by Figure 6, the design basis shown in Table 3 was employed.
For this study it was assumed that 5% of the calcium sulfate fed to the
reactor would be elutriated in the off-gas and would be recycled. Also it
was assumed that 5% excess air would be supplied to the reactor to
maintain an oxidizing zone in the upper part of the lower fluidized bed.
This much excess air was usually supplied to either of the two
19
Figure 6. Two-stage fluidized bed and heat recovery system for large-
scale applications
experimental reactors while demonstrating the two-zone version of the
process.
For the ‘design study, four cases were considered based on feed with
different amounts of water of crystallization and two types of fuel,
methane and petroleum coke. The fuel and air requirements of these
different cases are indicated in Table 4 as well as the sulfur dioxide
content of the off-gas. By feeding calcium sulfate hemihydrate it would
be possible to produce a product gas stream containing 11-12% sulfur
dioxide after drying. Feeding calcium sulfate dihydrate would increase
the fuel and air requirements by 11% and 15%, respectively, and reduce the
sulfur dioxide concentration of the dried product gas stream by 11%.
20
Table 3. Design basis for calculating fuel and air requirements
Parameter Value
Reactor feed
Reactor fuel
100% CaS04.112 H20
or 100% CaS04.2 H20
100% CH4 (natural gas)
or 100% C (petroleum coke)
Ambient temperature 77'F
Reaction temperature 2100'F
CaS04 desulfurization 98%
CaS04 recycle 5%
Excess air 5%
Table 4. Fuel and air requirements for converting CaS04 to CaO and SO2
at 2100'F
Feed mC m Air
Fuel SO2 cone., %
m CaS04 m CaS04 wet dry
CaS04*1/2 H20 CH4 0.99 7.48 9.4 12.3
CaS04-2 H20 CH4 1.11 8.61 7.4 10.9
CaS04*1/2 H20 Coke 1.94 7.24 10.6 11.3
CaS04.2 H20 Coke 2.17 8.38 8.3 10.0
21
Recently, the bench-scale reactor was converted to a two-stage
fluidized bed system to demonstrate the method of feed preheating
described above. Preliminary operation of the system has produced very
encouraging results.
CONCLUSIONS
Two versions of a process for recovering sulfur dioxide and quicklime
from either phosphogypsum or waste gypsum were demonstrated with a bench-
scale fluidized bed reactor. In one version coal was incorporated in the
feed to serve as a solid reductant whereas in another version powdered
coal or natural gas was the source of carbon monoxide and hydrogen which
served as gaseous reductants. Similar results were achieved with both
versions.
The second version was also demonstrated in a small pilot plant
operating on phosphogypsum or waste gypsum and using natural gas as a
source of energy and reductants. The results were similar to those
obtained with the bench-scale reactor under comparable conditions.
Operation of the pilot plant for periods ranging up to 30 hr. revealed
only one significant problem and that was due to build up of scale in the
off-gas system. Some success was achieved in controlling the build up of
scale and further measures for controlling scale are under consideration.
The quicklime produced from phosphogypsum was found to be a promising
soi l stabi l iz ing agent. Moreover the quicklime made from waste gypsum was
found to undergo hydration at a rate comparable to that of a commercial
quicklime. Therefore the by-product lime could be used in a number of
applications.
22
A preliminary design study based on the experimental results and a
practical method of heat recovery showed that a product gas stream
containing 11-12% sulfur dioxide could be produced from calcium sulfate
hemihydrate by the indicated process. Fuel and air requirements for such
a process were also noted.
ACKNOWLEDGEMENT
Experimental work on phosphogypsum was supported by the Engineering
Research Institute, Iowa State University, with funds made available by
the Florida Institute of Phosphate Research through a contract with
Zellars-Williams, Inc. (subsidiary of Jacobs Engineering Co.).
Experimental work on waste gypsum was also supported by the Engineering
Research Institute with funds provided by Hercules Inc. and the U.S.
Department of Defense. Additional support was provided by the Iowa State
Mining and Mineral Resources Research Institute funded by the Department
of the Interior's Mineral Institutes program and administered by the U.S.
Bureau of Mines under Allotment Grant G1164119.
REFERENCES
American Society for Testing and Materials. 1986. Annual Book of ASTMStandards. Cement; Lime; Gypsum. Vol. 04.01: 80-98. Designation:C110-85. Philadelphia, PA.
Chu, T. Y. and D. T. Davidson. 1960. Some Laboratory Tests for theEvaluation of Stabi l ized Soi ls . In: Methods for Testing EngineeringSoi ls . D. T. Davidson and associates. Iowa Highway Research BoardBulletin No. 21.
Morris, C. E., T. D. Wheelock and L. L. Smith. 1985. Processing wastegypsum in a two-zone fluidized bed reactor. Presented at nationalmeeting of the American Institute of Chemical Engineers, Chicago, IL,Nov. 10-15, 1985.
23
Smith, L. L., M. L. Fortney, C. E. Morris, T. D. Wheelock andJ. A. Carrazza. 1984. Resource recovery from wastewater treatmentsludge containing gypsum. Proceedings of the 1984 National WasteProcessing Conference. American Society of Mechanical Engineers.441-450.
Swift, W. M. and T. D. Wheelock, 1975. Decomposition of Calcium Sulfatein a Two-Zone Reactor. Industrial & Engineering Chemistry, ProcessDesign and Development. 14: 323-327.
Wheelock, T. D. 1978. Simultaneous Reductive and Oxidative Decompositionof Calcium Sulfate in the Same Fluidized Bed. U.S. Patent 4,102,989.
Wheelock, T. D. and C. E. Morris. 1986. Recovery of Sulfur Dioxide andLime from Waste Gypsum. TIZ-Fachberichte. 110(1): 37-46.
24
ENERGY SAVING PROCESS FOR
THERMAL DECOMPOSITION OF PHOSPHOGYPSUM AND OTHER
CALCIUM SULPHATES FOR THE
PRODUCTION OF H2SO4 AND
CEMFNT CLINKER BY APPLYING
THE CIRCULATING FLUID BED - CFB -
Authors: Dr. K.H. Kuehle, Head of Cement Plants Section
Mr. K.R. Knoesel, Manager
from LURGI GmbH
Gervinusstrasse 17/19
6000 Frankfurt/Main
F.R.G.
Presented at the2nd International Symposiumon PhosphogypsumMIAMI, Florida - USA
December 1986
25
INTRODUCTION
LURGI GmbH, worldwide well known as one of the biggest Engineering
Companies has experience with the Circulating Fluid Bed technology
since more than 20 years.
During these two decades LURGI installed over 50 industrial scale
applications of the CFB mainly for calcining of Alumina hydroxide and
cement raw meal as well as combustion of low grade coal for power
generation.
The know-how and experience gained by LURGI with the CFB led us to apply
this technology also for the thermal decomposition of Phosphogypsum and
other calcium sulphate bearing materials in order to produce simul-
taneously highly concentrated SO2 off-gases and cement clinker or any
other marketable solid by-product.
With regard to the european situation, we should also point out, that
there we mainly have to consider not Phosphogypsum but flue gas desul-
furization by-products. The projected quantity only for West-Germany has
been estimated to about 4.0 Mio. tpa of solid by-products coming from
desulfurization of coal fired power plants. From the year 1990 this huge
amount of gypsum should be utilized for various purposes.
26
PRINCIPLE OF THE FUNCTIONING OF THE CFB
With the following picture we would like to explain how the CFB is
working
(Figure 1)
27
In the literature the heat and mass transfer is stated as a function
of Reynolds (Re) and Prandtl (Pr) or of the Schmidt (SC) figure.
Nu = K1 x Rea x Prb
or
Sh = K2 x ReC x Scd
This means that the slip velocity (relative speed) exerts a con-
siderable influence on the particle as a function of the Reynolds
number.
In other words, the higher the slip velocity the better the heat
transfer from the gases to the solids.
In the conventional fluid bed "A" with a limited bed surface, the mean
slip velocity corresponds to the gas flow velocity through the
reactor.
The limited surface of the fluid bed is dissolved and case "B" results
if the flow velocity is increased. The fine particles are discharged
and the remaining particles are fluidized with a higher slip velocity.
This causes an intensification of the heat and mass transfer. The main
portion of the material does not leave the fluid bed.
Another increase in the fluidization velocity results in a maximum
slip velocity, the so-called case "C".
The entire column-shaped reactor is simultaneously filled with the
material. The density of the suspension increase from the reactor top
towards the bottom. The material discharged at the reactor top is
separated from the gas stream in an succeeding cyclone an then returned
to he reactor.
28
Another increase of the fluidization velocity results in the pneumatic
transport "D".
The slip velocity between the transport gas and particles flows down
and approaches value zero.
This is associated with a deterioration of the heat and mass transfer.
OBJECTIVES OF THE NEWLURGI-CFB APPLICATION
It is our declared aim to introduce into the market a technology
allowing to recover from waste materials like Phosphogypsum-Flue
gas-Desulphurization by-products etc. valuable raw materials and to
solve by these means the increasing problems of gypsum disposal and
further-more to save natural raw material resources.
Since any newly developed process technology should be competitive in
terms of technical and financial feasibility, it goes without saying
that our process shall be competitive with others not only technical
wise but also in terms of investments and production cost.
Based on a comparison with the well known "MÜller-KÜhne"-process we
are of the opinion that we fulfilled all the targets, since the
overall heat consumption is approx. 30 % lower, the electrical power
consumption approx. 40 % lower. The maintenance cost is lower because
the CFB is a statical device and the necessary fuel quantity for
decomposition and clinkering is splitted between the CFB and the
rotary kiln, therefore the latter can be of smaller dimensions since
it is only designed for clinkering purpose. This will further lead to
lower investment cost.
29
Some specific figures, based on a 1200 mtpd cement clinker production
and also 1200 mtpd of sulfuric acid are showing the economics of our
process. The overall heat consumption including gypsum drying -
calcining - decomposition and clinkering is less than 1900 Kcal/Kg
clinker or 7543 Btu/Kg clinker.
The power consumption for the drying - calcining - decomposition and
clinkering departments is less than 50 Kwh/To clinker or 67 HP/To
clinker.
The SO2 concentration in the CFB off gases is about 15 Vol. % based on
dry gases.
All off gases from the CFB - rotary kiln and clinker cooler are used
for drying - calcining and preheating purposes.
This closed heat cycle of the gypsum decomposition and clinkering
units as well as the drying and calcination stages ensures optimum
utilization of the waste heat, generated in the overall process.
DESCRIPTION OF THE CFB-PROCESS
Our new process consists mainly of the following three sections:
- Gypsum drying, calcining and preheating
- Decomposition and
- Clinkering
30
The next picture shows a CFB arrangement for gypsum decomposition and
production of a suitable cement clinker raw material
The CFB consists of the column-shaped reactor, the cyclone separator
and the seal pot mounted between cyclone and reactor. The entire feed
is in a highly expanded fluid bed in the reactor. The density of
suspension decreases from the reactor bottom to the top. The solids in
the gas are separated in the recycling cyclone and returned back to
the reactor by way of a seal pot. A partial stream of the treated
material is withdrawn according to the feed rate.
31
The dryed and calcined raw mix (Phosphogypsum - Clay - Sand etc.) is
fed into the CFB Reactor. The necessary reduction agent (either petrol
coke - or any other carbon-bearing material) is fed to the reactor
separately.
Approx. 2/3 of the reactor volume are operated under reducing
atmosphere.
The necessary combustion air is splitted into primary and secondary
air. The first is blown through the reactor bottom and the latter is
introduced via a circular gas duct into the upper part of the reactor.
The normal operating gas temperature of the CFB range between
1742°F (950°C) and 2012°F (1100°C).
One of the main advantages of the LURGI-Process for gypsum
decomposition is the separate gasside operation of the CFB and the
rotary kiln. This enables an accurate control of the operating
conditions required for gypsum decomposition or clinker production.
Adjustment of the necessary reducing atmosphere in the CFB reactor is
very simple due to the staged addition of the combustion air as
primary and secondary air.
The decomposition product is continuously fed to the rotary kiln via
the material discharge end, part of it being recycled into the reactor
via the seal pot.
The exact control of the product retention time in the CFB ensures
that the material routed to the rotary kiln is practically 100 %
decomposed so that only the clinkering reactions have to be carried
out in the latter.
32
DESCRIPTION OF OUR R + DPHOSPHOSGYPSUM DECOMPOSITION
PROGRAMME
In Dec. 85 we started a R + D program with the main target to prove
that the CFB is a proper device permitting to decompose thermally,
Phosphogypsum and all other calcium sulphate bearing material in the
presence of a reducing agent and either together with specific
additives for the cement clinker production or only using
Phosphogypsum.
One of the raw mix tested was containing:
84 % Phosphogypsum
12 % Clay
4 % Sand.
Additionally approx. 0.08 Kg Anthracit per Kg Phosphogypsum. The
working gas temperature of the CFB Reactor has been approx. 1940°F
(1060°C). Under these conditions we reached a decomposition degree of
nearly 99 %.
Further tests carried out with raw material mixtures containing
Phosphogypsum and Sand or Phosphogypsum with alumina A12O3 at various
temperatures gave a decomposition degree of about 85 %.
According to our experience it seems that the most economic way to
make use out of Phosphogypsum is to produce two products, i.e. highly
concentrated SO2 off-gases and cement clinker.
33
Another series of tests carried out during the month of August 1986,
confirmed the results obtained in Dec. 85.
The following picture shows our Pilot CFB facility.
When using Phosphogypsum as a raw material for cement clinker
production, one should absolutely investigate the quality of such a
product in terms of Fluorine and Phosphate P2O5 content; because the
quality of the clinker depends upon the amount of such impurities.
A high amount of P2O5 is affecting the early strength of the cement,
as it interferes with the formation of C S Tricalcium silicate, and a3
high amount of Fluorine in the clinker may cause setting failures.
34
Normally the Phosphogypsum should not contain over 0.6 % of P2O5 and
0.1 % of Fluorine based on dry Gypsum.
Several methods have been developed to reduce to an acceptable limit
these impurities. So far as our process is concerned the final step
to eliminate Fluorine is to use an indirect heated rotary calciner
through which we are routing the CFB SO2 rich gases. By these means we
are Teaching two targets, first using the heat content of the SO2 rich
gases to preheat the raw material mixture up to approx. 450 0C or
842 0F and second to evaporate the non water soluble fluorine content
by adding sulfuric acid and simultaneously to calcine the gypsum. The
resulting waste air stream containing water vapour, HF, SO3 and dust
is routed to a scrubber followed by a wet type ESP.
In addition to these impurities mentioned, certain Phosphogypsum may
also contain radioactive substances. This should be considered when
planning to use Phosphogypsum.
The next picture shows an example of a commercial scale plant
(Figure 4)
The raw mix consisting of hemi-hydrate and specific additives is fed
to the rotary calciner.
Since the calciner is an indirect working heat exchanger the SO2 rich
gases which are passed through the calciner are not mixed with the air
and vapour stream.
The calcined raw mixture is then routed to a three stage cyclone
preheater and behind the stage III fed into the CFB reactor.
From the CFB discharge the material mixture is continously fed to the
short rotary kiln where the clinkering takes place.
The clinker leaving the rotary kiln is then cooled down by
conventional means either grate cooler or rotary or planetary cooler.
As mentioned before the heat content of all off-gases is recovered for
either preheating - drying or calcining purposes.
36
CONCLUSIONS
According to our actual state of knowledge it is absolutely certain
that the CFB is an appropriated device for the thermal decomposition
of Phosphogypsum. Therefore we are of the opinion that we are able to
offer various solutions for the gypsum recovery.
The next step to undertake is to install together with an industrial
partner a commercial scale plant.
37
FEDMIS SULPHURIC ACID / CEMENT
FROM PHOSPHOGYPSUM PROCESS
by D A CLUR M Sc
Deputy General Manager Development
Fedmis Division of Sentrachem Ltd
ABSTRACT
Fedmis Division of Sentrachem Ltd is one of the largest fertilizerproducers in the Republic of South Africa. It has been producingphosphoric acid from Foskor rock concentrate at Phalaborwa since 1964.The phosphogypsum is dumped on nearby land. In 1970 the only economicalsolution to this disposal problem was the production of sulphuric acidand portland cement from this byproduct.
The 70 000 tons/year plant was commissioned in 1972 and was uniquein two respects. It was the only cement/acid plant in the world(i) based solely on phosphogypsum, and (ii) using a pulverised coal-fired rotary kiln. As a result, considerable development work by Fedmiswas necessary to ensure the technical and economic viability of theprocess. The plant produces 98% H2SO4 and cement which compares favour-ably in quality with conventional limestone-based cements.
An 800 t/day plant would cost about R120 million today, and undercurrent South African conditions a satisfactory return can be expected.Forecast sulphur shortages and ever tightening environmental controlsfavour the future of the process.
INTRODUCTION
Fedmis Division of Sentrachem Ltd is one of the largest fertilizerproducers in the Republic of South Africa. Its phosphate operation islocated at Phalaborwa in the N.E. Transvaal. Phalaborwa (population50 000) owes its importance to the occurrence of valuable minerals,notably phosphate and copper ore, in a volcanic pipe covering a relativelysmall area close to the town. Palabora Mining Co (PMC) exploits thecopper and the Phosphate Development Corporation (Foskor) produces 36%and 39% P2O5 rock concentrate from the igneous fluo-apatite (Foskorite).
Fedmis has been producing fertilizer phosphates from this concentratesince 1964, when it commissioned the first South African phosphoricacid and triple superphosphate plants at Phalaborwa. These intermediatesallowed significant up-grading of local fertilizer concentrations.
39
Today ammonium phosphates and monocalcium phosphate for animal feedsare also produced at the factory. We are now operating two Prayondi-hydrate plants and a Central-Prayon hemi-hydrate plant, with a totalcapacity of about 390 000 t P2O5/year.
Phosphate rock concentrate is obtained from Foskor as a slurry. Sulphuricacid is obtained from our own sulphur-burning and cement/acid plants,from nearby PMC's copper smelter and from the gold mines' pyrites plants.The high local and imported sulphur price has made mines acid competitivein spite of the loss of steam by-production.
As our phosphoric acid plants are landlocked, the phosphogypsum isdumped on nearby land. The effluent is completely contained. The seepagefrom the dumps is collected in a system of impounding dams and recycledto the plants. To date, about 12 million tons phosphogypsum has beendumped in this way, covering a land area of about 200 hectares. Thisdisposal problem is growing by about a million tons annually. As disposalcosts are increasing and suitable land close to the factory is becomingmore difficult to obtain, we are continually investigating possible waysof profitably utilising our phosphogypsum.
As can be seen in Table 1, we are currently putting roughly 40% ofour current phosphogypsum production to profitable use.
In 1970, however, the only profitable solution to our disposal problemwas the cement/acid process. Following extensive development work, whichculminated with a 5-day trial on the Chemie-Linz cement/acid plant, itwas decided to build a 300 t/d unit at Phalaborwa, using Chemie-Linztechnology. This plant has been operating for 15 years, and the purposeof this paper is to discuss this experience and the economics andfuture of the process.
40
PROCESS
The cement/acid plant, commissioned in 1972, was unique in two respects :It was the only plant in the world (i) based solely on phosphogypsum and(ii) using a coal-fired rotary kiln. Both of these increased Fedmis riskas the impurities could affect both the operation of the plant and thequality of the cement produced. These points will be referred to later.
The plant, which is shown diagrammatically in Fig 1, is based on thewell-known Muller-Kuhne process.
RAW MEAL
Calcium sulphate hemi-hydrate is calcined to the anhydrite in a turbo-tray drier and blended in correct proportion with milled reductant(usually coke), clay soil (mined and homogenised on the Fedmis property)and PMC magnetite, to provide the raw meal fed to the kiln. As a highash (18%) coal is used for kiln firing, this must be taken into accountwhen formulating the raw meal composition. Good kiln operation and hencegood clinker production depends on a constant raw material compositionand accurate metering to the kiln. To assist in this respect, thepulverised coal is milled prior to metering to the firing lance.
ROTARY KILN
The 107 m long rotary kiln has an internal diameter of 4, 2 m and isprovided with a number of refractory Dietze Bridge heat exchangers anda bank of chains (near the back end) to improve the thermal efficiencyof the process. Better thermal efficiencies can be achieved with acyclone-type preheater, as used on conventional cement plants, but sucha system only became available after our plant was commissioned.
In the kiln the anhydrite is reduced according to the following equations :
That is, the reduction occurs in two stages via calcium sulphide, anexcess of which can lead to problems in kiln operation (due to itsrelatively low melting point).
The lime produced subsequently reacts with the clay minerals present(Al2O3, Fe2O3, SiO2) in the clinkering zone at 1450°C to produce cementclinker.
The kiln is operated under slightly reducing conditions to ensure completereduction of the sulphate by the coke. Energy-wise, about 8,3 GJ/tclinker is consumed, compared with about 3,4 GJ/t clinker for a typicallimestone cement preheater kiln. The Fedmis consumption could bereduced to about 5,8 GJ/t clinker by retro-fitting a preheater, and thisis seriously being considered. The higher energy consumption of thecement/acid process arises mainly from the difference in heats ofreaction of the two processes ((∆ H° limestone cement = 42, 6 Kcal/mole CaCO3).
storage.The clinker is cooled in a grate cooler and supplied to the cement grindingplant via intermediate
SULPHURIC ACID PLANT
The dust-laden gas from the kiln, containing about 7% SO2, is fed to ametallurgical-type gas cleaning system and contact plant producing 98%H2SO4 . The recovered dust is recycled to the kiln. The sulphuric acidplant, in addition, is designed to burn sulphur to its full capacity, asa precautionary measure. Although this facility is seldom used, a minimumquantity of sulphur is still burnt in parallel with the kiln SO2. By thismeans the sulphur burner is kept hot and the converter can be kept on lineduring kiln stoppages. A preheater, together with its liquid fuel costs,was hence avoided.
Approximately 1 ton H2SO4 is produced per ton of clinker produced.
CEMENT PLANT
This is a 600 t/d conventional cement grinding unit. The clinker isground together with 3-4% of Fedmis retarder gypsum (and blast furnaceslag when required) in a closed circuit ball mill, to produce good qualitycement. This is despatched both in bulk and in 50 kg valve bags. Anyspare capacity is used for cement production from limestone-based clinker.
DEVELOPMENT
The Fedmis process includes hemihydrate crystallisation, calcinationto anhydrite, production of clinker, sulphuric acid and finally cement.These units are all interdependent, and, as most of the technology wasnew, we experienced a long learning curve. Most of these units requireddevelopment work. Either equipment had to be redesigned (eg to avoidserious build-up in the turbo-tray calciner), or additional capacity wasrequired (eg in the gas cleaning section).
Operating experience showed that good quality clinker depended on accuratecontrol of raw meal composition and proportioning, and skilled kiln burning.In the clinkering operation the prime objective is to reduce the SO,content to a value such that the kiln and its refractory lining is notover-heated, while at the same time avoiding over-reduction and the presenceof a significant sulphide content in the clinker.
The presence of significant amounts of CaSO 4 or CaS in the clinker mustbe avoided. Both these compounds have lower melting points than theclinkering temperature, and can lead to a clinker melt leaving the kilnwith a severe effect on the clinker cooler. Experience has taught thekiln operators to recognise the appearance of such liquid phases byobserving the conditions in the clinkering zone. We have found no instru-mentation to replace this operator art. In the short term, these upsetconditions can be corrected by increasing or decreasing the coal firingrate, depending on whether more or less reduction is needed. W R Scurr (1983)gives further details on coal-firing of our kiln.
Although the life of the refractory lining in the clinkering zone of thekiln has improved significantly over the years, we still have some way togo to match that found in conventional limestone cement kilns. Work iscontinuing in this direction.
Generally, however, our cement/acid plant is operating well, and mainte-nance is normal for this type of operation. Work is in progress to expandthe capacity of this plant by 25%.
PHOSPHOGYPSUM
As the success of the process depends on the quality of the cementproduced, the quality of the phosphogypsum feedstock is important. Theimpurities which adversely affect cement strength development are phosphates and fluorides. However, some fluoride in the presence of phosphate isbeneficial (Gutt and Smith, 1970).
Phosphates reduce early cement strengths by forming a dicalcium silicate- tricalcium phosphate solid solution at the expense of tricalciumsilicate, the rapid hydration of which is largely responsible for theearly strength development. Small amounts of fluoride tend to counteractthe deleterious effect of phosphates by a complex mechanism in which thecalcium silicates are formed via intermediate silicofluorides. Work bythe British Building Research Station in the late 1960's, concluded that acement meeting the requirement of British Standard Specification BS 12 :1958 could be made from phosphogypsum by the cement/acid process, providedthat the P2O5 and F levels in the clinker did not exceed 2, 5% and 1%respectively. This corresponds to limits of 1, 2% P2O5 and 0,45% F inthe phosphogypsum (dry basis).
Typical analyses of our phosphogypsum and hemi-hydrate are shown inTable 2 below :
43
As both the above criterion and a full-scale plant trial at Chemie Linzindicated that the P2O5 level in our phosphogypsum was just too high toconsistently produce an acceptable cement, we converted one of our di-hydrate-based phosphoric acid plants to produce the purer hemi-hydrate(0,6% P2O5) as the process feedstock , and this has been successfullyused ever since. A reduction in calcining capacity was also possiblewith this material.
The unusually high lanthanide content in our phosphogypsum (0, 23 -0,272 Ln2O3) was shown by the British Building Research Station to haveno adverse effect on cement strength development (Gutt and Smith, 1970).
CEMENT QUALITY
Fedmis cement has carried the South African Bureau of Standards mark since1974. Local standard SABS 471 is based on BS 12. At that early stage,however, a portion of our cement produced during start-ups and out-of-control periods was separated as second grade material for use as a roadstabilizing cement. But, by 1978, our entire cement production compliedwith and carried the SABS mark. To be competitive, however, our cementhas to meet the high compressive strength specifications of the industry.As can be seen in Table 3, this has been achieved. Our cement is groundsomewhat finer than conventional cement, to ensure that our good strengthdevelopment is maintained.
44
It is interesting to note that the grindability on our cement clinkeris no different from that of conventional cement clinker. Althoughnatural gypsum was originally used as cement retarder, this has nowbeen successfully replaced by our own retarder, made from phosphogypsumby a Fedmis process.
We are now also making a cement containing 15% blast furnace slag.
The quality of Fedmis cement compares favourably with that of the locallimestone-based cements, and is used in all classes of building construc-tion and civil engineering.
ECONOMICS
When considering the economics of the cement/acid process, it isperhaps better to consider it as a cement producer with sulphuric acidas the by-product, rather than the commonly accepted opposite. Thistends to bring the relatively high capital and energy cost of the processinto perspective and focuses attention on the fact that roughly 80-90%of plant operation and control laboratory effort goes toward achievingthe desired cement quantity and quality.
CAPITAL COST
Our 300 t/d cement/acid plant cost about R9 million in 1972 when R1 = $1,4. The plant is small by cement standards, and reflects the state of develop-ment of the process during the 1960's. Since then, however, most of thetechnical problems associated with the use of phosphogypsum as feedstockhave largely been solved, and single stream plants over 1000 t/d havebeen mooted. For the purpose of this discussion, an 800 t/d plant has beenassumed as a reasonable scale-up of our present facility.
The capital cost of such a plant has been estimated to be about R120million under present South African conditions (equivalent to U.S. $100million by engineering factor). A breakdown of this cost is given below :
R millions
Pre-heater kiln plant 85Hemi-hydrate calcinationRaw meal preparationCoal preparationPre-heater kilnGas treatmentStorage
Sulphuric acid plant 20
Cement grinding plant 15
R120
Roughly speaking, this cost is five times that of a sulphur-burning, ortwice that of a pyrites-burning sulphuric acid plant of the same capacity,but it should be remembered that cement is also produced by our process.
45
PROFITABILITY
The economic viability of the process depends on the location of theplant and the cement, acid and fuel prices obtaining in that market,i.e. the process is site sensitive.
The profitability of the Fedmis plant is favoured by the relativelyhigh cost of cement in the N.E. Transvaal market, owing to the considerabledistance of the conventional cement plants from this market (at least 500km). The price for cement in this market is about R110/ton. On the acidside, our profitability is favoured by the relatively high cost ofsulphur, which is imported from Vancouver. Over recent years the weakenedRand (currently worth 49 U.S. cents) has made sulphur an even moreexpensive alternative: currently R350/t delivered. Limited quantitiesof local sulphur are available ex Sasol at import parity prices. Hightransport costs have until recently made pyrites-based acid uncompetitiveat Phalaborwa. Low cost coal (R35/t delivered), mined about 300 kmfrom the factory, also contributes to the viability of the Fedmis plant.
A disadvantage of this process compared with a sulphur-burning acidplant is the lack of steam production. Steam raising plant is thereforenecessary for phosphoric acid concentration. This difficulty alsoapplies when using bought-out sulphuric acid (eg from the gold mines).
All the sulphuric acid produced on our plant is recycled to phosphoricacid production,also,
and, as the acid plant is designed to burn sulphurif necessary, the acid is valued at the equivalent sulphur price.
Again, in discussing the profitability of the process, we refer to an800 t/d unit.experience.
Costs are reasonable estimates based on our operating
RAW MATERIAL AND UTILITIES CONSUMPTION
These are given in Table 4 below :
46
INCOME STATEMENT
As a guide to the economics of the process, an income statement whichcould be expected for an 800 t/d cement/acid plant operating under SouthAfrican conditions at Phalaborwa, is shown in Table 5 below :
During 1986, largely due to the depreciation of the Rand, the cost ofsulphur delivered to Phalaborwa was over R400/t or R140/t H2SO4 equivalent.At this sulphur price, the DCF return increases to almost 30% (at the samecement price) and shows the importance of rate of exchange on the processprofitability.
The cost structure is characterised by a high marginal contribution andhigh fixed costs. The process is therefore vulnerable to under-utilisa-tion of plant, personnel and infrastructure. As the acid is recycled,the only external effect on utilisation would be cement sales, and it isvital therefore that a steady demand for cement is secured. On the otherhand, significant economies of scale are in principle possible. Hencethe interest in larger plants. One of the advantages of our presentplant expansion programme will be better cement plant utilisation.
The cost of energy (coal plus electricity) is about 15% of turnover.The relatively low cost of coal (R1, 31/GJ) contributes significantly tothe viability of the process.
FUTURE
As the technical problems of producing a good quality Portland cementfrom, phosphogypsum have largely been solved, the future of the processwould seem to depend on local economic and environmental factors.
Over the years, interest in the cement/acid process has varied with thesulphur price. Sulphur shortages forecast by organisations such as Agri-chemicals Economic Research of Vancouver, Canada (1986), which couldcause prices to increase significantly, would favour the process, parti-cularly where sulphur is imported from distant sources. Foreign exchangefactors could also play an important part here.
Environmental pressures on phosphoric acid producers are increasing,and disposal of phosphogypsum is being more stringently controlled.Disposal costs are therefore increasing. A few years ago a local pipe-line for sea disposal of phospho-gypsum was reported to have cost aboutR50 million. This is about half the cost of an 800 t/d cement/acid plant.
The Fedmis know-how for the operation of the cement/acid process entirelyon coal as its heat source, could in many situations improve the overalleconomics of the process compared with oil or other liquid fuels.
We believe that our coal-based cement/acid process has a worthwhilecontribution to make to the solution of the phosphogypsum problem, par-ticularly in areas where distance from sulphur and cement suppliersmakes these commodities relatively expensive. Fedmis would be happy todiscuss its know-how with any prospective user of this process.
ACKNOWLEDGEMENT
The author wishes to thank Fedmis Division of Sentrachem Ltd forpermission to present this paper.
48
REFERENCES
GUTT, W and M A SMITH 1970 The use of phosphogypsum as a rawmaterial in the manufacture of Portland cement. EN 58/70.Building Research Station, Ministry of Public Buildingsand Works, Great Britain.
GUTT, W and M A SMITH 1970 Cerium as a minor component in cementmanufacture. CEMENT TECHNOLOGY 1 : 3-7.
SCURR, W R 1983 The use of coal for recycling phosphogypsum; theexperience of Sentrachem Ltd. Raw Materials in South Africa,International Fertilizer Industry Association.
49
PHOSPHOGYPSUM TO SULFURIC ACID
WITH COGENERATION - A COMPETITIVE EDGE
T. J. KENDRON
DAVY MCKEE CORPORATION
G. M. LLOYD
FLORIDA INSTITUTE OF PHOSPHATE RESEARCH
Presented at theSecond International.
Symposium on PhosphogypsumMiami, Florida - December, 1986
51
1. ABSTRACT
Since 1982, the Davy McKee Corporation (DMC) and the FloridaInstitute of Phosphate Research (FIPR) have been involved in acooperative program involving the utilization of phosphogypsum. Theobject of this joint effort is to demonstrate a practical process for thethermal decomposition of phosphogypsum for the production of sulfuricacid and a saleable, solid byproduct. The innovative application of thecommercially proven DMC circular grate, used as the process “reactor”for this phosphogypsum utilization system, has been extensively testedusing the same laboratory equipment and procedures used to completeover 60 commercial grate designs. The circular grate system which hasbeen successfully applied to the iron and steel industry, exhibitssignificant advantages, both technically and economically, whencompared to previously demonstrated or other proposed phosphogypsumdecomposition processes.
The phosphate industry and involved environmental agencies ailappreciate that a successful phosphogypsum decomposition process willprovide the capability to control costs of sulfuric acid and reduce theimpact of phosphogypsum deposits on the environment.
The DMC/FIPR process incorporates a mixture of phosphogypsum, asolid carbon source, waste phosphatic clays, pyrites and other additiveswhich depend upon specific process parameters. The mixture is fed ontothe rotating, circular grate and is processed in a series of sealed zones onthe grate. The raw product gas, high in SO 2, is collected for use as thefeed gas for a conventional metallurgical-type sulfuric acid plant. Thesintered, solid byproduct, which remains on the grate after reaction, isdischarged in a dry form from the grate by a tilting pan mechanism.The solid byproduct is essentially inert and has been evaluated for anumber of possible uses. The use of pyrites has further strengthened thebyproduct solid and has given it mechanical and chemical propertieswhich make it suitable for large aggregate use in road construction.Depending on the feed mix “recipe”, other uses for the solids beingevaluated are: in skid-resistant road surfaces, as a soil cement and inconsolidation of phosphatic slimes.
Recent ly FIPR and DMC have been __ evaluat ing a coal-based,cogentration system which can be integrated with the circular grate
phosphogypsum process. This system allows the use of a wide range ofcarbon sources such as coal and coal wastes and the production ofenergy equivalent to or exceeding that produced in sulfur-burningsulfuric acid plants. The majority of this paper discusses the results ofthe evaluations completed to date.
The DMC/FIPR process when integrated with the coal/cogen systemderives advantages from both the circular grate, previously demonstratedin the ferrous industries, and coal processing techniques developed inthe energy and chemicals sectors. The major advantages of this
integrated system include:
52
- Circular Grate:
Grate sintering provides low capital and operating cost perunit of material processed.
Circular des ign offers mul t ip le react ion zones wi thadequate sealing to allow varying conditions and toprevent fugitive air leakage.
Reactants arc thoroughly mixed in prepared feed.
Feed is not redistributed once placed on grate.
. Coal/Cogen:
Steam, electric power, clean fuel gas and/or process heatavailable for use in other process facilities.
Low grade, high sulfur carbon sources such as coal andcoal wastes are suitable.
Under the present FIPR contract extension, DMC is completing capitaland operating cost estimates for a commercial scale circular grate system.Chemetics International Company is assisting with the technical and costinformation being developed for the gas cleaning system and sulfuricacid plant modifications.
The current capital and operating cost estimates for a circular gratephosphogypsum facility integrated with the coal/cogen system willproduce sulfuric acid for about $17/ton on a breakeven basis andrequires a value of about $31/ton to achieve a 15% after tax return oninvestment.
In summary, successful commercialization of this technology wouldprovide several significant socio-economic benefits:
. Stabilize phosphate industry raw material prices
. Improve competitiveness opposite foreign manufacturers
. Reduce environmental impact of gypsum and waste clays
. Decrease acid-drainage problems of coal and certain metalsoperations by consumption of waste pyrites and carbon containingwastes
. Provide alternative sourcing for the projected world shortage ofsulfur supplies.
53
2. PROGRAM BACKGROUND
The technology for the production of sulfuric acid from calcium sulfatehas been in various stages of development and commercial applicationsince World War I. (1) The literature has dealt with numerous processapproaches for this, method of acid production with particular interestin the recycling of phosphogypsum for use as a raw material. (2,3,4,6)
It is estimated that over 400 million tons of phosphogypsum are(4)currently stacked in piles in the Central Florida area of the U.S. and
that domestic U.S. stockpiles arc accumulating at a rate of about 30(5)million tons per year.
In 1982, a cooperative program was initiated, combining the resources ofthe Florida Institute of Phosphate Research (FIPR) and Davy McKeeCorporation (DMC). The ultimate goal of this program is to demonstratean economically feasible process for the thermal decomposition ofphosphogypsum for production of sulfuric acid and a saleable solidbyproduct. The program was recent ly expanded to include theinvestigation of an integrated coal-based/cogeneration system to simultaneously provide a suitable carbon source for the circular grateprocess and to replace the energy lost due to the elimination of sulfurburning. The reporting of results from this ongoing evaluation are themajor subject of this paper.
2.1 Phase I and Phase II
The Phase I, completed in 1982, essentially demonstrated that thecircular grate approach could be successfully applied to thethermal decomposition of phosphogypsum into sulfur dioxide andvitrified solids, and more specifically, that a sintering processcould be applied to convert phosphogypsum into a sulfur dioxidecontaining gas and a sinter or aggregate. Although additionallaboratory investigation was deemed to be required prior todrawing more definitive conclusions, the sintering processexhibited promise as the most viable of the circular gratetechnologies for the decomposition of phosphogypsum due to its(1) potentially high grate factor (system capacity); (2) thermalefficiency; (3) SO2 content of the gas; and (4) degree ofdecomposition demonstrated.
In April 1984, FIPR approved funds for the Phase II program.The title of this project is “Phosphogypsum Conversion UtilizingCircular Grate Technology, Phase II”. The Phase II scope of workis separated into two parts: Part 1 - Laboratory Investigation,and Part 2 - Economic Analysis.
The specific objectives’ for Phase II, Part 1 were to improve onthe results obtained in Phase I of the program. The key goalsconcentrated on reducing the projected capital and operatingcosts.
54
To initiate the Phase II activities, raw materials were selected andprocured using the same criteria and techniques employed inPhase I. Representative samples of phosphogypsum, petroleumcoke, coke breeze, tailings sand, phosphatic clays and limestonewere collected and t r ans fe r r ed to Cleveland ResearchLaboratories, Inc. (CRL), Cleveland, Ohio, the subcontractorselected for the sintering test laboratory work.
A total of 86 sintering tests were completed during this phase andcan be generally categorized in the following groups: (1) benchmark testing, (2) updraft process development, (3) downdraftprocess development, (4) aggregate byproduct development, and(5) solid byproduct sample production.
Additional testing was completed by CRL on the sintered solidsto evaluate certain mechanical properties with respect to use as aroad construction material. Other subcontractors, i.e., Ardaman &Associates, Florida Mining and Materials, Pembroke Laboratory,and Macasphalt completed tests on the solid byproduct to evaluateits mechanical and chemical acceptability for use as roadconstruction materials and as a low-grade lime.
The key results of the Phase II laboratory investigation can besummarized as follows:
Circular Grate and Gas Investigations
. The circular grate system design capacity which resultedfrom the Phase I work was significantly increased in PhaseII which greatly lowers the grate’s projected installed cost.
. Predrying of the feed material is not required prior toignition.
. A 12' bed depth (up from 8" in Phase I) is operable withdeeper beds (up to 16") probable.
. Sulfur recovery from the raw feed material will be in therange of 80% to 90%.
. The SO2 strength of the offgas is sufficient for sulfuricacid production.
Solid By-Product Utilization
. Tests on the large aggregate for abrasion resistance showedthe mater ia ls as produced were marginal for use asaggregate.
55
. The material produced using additives of limestone andclay exhibited properties similar to flyash. Use in soilcement or as a flyash replacement in cement for road useis possible.
. Potential chemical uses for this material are:
. Neutra l izat ion of phosphogypsum or otherphosphate industry wastes.
. Consolidation of phosphatic slimes
. As a partial replacement for high grade limestoneor lime currently used in SO2 scrubbing systems
2.2 Pyrites Addition Testing
After complet ion of the Phase I I es t imate and subsequenteconomic analysis of the commercial-scale phosphogypsum tosulfuric acid facility, it was obvious that certain refinementswould lead to much improved economic viability. Both capitaland operating costs (per ton of products produced) would besignificantly lowered if four objectives could be achieved:
1. Increase mechanical strength of the byproduct solids.
2. Reduce the carbon usage.
3. Increase SO2 strength from 6% to 9% or above (on alevelized basis with SO 2/O2 ratio approximately 1:1).
4. Increase sulfur removal from phosphogypsum.
A single answer to this was hypothesized by DMC and FIPR;pyrites addition to feed mix. A preliminary survey of pyritesavailability and projected cost for Florida delivery provided somesurprising results Large amounts of pyrites and pyritic-richwaste materials containing some carbon and coal ash are availablefor economic delivery to Florida. Obviously, the carbon and ashconstituents are not detrimental to the circular grate process withthe high heat input requirements and aggregate production mode.A series of sintering tests utilizing pyrites in the feed mix wereconducted in early 1986. These tests confirmed the attainment ofall four objectives. The increase in mechanical strength andstability of the byproduct solids was most significant. Theleachability of trace heavy metals was also tested and found to belower than specified for the EP toxicity tests. A fact sheetsummarizing the tests completed on the solids is included in theAppendix.
56
This development is considered a significant breakthrough forthis process and DMC and FIPR have appl ied for patentp r o t e c t i o n .
2.3 Coal/Cogeneration Study
In July 1986, a supplemental study was initiated by FIPR andDMC to evaluate the integration of a coal-based, cogenerationsystem with the circular grate phosphogypsum process. The goal of the study is to further improve the promising econoprojected for the phosphogypsum/pyrite circular grate system andto provide even greater proces flexibility. The incentives for theincorporation of this system are:
1 . Elimination of the importation of petroleum coke, electricpower, circular grate ignition fuel and a portion or all ofthe pyrites.
2. Production of steam and/or electric power equivalent to/orgreater than a sulfur burning sulfuric acid plant.
3. Broadening the suitable carbon and sulfur sources to include high sulfur coals and/or coal wastes.
4. Greater utilization of existing sulfur burning, sulfuricacid equipment including cogeneration facilities.
As in the case of pyrites addition, DMC and, FIPR have appliedfor patent protection for the coal/cogeneration system.
A. number of screening evaluations were completed in, order toestablish the best overall system for the final study. Theoptimum system has been selected and the detailed estimate andfinal report are being completed and should be issued to FIPR inJanuary 1987. Section 3 presents the results of the screeningevaluation and Section 4 presents a summary of the economicevaluations completed to date.
2.4
Two major advantages of the c i rcular grate reactor are i tsflexibility to employ variable feedstocks and its overall energyefficiency. Thus the design of the reactor can incorporate anumber of feed and recycle streams which include extensivewaste heat recovery schemes. In order to fully evaluate theseprocess alternatives, it was decided to use the ASPEN program tosimulate the reactor system.
57
ASPEN (Advanced System for Process ENgineering) is a processsimulator developed by MIT under the funding of the U.S.Department of Energy. It employs a modular approach to processsimulation which treats each equipment item or unit operation asa block and connects process streams between the blocks. Thematerial and energy balances are performed based on the user’sspecification of the feed streams and unit operation blocks. Someapparent advantages in using the ASPEN simulator for designinga new process from laboratory test work are:
1. Test results and process “know-how” can be directlyapplied to the reactor block specification. Material andenergy balances can be performed to evaluate the validityof test data, and the impact of the test data variation onthe entire plant can be efficiently established.
2.
3.
The
Test systems are usually designed to emulate the novelportion of the process without total process integration.The ASPEN simulation allows the evaluation of the impactof integration, such as continuous recycle and heatrecovery, on the overall system.
With a developing process, ASPEN allows the user tomodify process schemes, alter feedstocks and convergerecycle streams to achieve a complete material/energybalance. The “what ifs” o f an a l t e rna t ive can bequantitatively and efficiently evaluated.
simulation presented in Figure 2.4 is currently beingexpanded to directly include integration of the coal processing,gas cleaning and sulfuric acid units to provide overall heat andmaterial balances.
58
3. COAL/COGENERATION TECHNICAL INFORMATION
The completion of an economic evaluation for a complete 2800 STPDsulfuric acid plant based on the decomposition of phosphogypsum withpyrites added to the mix resulted in a breakeven sulfuric acid price of
(8).$35/ton This price reflected the capital investment for the circulargrate system, gas cleaning and a metallurgical-type sulfuric acid plant.This scheme produced about .3 tons of steam per ton of sulfuricproduced, or slightly less than 25% of the steam produced by a typicalsulfur-burning acid plant. FIPR suggested that DMC explore possiblemethods of producing more available energy for export, keeping in mindthe importance of the overall process economics. FIPR also suggestedthat a retrofit approach, i.e., maximize the use of existing sulfuric acidplant facilities, be considered as this is the most likely scenario for anexisting phosphate complex.
The proposed “coal/cogeneration” process configuration was derived aspart of an investigation to utilize coal to replace petroleum coke or cokebreeze. Previous phosphogypsum sintering testwork with coal as acarbon source had indicated that the higher levels of volatile matter,greater than 15%, interfered with the decomposition reaction andp r e s e n t e d d o w n s t r e a m g a s h a n d l i n g p r o b l e m s w h i c h w o u l d b eunacceptable in commercial grate operation. DMC proposed todevolatize the coal to produce a low volatile, reactive char for use onthe circular grate. This processing step produces a number of beneficialsecondary effects wh ich fu r t he r enhance t he v i ab i l i t y o f t h ecoal/cogeneration circular grate system for certain installations.
3.1 General Description
The removal of volatile matter from coal can be approached witha number of commercia l processes used pr imary for coalgasification. The most common, commercially demonstratedprocesses are: (I) fixed bed, (2) fluid-bed, and (3) entrained flow.E a c h o f t h e s e p roces se s ha s i nhe ren t advan t ages anddisadvantages. The nature of the feedstock, the desired energysplit between char and by-product gas stream, the characteristicsof the gas stream and the need for oxygen or oxygen-enrichmentare some of the key criteria for process selection.
It was decided early in our study to utilize a fluid-bed system forthe following reasons:
1.
2.
Coal feed size acceptable ranges from fines to 3/8”.
A wide variety of coals, including caking coals, lignites.and coal wastes, is suitable for use.
3. A uniform bed temperature, which can range from 1650°F- 2000°F, results in complete cracking of volatiles so tars
60
and oils are not contained in by-product gas.
4. Fluid-bed gasifiers have operated commercially with air.
A number of fluid-bed systems arc commercially offered(9) andfinal selection will depend on site specific parameters.
Other considerations such as need for higher BTU gas, synthesisg a s o r ava i l ab i l i t y o f l ow cos t oxygen cou ld r e su l t i nconsideration of processes other than air-blown fluid-bedgasification.
Fuel Gas Utilization
The efficient USC of the by-product fuel gas is important to theeconomic viability of coal/cogen configuration. A number ofapproaches, i.e., auxiliary steam boilers, reuse of sulfur furnacesas boilers and a gas turbine were evaluated. Using the criteriaestablished, it appears that the gas turbine provides by far themost flexible, efficient system
The fuel gas can be utilized to produce a mix of electric power, steam at various pressure levels and process heat. In the selected
case, the heating of the SO2 gas upstream of the sulfuric acidconverter is achieved with waste heat from the gas turbineexhaust.
3.2 Commercial Plant Description
As part of the continuing work on the circular grate process,DMC and FIPR art completing an engineering study whichevaluated a full-scale, retrofit of a coal-based, phosphogypsum tosulfuric acid facility. A block flow diagram, Figure 3.2.1, of theprocess route used for the evaluation is included in this section.
Key assumptions for the evaluation are summarized as follows:
1. High sulfur, eastern bituminous coal is both the carbonsource for the circular grate reactions and the energysource for the complex.
2 Adequate electric power is produced to provide power forthe new facilities and the existing process units.
3. Clean SO2 gas is produced to allow production of 2800STPD of sulfuric acid.
4. An existing sulfuric acid plant with a cogeneration unit isavailable.
The raw material preparation, circular grate system and gasc l e a n i n g s e c t i o n s w e r e d e s c r i b e d i n d e t a i l i n a p r e v i o u s
61
publication (8) released in April 1986, at the spring nationalAIChE meeting. T w o m a j o r c h a n g e s i n t h e e q u i p m e n tarrangement concept as presented in April have been incorporatedinto the retrofit design basis. In order to minimize plot arearequirements, the layout of the raw material preparation sectionhas been modified to a more vertical arrangement similar to DAPgranulation plant designs. Another area saving approach involvescombining the sintering and sinter cooling operations onto thecircular grate unit. The study plot plan, Figure 3-2-2, includesthese revisions in addition to the coal processing, fuel gascleaning and gas turbine systems. It should be noted that interestin the use of coal-based fuel gas for combined cycle operationshas increased in recent years as an attractive solution for costeffective utilizationacceptable manner.(9)
of high sulfur coals in an environmentally
Coal Processing
The coal processing section includes a number of processoperations which convert the incoming coal to. (1) a low volatilechar suitable for use on the circular grate and (2) a clean fuelgas.
The facility is designed to receive coal by rail. Unit trains can .be unloaded and the coal stored in both live and dead storage.The reclaimed coal is crushed to "l/4” x 0 and transferred to thecoal devolatization unit. This unit consists of a single, fluid-bedreactor which produces the char and hot raw fuel gas from thecrushed coal, air and steam. The char is cooled and transferredto the feed preparation unit in the circular grate facility.
62
Fuel Gas Processing
The raw fuel gas i s cooled to recover sensible heat andparticulate is removed prior to the removal and recovery ofsulfur compounds. A number of raw gas sulfur recovery systemsare available. The LoCat system (11) as supplied by ARI offers anumber of advantages for our system and has been selected as thebasis for our design.
The clean fuel gas is used as ignition fuel for the circular grate,but most is compressed and used as fuel for a gas turbine system.The sulfur removed from the raw fuel gas used on the circulargrate to supplement SO2 production or sold to other users.
Gas Turbine System
The gas turbine system includes a gas turbine/power generationpackage and a heat recovery system. A G.E. Frame 5 producing anominal 30 MW is the basis for our design. The compressed fuelgas is combusted in the turbine and the hot exhaust gas is utilizedto raise steam and preheat the feed gas for the sulfuric acidplant. The electric power produced is distributed for overall
plant use.
Cost E .valuation Basis
The capital and operating costs which have been developed andare the basis for the Section 4 evaluations include the following(see Figure 3.2.1):
Capital Cost: All new facilities located betweenlimit designations A-A and B-B.
Major OperatingCosts: Cost of all imports crossing limit A-A
and B-B.
Credits include only items crossinglimit designation C-C.
65
4. ECONOMIC EVALUATION
The retrofit plant as described in Section 3.2 has been evaluated for its(11)economic viability. The economic model ECONOCALC was used to
prepare the evaluation presented in this section.
Table 4.1 provides a complete list of economic assumptions including allmajor raw material and utility consumptions. This information shouldprovide interested readers with the data to complete independentevaluations.
It should be noted that the new corporate Federal Income Tax rules wereused in this evaluation.
The results of the base case evaluation and the various sensitivities areincluded in Tables 4.2 and 4.3. The breakeven analysis calculates arequired value for sulfuric acid which includes all production costs andrepayment of capital without a return on the investment. The requiredvalue for sulfuric acid to support a 15% DCF return is approximately$31/short ton (100% H2SO4 basis).
66
TABLE 4.1ECONOMIC ANALYSIS BASIS;
DMC/FIPR PROCESSCOAL/COGEN CIRCULAR GRATE
Plant Location: Adjacent to an existing phosphoric acid complexwhich inc ludes a sulfuric acid unit withcogeneration capability, Gulf Coast U.S.
Total Installed Cost for New
Facilities, MMS (4th Quarter 1986) 85
Working Capital, % TIC 10
Investment Equity, % 100
Construction Period, Months 30
Federal Income Tax, % 34
Tax Depreciation Life (ACR), Years 3
Plant Life, Years 20
Production, Days/Year
Escalation, %
Sulfuric Acid Production, Tons/Day
Aggregate Production, Tons/Day
All weights in short tons
Sulfuric Acid Units Based on 100% Acid
330
3
2800
2510
Costs in U.S. Dollars
TABLE 4.1 fcont’dl
Cost ItemS:
Raw Materials
Coal (to char) .I1 35 3.85 Pyrites (85% FeS2) -35 26 9.10 Phosphogypsum 1.35 1 1.35 i Binder .lO 4 0.40
Ytilities
Coal (to energy) Cooling Water Pond Water Boiler Feed Water
Operating Maintenance
Personnel
Operating Maintenance Supervision/Administration
Credit Items
Ton/Ton fi2sIQ4
Ton/Ton H2=4
Unit Cost/Ton E2=4
-18 35 6.30 9.10 ;005 .05
18.43 .003 -06 -10 .300 .03
Yd
Aggregate Power to P205 Plant Steam to P205 Plant
H2SO4 t0 P205 Plant
Ton ,
Ton
Ton
Unit
f/Unit
5 .90 50 .13 9 .52
Calculated Value
Cost/Ton E2s0,
Cost/Ton a2se4
-65 3.68
Cost/Ton a2=4
1.95 1.83 .54
Production/ -2%
68
TABLE 4.2 BASE CASE
ECONOMIC EVALUATION RESULTS
After Tax DCF % Return
Breakeven.
5%
10%
15%
H2ssZ, Reauired Valug $/ST
17.15
20.40
25.10
31.00 -’
69
Caoital Cos
$85 MM
Coal Price
!§35/Ton
$26/Tori
wte Valuq
SS/Ton
. 1c Powct
Credit
$5O/Mw
Operating Costs/ Product Values
3%
TABLE 4.3 SENSITIVITIES
+20% 18.00 -20% 16.30
$25/Tori 14.25 ’ %30/Tori 15.75 $40/Tori 18.50
%22/Tori 15.75 $28/Tori 17.85 f32/Ton 19.25 $36/Tori 20.65
SO/Ton 21.65 $3/Tori 18.90 $8/Tori 14.50 $1 O/Ton 12.70
$4O/Mw
$6O/MW
0%
5%
Reauired Value of Y2Q for Breakeven
$/ST
18.45
15.85
18.40
16.40
70
5. OTHER CONSIDERATIONS
Intermediate Cogeneration
Recent site specific evaluations have demonstrated the importance ofdetailed energy surveys when considering retrofitting the circular grategypsum process. One approach, which can provide an intermediate stepto replace energy lost by not burning sulfur, is the installation of a gasturbine system based on natural gas as a fuel. This approach wouldresult in lower initial capital costs than the complete coal/cogenerationroute but this must be evaluated based on natural gas and petroleumcoke price tradeoffs.
It should be noted that the appropriate gas turbine can, with minimummodifications, be converted at a later date for use on coal-based fuelgas.
A blockflow diagram, Figure 5.1, of this “intermediate” cogeneration isincluded in this section. Using the same economic basis as for thecoal/cogeneration and assuming $4.00/MM BTU, it is projected that abreakeven required value of $22.50/ton would be required for sulfuricacid for the natural gas/cogeneration system.
A natural gas price of $2.00/MM BTU, which is currently available insome Gulf Coast states, would result in a $17.00/ton sulfuric acidrequired value. The relative prices and projected prices of energysources, i.e., natural gas versus the various solid carbon sources, must beevaluated prior to the selection of the most cost effective cogentrationroute.
Coal Waste Utilization
One extremely attractive alternate source of raw material for thecircular grate phosphogypsum to sulfuric process would be coal wastes.Initial surveys indicate that large amounts of high sulfur/pyritic coalwastes are available and their utilization would help alleviate a growingenvironmental concern in the coal industry. The price for such materialcan generally be considered the loading and transportation costs only.The inherent flexibility of the fluid-bed coal devolatization unit wouldallow the use of a wide range of coal wastes at no increase in capitalcost. Depending on the amount and form of the sulfur contained in thecoal waste it is conceivable that the need for a separate source of pyritescould be substantially reduced or even eliminated. One source which weplan to evaluate contains 9% sulfur and a heat content of about 8000BTU/LB. Using the same economic basis as the coal/cogeneration basecase in Section 4 but assuming a price of $22/ton for this coal waste anda 40% reduction in pyrites requirements, the breakeven value of sulfuricacid would be $5/ST lower than the base case or about $12/ST. Ofcourse each raw material source will have to be evaluated on anindividual basis since parameters such as heating value and moisture canhave significant impact on the overall economics.
71
6. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Hull, W.Q., Schon, F. and Zirngibl, H.: “Sulfuric acid fromanhydride”. Industrial and Engineering Chemistry 49, (8) (August1957) p. 1204.
Scurr, W.R.: “The use of coal for recycling phosphogypsum, theexperience of Sentrachem Limited”. Raw Materials in SouthAfrica, International Fertilizer Industry Association LimitedMeeting, Johannesburg, South Africa (March 1983).
Gosch, H.W.: “Production of sulphuric acid and Portland cementfrom phosphogypsum”. Raw Mate r i a l s i n Sou th Af r i ca ,International Fertilizer Industry Association Limited Meeting,Johannesburg, South Africa (March 1983).
Gardner, S.A.: “The conversion of phosphogypsum into usefulproducts utilizing the Davy McKee circular grate processingsystem”. Raw Materials in South Africa, International FertilizerIndustry Association Limited Meeting, Johannesburg, SouthAfrica (March 1983).
Saylak, D., Unger, E.L.: “The use of phosphogypsum as a roadbasematerial”. The Fertilizer Institute Environmental Symposium(October 1984).
Naff, D.B.: “The c lass i f ica t ion of phosphogypsum forenvironmental purposes”. The Fertilizer Institute EnvironmentalSymposium (October 1984).
Nolan, P.D., Vicard, J.F.: “Advances in gas cleaning technology”.International Sulfide Smelting Symposium; the MetallurgicalSociety of AIME, San Francisco, California (November 1983).
Kendron, T.J., Marten, J.H., Lloyd, GM.: “Phosphogypsum - aproblem becomes an opportunity”. American Institute ofChemical Engineers, Spring National Meeting, New Orleans,Louisiana (April 1986).
Simbeck, D.R., Dickenson, R.L.: “Integrated gasification combinedcycle for acid rain control”. Chemical Engineering Progress(October 1986).
Hardison, L.C.: “Chelated iron chemistry for H S oxidation comesof age”. American Institute of Chemical E ngineers, CentralFlorida Section 1986 Joint Meeting, Clearwater, Florida (May1986).
Mills, J., Atcheson, M.: “ECONOCALC 1, project and ventureeconomics and analysis”. Gulf Publishing Company, Copyright1985.
73
7. AUTHORS
T. J. Kendron (Tim) - DMC - Technical Manager, 4715 South FloridaAvenue, Lakeland, Florida 33807-5000, (8 13) 646-7 172.
Mr. Kendron, (B.S.M.E. 1969 Illinois Institute of Technology, Chicago,Illinois) has been involved with process development, design engineering,procurement, construction and operation of various petrochemical, coalconversion and fertilizer plants since receiving his degree in mechanicalengineering. His work at Davy McKee has included direct involvementin the engineering of large-scale methanol facilities, phosphoric acidcomplexes, and coal utilization processes. Mr. Kendron has participatedin the advancement of chemical processes from bench-scale testingthrough successful commercial demonstration.
G. M. Llovd (Mike) - FIPR - Research Director-Chemical Processing,Florida Institute of Phosphate Research, 1855 West Main Street, Bartow,Florida 33830, (813) 533-0983.
Prior to FIPR, Mike worked at Agrico Chemical Company and heldpositions in research, process engineering and production of fertilizers,phosphorus, and both wet and furnace phosphoric acid. He received aBChE degree from Clemson University in 1950.
The authors would like to acknowledge their respective organizations, DavyMcKee Corporation and FIPR; the Cleveland Research Laboratories, Inc. andthe Florida Department of Education for significant contributions to thecontinuing development of this process. We would also like to take thisopportunity to acknowledge the valuable contributions to this program of Dr.D. Borris, Executive Director of FIPR and J. H. Marten, Vice President ofTechnology of DMC.
74
AGGREGATE FACT SHEET
DESCRIPTION:
The solid by-product produced on the circular grate during the desulfurization ofphosphogypsum has been evaluated as a road construction material. The solid isproduced on the grate in large sintered slabs, reduced in size to 6" x 0, cooled andfurther crushed depending on ultimate use. A portion of the fine materialproduced is recycled for internal use. The larger solids can be sized to meet ASTMStandard C136 - sieve analysis of fine and coarse aggregates.
COMPOSITION:
The solid by-product is a complex matrix of various inert minerals. FIPR hascompleted x-ray diffraction analysis and reported the following:
Maior Phases (>20%)
No sulfides were detected. Sulfur ranges from 1-3 wt% and carbon ranges from .5to 1 wt%.
75
MECHANICAL TESTING:
Specific Gravity: 2.9 - 3.1
Absorbtion of Water; 50% less than higher grade lime-stone
Abrasion Resistance ASTM C 131-81-Resistance to Degrad-ation of Small-Size Coarse Aggregateby Abrasion and Impact in LosAngeles Machine
Class A (1-1/2” x l/4”) 38% passingfinal screeningClass C (3/8” x l/4”) 31% passingfinal screening45% or less passing final screeningis acceptable aggregate
Aggregate Mixes
CHEMICAL TESTING:
Two Asphaltic concrete design mixesperformed; D.O.T. Type S-1 andIII. These two mixes encompassmajority of the mixes, includingS-1,S-3, Type II ABC-l, ABC-2, ABC-3,SAHM and have the most stringentrequirements. These mixes metor exceeded all the criteriaestablished by the Florida D.O.T.in their current specifications.
An EP Toxicity test has been completed to evaluate the leachability of the eightprimary heavy metals (As, Ba, Cd, Cr. Pb, Hg, Se, Ag). All metals reporting to theleachate are substantially less than the standards set for this test which arecorrelated to EPA drinking water standards.
KINETICS OF REDUCTION OF ANHYDROUS
CALCIUM SULPHATE TO CALCIUM SULPHIDE,
WITH CARBON MONOXIDE, WITHIN THE
CHEMICAL CONTROL REGION
M.Alice Cabral de Goes*
Roberto C. Villas Boas**
INTRODUCTION
The rising prices of sulphur and the logistical problems
associated with its supply have arised growing concers regarding
alternative sources of sulfur.
The raw materials that have been receiving attention are anhydrate,
natural gypsite and phosphogypsum, besides pyrites and gases from non-
ferrous metals industries.
The calcium sulphates represent substantial sources of sulphur.
However their industrial process feasibility depends on a given process
design, on the market demand for the resulting by-products, as well on
the capital and operating cost of the industrial plants.
The fundamental step of several existing process schemes to
recover sulphur from anhydrate, gypsite and phosphogypsum is the
reduction of calcium sulphate to its sulphide.
OBJECTIVE
The determination of the intrinsic rates of reduction of calcium
sulphate into calcium sulphide using carbon monoxide gas as the reducing
agent, at temperatures ranging from 700°C to 900°C, within the chemical
control region.
* Met.Eng., M.Sc., Research Engineer CETEM, Rio de Janeiro, Brasil
**Min.Eng., M.Sc., D.Sc., Assoc.Professor UFRJ, Diretor PAA-Engenharia,
Sao Paulo, Brasil
77
EXPERIMENTAL
The reduction experiments were carried out in a "Stanton Redcroft,
MF-H1" thermobalance under isothermal conditions, being the reduction
kinetics determined through weight loss measurements.
Anhydrous calcium sulphate was obtained roasting phosphogypsum
(Serrana Mineracao S.A.) at 7OO°C. After roasting it was screened to
different size fractions, being the -100 +150 mesh fraction chosen for
this study.
The CaSO4 crystals, with a shape factor Fg=1 were carefully loaded
in the thermal balance Pt-basket, as to form a thin layer.
After the temperature has been stabilized, the nitrogen flow was
discontinued and immediately thereafter a flow (28.102 ± 0,5 l02 cm3.
min-1(298K; 1,01.105 Pa)) of purified carbon monoxide gas was introduced
into the reaction tube. At the end of the experiments the furnace was
turned off and the reduced specimen was allowed to cool under a nitrogen
atmosphere.
RESULTS
For each experiment, a fraction of reacted anhydrous calcium
sulphate"X"was ploted as function of the reduction time"t",as shown in
figures 1, 2, 3 and 4.
The curves for the interval 0,2 < x < 0,8 show a constant slope
which is the reaction rate, as:
Where, ΓCO = rate of reaction for CO (mol,min-l), = slope of the
"X versus t" (min-1), X = , = weight loss at a given reduction
time t (mg), = maximum weight loss corresponding to complete
reduction of the sample.
78
From these curves the experimental determination of the reduction
reaction rates were thus obtained.
MATHEMATICAL ANALYSIS
The model utilized is based in the work of John, Evans and
Szekely, which may be regarded as a generalization of the numerous
models that have been proposed to represent the diffuse reaction zone of
reacting porous solids. Henceforth we shall refer to the model as the
"grain model".
Let's consider a porous pellet of volume Vp and of superficial
area Ap, made up of individual grains each of which has a volume and
surface area of Vg and Ag, respectively. These pellets and grains may
take the shape of a sphere, cylinder or a slab and are then assigned
shape factors, Fp for the pellets and Fg for the grains, of 3, 2 and 1,
respectively.
For an isothermal system, irreversible reaction and of first order
79
with respect to gaseous reactant (in our case n = 1,15 + 0,10),
neglecting structural changes, assuming that diffusion through the
product layer of the individual grains and the resistence due to
external mass transport are not rate-limiting:
The dimensionless representation of the governing equations shows
that the concentration driving force for reaction and the local
extent of reaction of the solid reactant are related to the position
h and time t* through a single parameter s . This parameter s
incorporates both kinetic and structural properties and is the measure
of the relative magnitude of chemical reaction and diffusion rates.
As s approaches zero, pore diffusion presents a negligible
resistance to progress of reaction. Under this condition the reactant
concentration is uniform throughout the pellet and is equal to that in
the bulk ( = 1). Therefore, is independent of . Thus, for
Fg = 1 (as observed experimentally)
80
Figure 5 shows the experimental (X,t*), as well the theoretical
(X,t) according to values of and geometric forms of pellets and
grains (Evans). For the reaction is chemically controlled,
and thus those experiments where s < 3 were considered to obtain the
parameters for the Arhenius relation:
with a correlation coefficient equals 0.94.
BIBLIOGRAPHY
Cabral de Goes, M.A. - M.Sc. Thesis, COPPE/UFRJ, October 1984,
Rio de Janeiro, Brasil.
81
^
THE M. W. KELLOGG COMPANY
KEL-S PROCESS
A. G. SLIGER
TECHNOLOGY MANAGER
THE M. W. KELLOGG COMPANY
8 3
ABSTRACT
The large quantity of gypsum being produced both by phosphoric acid plants andby flue gas desulfurization units is creating a waste disposal problem. Withthe objective of solving the waste disposal problem while at the same timeproducing useful products, Kellogg undertook a program to develop a process thatwould accomplish the objective. As a result of this development work, the Kel-SProcess was conceived and shown to be technically feasible and with attractiveeconomics possible. The end productsfinely divided calcium carbonate.
are elemental sulfur and a high purity,The key to the Kel-S Process is a reduction
step using some form of solid carbon such as coal or coke. Kellogg has enoughinformation available to design a demonstration plant and would be interested inpursuing a joint development with a partner for whom the process could solve agypsum disposal problem.
84
THE M. W. KELLOGG COMPANY KEL-S PROCESS
BACKGROUND
In the late 1960's, projections by Kellogg and others indicated that flue gasdesulfurization (FGD) systems would be installed on a large scale starting inthe 1970’s and continuing for many years. Since the majority of the projectedFGD units were expected to be of the "throwaway" type, thus generating millionsof tons per year of calcium sulfite/sulfate sludge, a waste disposal problem wasanticipated. In addition, calcium sulfate from existing and planned phosphoricacid plants would add millions of more tons per year to the disposal problem.In fact, the phosphogypsum quantity was expected to exceed the FGD sludgeproduction by several-fold.
With the expectation that the vast amounts of waste gypsum being generated wouldcreate the need for an efficient disposal process, Kellogg undertook a programto develop such a process. Since Kellogg also was developing a throwaway-typeFGD process at the same time, the initial thrust was to develop a system thatwould complement the scrubbing unit, thus providing a complete package forutility application. The objective of Kellogg's development work in this areawas to develop a process that would convert calcium sulfite/sulfate into sulfurand calcium carbonate which could be recycled back to the FGD scrubber. Afurther objective was to develop a system that used only coal, or other forms ofcarbon, as the reductant. The result of this development work was the Kel-Sprocess.
BASIC PROCESS CHEMISTRY
Kel-S is based on concept of reducing the sulfites/sulfates to sulfide,converting the sulfide to the water soluble hydrosulfide, separating thedissolved hydrosulfide from insoluble impurities, followed by carbonation of thehydrosulfide to produce ultra-pure calcium carbonate and gaseous hydrogensulfide. The calcium carbonate can be recycled to the FGD unit or recovered forother end uses while the hydrogen sulfide is sent to a standard Claus plant forsulfur recovery.
Basic chemical reactions involved are:
85
PROCESS DESCRIPTION
A block flow diagram shownthe Kel-S system. Severalone individual application
in Figure 1 illustrates the basic process scheme ofvariations are available, the final selection for anydepending largely on specific site conditions.
Referring to the block flow diagram, calcium sulfite/sulfate is mixed with coal(or some other form of carbon) and formed into pellets via standard techniques.The green pellets are dried and then enter a rotary kiln which is fired withsupplemental fuel, preferably coal. As the pellets pass through the kiln,conditions are maintained to reduce the sulfite/sulfate to sulfide. The hotpellets exit the kiln and after optional heat recovery, are slurried with waterand fed into a dissolving tower where contact with H2S converts the insolubleCaS into soluble Ca(HS)2 which dissolves in the water. Effluent from thedissolving vessel is filtered to reject inerts, unreacted material and any othersolid contaminants that were in the original feed, and a clear solutioncontaining dissolved sulfur values is produced.
The clear filtrate from the dissolving step is sent to a carbonation tower whereit is contacted with CO2 and produces CaCO3 and H2S according to equation 3).Calcium carbonate precipitates as a finely divided, very reactive material whichcan be recycled to the FGD system to absorb SO2. One-half of the H2S producedis sent to the dissolving vessel described earlier and the other half,representing the net sulfur from the original feed, is sent to a standard sulfurrecovery unit, such as a Claus plant, for conversion to elemental sulfur.
DEVELOPMENT RESULTS
BENCH SCALE
The key step in the Kel-S system is1). During the early development phase, a number of methods were evaluated
reduction to sulfide as shown in equation
experimentally with varying degrees of success. It was found that the bestmethod was to form pellets comprising a mixture of coal (carbon) and gypsum andthen heating the pellets to high temperatures which caused the reductionreaction to take place. Critical variables were found to be the coal/gypsummixture, temperature and reaction time. Also important was the method offorming the pellets since sufficient strength had to be developed for processingin the rotary kiln chosen as the reactor.
The bench scale setup used for Kellogg's test work is shown in Figure 2. Theprocess variable which is probably the most important is the amount of coalneeded to reduce the gypsum to sulfide. If too much is used, the cost isincreased and the final product will contain excess carbon. If not enough coalis added to the pellet, complete reduction will not be achieved. Ideally, justenough carbon will be added to achieve complete reduction with all carbon beingconsumed. After considerable experimental work with various combinations, amixture was developed which gave essentially complete conversion of sulfate andnearly 100% utilization of coal. Results of these tests are shown in Figure 3.As is evident from this figure, conversion of gypsum increases rapidly withincreasing coal concentration and then levels off at nearly complete reductionwhen sufficient coal is present. Conversion of coal is nearly complete untilall of the gypsum is reduced and then starts decreasing as more coal is added.The preferred operating point obviously is the one which achieves completeconversion of both coal and gypsum.
FIGURE 2
FIGURE 3
From previous Kellogg work and literature information, a reaction temperaturerange of 1500°F to 1700°F was knowncarbon. A temperature of 1700°F was
to be optimum for reduction of gypsum withselected as the base and the effects of
pellet heat-up times and run times were evaluated.CaSO4 left in the pellets as a function
Figure 4 shows the amount of
two different rates.of contact time for pellets heated at
For rapidly heated pellets -- 2-minute heat-up -- thereduction was less than for pellets which took 35 minutes to reach reactiontemperature.in Figure 5.
The average temperature profile for the 35-minute heat-up is shownA possible reason for the difference is that at about 1400oF, the
volatile matter in the coal is evolved so rapidly that a portion of the carbonin the coal escapes unreacted.
LARGE SCALE TESTS
Reduction
The bench scale test results confirmed earlier results obtained both in thelaboratory and batch kiln tests run for Kellogg by an outside company. It wasnecessary to determine if these resultscontinuously operating equipment using actual
could be achieved in larger,scrubber sludge. Additional work
was done under a joint funding arrangement with EPA.
For the large scale tests, it wassolids would be needed.
estimated that some 30-40 tons of sludgeAt the time (mid-1975), very few FGD scrubbers were
operating that were producing sludge suitable for Kel-S feed; i.e., sludge thatwas free of other material such as stabilizing additives.sludge from Commonwealth Edison's
After some searching,Will County Station scrubber was chosen and
100 tons of waste slurry were obtained.during which
The program was delayed for two years
oxidized.time the sludge, stored in an outside reservoir, had become
Figure 6 shows the change in composition with time.
Eventually, Kennedy Van Saun (KVS) was subcontracted to carry out bothpelletizing and reduction tests at their test center (Danville, Penn.). Firstresults showed that the long storage and slow oxidation had increased the gypsumparticle size such that suitable pellets could not be formed. After testingvarious methods and additives, a proceduredeveloped.
that produced suitable pellets wasThe pellets were tested for strength by two common methods -- drops:
number of 18" drops onto a steel plate until breaking, and compression strength:weight required to break pellets placed between two steel plates. Pellets withwet strength of 5.2 lbs and up to 20+ drops before breaking were produced.These characteristics made the pellets satisfactory for drying and kilnprocessing. Two hundred pound batches of pellets were formed in a 4-ft diametercone pelletizer, stored overnight and thenFigure 7.
dried in KVS's flowdryer system,The dried pellets were belt-fed into the kiln for reduction.
The KVS test kiln used was a 24-in.kiln, shown schematically in Figure 8.
diameter by 30-ft long, oil-fired rotaryCommercially, a coal-fired kiln would be
used. The test objective was to define operating conditions that would produceessentially complete reduction and 100% utilization of the coal. Batches ofpellets suitable for a 2-day kiln test were made, dried and stored in drums.Nominal feed rates of 130-180 lb/hr were used and the hot end temperature was tobe maintained at 1750°F. Samples were taken from three locations -- at 10-ft
93
W Iv
@ RUN 1 H RUN 2 V RUN 3 Ir, RUN 4
I I I I
10 20 30 40
(REACTION) CONTACT TIME (MINI
KEL-S SLUDGE REDUCTION TEMPERATURE DURING REACTION
FIGURE 5
LEGEND
l CaC03
m CaS03 - %H20
A CaS04 - 2H20
601
50
z
40
e 2 g 30
20
10
t I I I 1 I I I I I I I
0 2 4 6 8 10 12 14 16 18 20 22
MONTHS
KEL-S SCRUBBER SLUDGE, EFFECT OF STORAGE
FUEL
VENT
SILO
COLD
1 TC = THERMOCOUPLE S = SAMPLE PORT
ID FAN #2
ID FAN #l
PREHEATER / HOT CYCLONE
f- BYPASS
-=+=il FIRING ZONE / DRIVE
ROTARY KILN TRAIN
FEED/ CHUTE
HOT CYCLONE
FEED
PRIMARY FAN SPI LLAiiE
KVS 249INCH OD B-+ 3&FOQT ROTARY PILOT KILN
FIGURE 8
from the discharge, 16-ft from the discharge and the kiln discharge itself.Kiln residence time was about 1-1/2 hours, which appeared to be excessive sincesome reoxidation was taking place.
A typical temperature profile for the kiln tests is shown in Figure 9. Theaverage reduction for one test run was 97.2% and the average residual carbon inthe kiln discharge was 0.02%. A number of individual samples in some runsachieved reduction of greater than 99%. It was concluded, therefore, that ifenough time had been available to optimize the operation, complete reductionwith essentially no residual carbon could have been achieved.
Carbonation
The following description is for a "one-step" Kel-S process wherein the finalproduct is a mixture of CaCO3, ash, unreacted carbon, and other impurities.
Reduced material (i.e., CaS) from the large scale tests was shipped to Kellogg'sR&D Center, Houston, and used as feed for carbonation tests. The tests wereconducted using batches of reduced material but continuous flow of reaction gas,i.e., CO2. A schematic of the test apparatus is shown in Figure 10. Theprocedure was to add a slurry of CaS and H2O, heat to reaction temperature underhelium flow, then switch to CO2. Samples of off-gas were taken every 3-minutesand analyzed by gas chromatography. At the end of the run, which was consideredto be when off-gas H2S concentration fell below 0.5%, the contents were cooled,discharged, and analyzed.
When efficient contacting was achieved between gas and slurry, the reaction wasvery rapid and essentially complete; i.e., greater than 99.9% conversion of CaS.The H2S in the offf-gas approached 100% when CO2 only was fed into the reactor.Figure 11 shows typical results. Note that there is an initial period in whichno CO2 or H2S was detected and no pressure change occurred. This delay is theresult of three separate factors:
0 Equipment related -- time required to fill up the volume of unoccupiedspace in the reactor and lines.
Typical results are shown in Figure 12. Note that the rate of H2S evolution isgreater than the CO2 feedequation 6). The overall reaction, however, is the same as the stoichiometry:
rate during part of the run which is explained by
Also indicated in Figure 12 isof CO2; i.e.,
the time corresponding to stoichiometry additionthe quantity of CO2 required to react with CaO and CaS. It was
determined from analysis of the test results that the limiting factor was therate at which CO2 could be fed into the reactor.unit operating under
A properly designed full-size
limitation.steady state conditions would not be subject to this
96
TEMPERATURE PROFILES AND RATE OF PRODUCT DISCHARGE, MIXTURE K3
0 He f
KEL-S CARBONATION, BENCH SCALE UNIT, FINAL DESIGN
A "two-step" Kel-S process produces CaC03 as a pure component and the remainingmaterial as a separate stream by separating the dissolution and carbonationreactions. Test runs were made to demonstrate the feasibility of the "two-step"system using the same equipment described earlier.
First, dissolution is achieved by feeding H2S gas into the reactor equal to thestoichiometric amount needed to convert the CaS to Ca(HS)2. A typical run isshown in Figure 13. The remaining solids were filtered and analysis showed noCaS present. The clear liquid was then contacted with CO2 using the sameprocedure described earlier. Results are shown in Figure 14 and are essentiallyidentical to the one-step carbonation of CaS. The precipitated calciumcarbonate was analyzed and found to contain 99.37% CaCO3.
Solids Separation
Denver Equipment ran air flotation tests on the material produced in the one-step carbonatioh runs to determine if CaCO3 could be separated from ash,unreacted carbon and other impurities. Results show that air flotation was notvery efficient in that high losses, of calcium were suffered in order to rejecthigh percentages of non-calcium compounds. Optimization probably would improveresults but the probability of obtaining a high purity CaCO3 at a high recoverylevel appears marginal.
The material produced by the dissolution reaction was tested by Bird Machine Co.to determine the feasibility of vacuum filtering and centrifuging for solidsseparation. Both solid bowl and perforated bowl tests were run and showedacceptable performance. Vacuum -filter tests were also carried out and likewisegave acceptable results. For the two-step process, it was concluded that eithercentrifuging or filtering would be acceptable for a commercial unit.
ECONOMICS
Based on the experimental data obtained, a number of flowsheets and economicevaluations have been developed. One of these was for disposal ofphosophogypsum in response to ansulfuric acid as end products.
inquiry for production of cement clinker andA comparison was made in 1981 to the OSW process
which is a commercial process for producing sulfuric acid and cement clinkerfrom either calcium sulfate anhydrite or phosophogypsum. The OSW processrequires that phosophogypsum be treated to remove phosphorus and fluorinecompounds to prevent clinker. contamination. Kel-S should remove some of thephosphorus and most of the fluorides priornot yet been verified by actual data.
to the clinkering step but this has
For a plant processing 7600 TPD gypsum feed (dry basis), the Kel-S investmentwas estimated to be 30% lower than the OSW investment.based on a complete plant in each case,
These comparisons are
clinker plants and a tail gasincluding Claus, sulfuric acid and
scrubbing unit. Operating requirements for thetwo systems are shown in Table I while a comparison of operating costs is shownin Tables II and III. In both cases the credits exceed the cost, excludinggypsum cost and capital charges.
At the time of this study (1981), the client reported that Kel-S costs lookedattractive but before any further action was taken, environmental requirementswere relaxed which eliminated the need for gypsum disposal.
101
104
105
106
CONCLUSIONS
Experimental work has shown that the Kel-S process is technically feasible andpreliminary cost analyses show that attractive economics are possible. Enoughinformation is available to design a demonstration plant using FGD sludge as thefeed. Some additional work is needed to confirm the disposition of fluorine,uranium and residual phosphate if phosophogypsum is used as the feed. Webelieve we have reasonable routes to handle these impurities but experimentalconfirmation should be made prior to building a demonstration plant.
As an engineering/construction company, Kellogg does not have a plant forinstalling a demonstration Kel-S unit.joint development of Kel-S with
Kellogg would be interested in pursuing
gypsum disposal or re-use needs.a partner for whom the process could solve
107
PROCESSING OF PHOSPHOGYPSUM TO SULPHURIC ACID
Iv. Gruncharov, Y. Pelovski, Iv. Dombalov, Pl. Kirilov,
N. Videnov
Higher Institute of Chemical Technology, Sofia 1156,
Bulgaria
The thermochemical decomposition of phosphogypsum (PHG)
to lime and sulphur dioxide permits the completion of the
technological cycle in the production of phosphoric and sul-
phuric acid and the development of a complex waste free tech-
nology. On this problem are dedicated various investigations
and patents but as yet none was applied in industry [1-4].
The essential variation in the experimental conditions and
in the structure and the chemical composition of PHG do not
allow the direct comparison of the obtained results and the
derivation of generalized criteria of the economic expedience
of implanting this
In the present
during the study of
process.
paper some of the final results obtained
the process of the thermochemical PHG de-
composition in a reducing gaseous medium, prepared by the pro-
cessing of North African phosphorites to phosphoric acid by
the dihydrate method are presented.
The thermodynamic equilibrium in the system CaSO4-CO-H2
defines the basic interrelation between the composition of
the gaseous and the solid phase. In the PHG decomposition this
equilibrium is only approximated because of the impurities
influence. The presence of specific admixtures in PHG deter-
109
mines the possibility to take place of a multitude of side
reactions concerned with the phase and structural composition
of the solid product. Some of the admixtures are able to cata-
lyze or to inhibit the process.
Our investigations on PHG from North African phosphorites
containing 38,6% CaO, 54,7% SO3, 0,7% P2O5, 0,2% F, 1,5% SiO2,
0,7% Fe2O3, 0,4% Al2O3, 0,2% MgO etc. in various gaseous media
applying thermal methods (TG-DTA, high temperature microscopy),
gas-chromatography, X-ray spectroscopy, SEM etc. [4-6] demon-
strated that at a definite gas phase composition the PHG could
be decomposed throughly to CaO or to CaO with an insignificant
content of CaS at temperatures 1O5O-11OO°C for 8 to 20 min.
The presence of CaS in the solid product depends on the redox
potential and the temperature of the process. The direct mea-
surement of the partial oxygen pressure by a solid electrolyte.
system [7] of gaseous mixtures containing l-4% HZ, l-4% CO,
5-20% H2O and S-20%, CO2 shows that in the temperature range
1OOO-11OO°C it varies from 3,38.10-12 to 5,34.10-16 atm. with
a limiting effect of the gaseous components ratio. Established
were the limits of the gaseous components ratios and respecti-
vely the values of the partial oxygen pressure at which CaS
is missing or is present in an insignificant quantity in the
solid product at various temperatures.
The kinetics of the process depends on the gaseous phase
composition, the temperature and the size of the PHG partic-
les. At temperatures less than 1050-1075°C and particles'
110
size less than 0,1 mm the effect of mass transfer is insigni-
ficant. When the temperature is increasedthis brings to struc-
tural and specific surface variation of the solid phase accom-
panied by an increased mass transport effect. In the decompo-
sition of PHG utilizing hydrogen (PH = 0,01 atm) as a reduc-2
ing agent wide ranges of the experimental data are described
by the Polany-Wigner equation. The activation energy of the
process is 213,3 kJ/mol. The thermochemical decomposition of
PHG by CO (PCO = 0,01-0,04 atm) is accompanied by a conside-
rable induction period /Fig. 1/. In this case of the decompo-
sition of PHC the activation energy is about 2 times lower.
The various solid products obtained from the partially
or completely decomposed PHG were studied. The specific sur-
face of the thermally treated PHG at the temperatures up to
1100°C in various gaseous media dereved from the gas system
H2- CO - CO2 - H2O - Ar has a pronounced extremal value at
350-500°C. This temperature range coresponds to the transition
of the soluble anhydrate form of the calcium sulphate into an
insoluble form. The specific surface of the solid products
decreases rapidly reaching 1-2 m2/g by the temperature rise
up to 800-900°C. After that up to 1100°C the specific surface
is almost invariable.
The chemical and X-ray difraction analysis of the solid
products confirm that the ratio of calcium oxide and calcium
sulphide is determined by the value of the redox potential
i.e. by the content of Hydrogen and Carbon monoxide as well111
Fig. 2 SEM photographs on the solid phase from PHG decompo-sition ( = 0,25) in gas atmosphere: 2.1 1%H2+99% Ar
2.2 4% CO+96% Ar
112
as the ratios H2/H2O, CO/CO2, CO2/H2 and H2O/CO.
By the SEM investigations on the solid products it was
found out that if the reducing agent is CO the new calcium
oxide phase is initiated on definite spots of the PHG partic-
les only. When the gaseous medium contens Hydrogen the appea-
rance of the calcium oxide is spontaneously on the total sur-
face of PHG /Fig. 2/.
Introducing additional quantities of the admixtures in
PHG usually brings to a decreased thermal stability and struc-
tural and compositional solid phase variation. As a result
some of the admixtures accelerate the process in the lower
temperature range and decrease the temperature limit of diffu-
sion enabling to carry out the process. The effect of some
other additives (carbonates, chlorides, sulphates, etc.) on
the kinetics of the process was studied and it was establi-
shed that some of them considerably intensify the process of
the thermochemical decomposition of PHG as the optimal tempe-
ratures decrease with 25-50°C.
The thermochemical decomposition of PHG was simulated on
a laboratory unit in a "Fluidized bed" reactor with a 60 mm
diameter at temperatures 1000-1200°C. At temperature of
1100°C the obtained degree of decomposition for 15 min. was
98,6%. On the ground of the results obtained and the determi-
ned optimal conditions semi-industrial stuides were carried
out in a reactor with a "fluidized bed" with 0,9 m diameter.
As a fluidizing agent the products of the incomplete burning
113
of natural gas with air were used. The PHG was preliminary
subjected to granulation as the size of the granules was
1-5 mm. The expense coefficient of air was 0,82-0,85 and the
gas flow rate was 3-3,2 m/s. Some of the results obtained are
shown in the Table.
Main results from PHG decomposition in a "fluidized bed"
furnace
The investigation carried out in semiindustrial condi-
tions on the total processing of PHG to lime and sulphuric
acid confirm the possibility for the industrial realization of
the technology. The increased content of P2O5 in lime results
from an additional contemination of PHG with superphosphate
during its granulation. Under normal conditions the content
of the active calcium oxide in lime increase to 70-75%.
The technical-economic estimation of the technology
114
/Fig. 3/ carried out shows that when the optimal exenergy
scheme can be realized the production of sulphuric acid by
this method could be economically efficient.
REFERENCES
1. T.D. Wheelok, D.R. Boylan, Ind. Eng. Chem., 52, 1960215,
2. N.L. Solodyankina, V.M. Borisov, V.A. Skorobogatov et al.,
Khim. Prom. 2, 155, 1982
3. I.G. Kostilkov, O.V. Rogachov, Khim. Prom. 5, 292, 1986
4. Iv. Gruncharov, Pl. Kirilov, Y. Pelovski, Iv. Dombalov, N.
Videnov, Nat. Scient. Conf. Mineral Fertilizers, Varna,
11-13.9.1982
5. Iv. Gruncharov, Pl. Kirilov, Y. Pelovski, Iv. Dombalov,
God. VHTI, Sofia, 29, 14, 1983
6. Iv. Gruncharov, Pl. Kirilov, Y. Pelovski, Iv. Dombalov,
Proc. of the 8th Int. Conf. on Therm. Anal. ICTA'85, Bra-
tislava, vol. 1, 173, 1985
7. Y. Pelovski, Iv. Gruncharov, Pl. Kirilov, Iv. Dombalov,
Compt. rend. Acad. bulg. Sci. vol. 39, No 10, 75, 1986.
115
-T--Q’
I ‘D
I
p-1 .n
116
RECOVERY OF SULFUR FROM PHOSPHOGYPSUMPART 1. - CONVERSION OF CALCIUM SULFATE TO CALCIUM SULFIDE
BYMargaret M. Ragin, Physical Scientist
andDonald R. Brooks, Metallurgist
U.S. Department of the InteriorBureau of Mines
Tuscaloosa Research CenterP.O. Box L
University of Alabama CampusTuscaloosa, AL
117
RECOVERY OF SULFUR FROM PHOSPHOGYPSUMPART 1. - CONVERSION OF CALCIUM SULFATE TO CALCIUM SULFIDE
By Margaret M. Ragin and Donald R. Brooks
ABSTRACT
In a cooperative effort between the U.S. Bureau of Mines and the
Florida Institute of Phosphate Research (FIPR), with input from the
phosphate industry, the conversion of phosphogypsum to sulfur is being
investigated. The first step in the conversion scheme involves the
reduction of phosphogypsum to calcium sulfide. The effects of
temperature, catalyst, reaction time, and type of reductant (carbon
monoxide and coal) are discussed. Results indicate that as the volatiles
content of the coal is increased, the reaction temperature may be
decreased to obtain a given conversion, or that if the reaction
temperature is maintained constant, the conversion can be increased. Best
results were 95- to l00-pct conversion using a high-volatile coal at 800
to 850° C. The addition of catalyst increased the calcium sulfide yield
when low- or medium-volatile coals were used as reductants, but had little
effect when high-volatile coals were used.
118
UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT
119
INTRODUCTION
The phosphate industry in Florida is a vital segment of the Nation's
economy. Phosphate fertilizer manufacturers produce wet-process
phosphoric acid, which is a major ingredient for fertilizers. The
phosphoric acid is made by the reaction of phosphate rock with sulfuric
acid, phosphogypsum being the byproduct. This phosphogypsum waste
material represents virtually all of the sulfur that has ever been used in
the fertilizer manufacturing process and is therefore a potentially
valuable mineral commodity.
Estimates indicate that over 500 million tons of phosphogypsum have
accumulated in Florida stockpiles. In 1980, phosphogypsum was being
generated at a rate of about 33 million tons a year. The projected
accumulation by the year 2000 would be over 1 billion tons (May and
Sweeney, 1982).
Processes have been developed to produce sulfuric acid from gypsum, and
commercial plants have been built in Germany, England, France, Austria,
South Africa, and Poland (Zellars-Williams, Inc., 1981). The processes
involve the reductive roasting of gypsum using coal, natural gas, carbon
monoxide, or hydrogen to produce sulfur dioxide, which is converted to
sulfuric acid. The Bureau has investigated the reductive roasting of
gypsum (Martin, et al., 1963), the conversion of calcium sulfide to
elemental sulfur (George and Riley, 1971), and sulfur dioxide emission
control (Marchant, et al., 1980). The latter process involved the
reaction of sulfur dioxide and hydrogen sulfide to produce elemental
sulfur (Claus Process).
120
Numerous investigations of the environmental aspects and possible uses
of phosphogypsum have been conducted by the U.S. Bureau of Mines, the U.S.
Environmental Protection Agency, the U.S. Geological Survey, the Florida
Institute of Phosphate Research, the phosphate industry, and many
universities (May and Sweeney, 1982; Zellars-Williams, Inc., 1981; PEI
Associates, Inc.; Miller and Sutcliffe, Jr., 1982; and Austin, 1980). In
general, these investigations have characterized the phosphogypsum,
established its environmental impact on air and water, and proposed
several high-volume potential end uses. Some research has focused on
conversion of phosphogypsum directly to sulfuric acid (Zellars-Williams,
Inc., 1982, Kendron
been focused on the
and Lloyd, 1985). Only a minimum amount of effort has
technology of converting the material to sulfur.
The consensus of the Florida phosphate fertilizer manufacturers is that
the conversion of phosphogypsum to elemental sulfur would be very
advantageous. The future need for additional phosphogypsum waste piles
would be eliminated, and the sulfur could be recycled back into fertilizer
production through existing sulfur-burning sulfuric acid plants. This
conceivably could eliminate the need to purchase additional sulfur. In a
cooperative effort with FIPR, the Bureau of Mines is investigating the
conversion of phosphogypsum to elemental sulfur. Part 1 of this
investigation discusses the reduction of phosphogypsum to calcium
sulfide. Part 2 will discuss the conversion of calcium sulfide to sulfur.
MATERIALS
The 1-in tube furnace experiments were conducted with reagent-grade
calcium sulfate (anhydrite) and an air-dried phosphogypsum waste from
central Florida. The chemical composition of the phosphogypsum waste is
121
given in table I, and a screen analysis is shown in table II, Commercial-
grade carbon monoxide and coal were used as reductants. The analyses of
the different coals used in these studies are shown in table III. A
chromium doped iron oxide and a 0.05-µm magnetite (Fe3O4) served
as catalysts.
For the 3-in rotary tube furnace tests, phosphogypsum samples were
obtained from two Florida phosphate producers. Chemical analyses of these
samples, A and B, are given in table IV. The samples were air-dried and
passed through a 20-mesh screen to ensure a uniform particle separation
permitting intimate mixing with the reductant particles. Coal sample 8
was obtained from a mine in Alabama and prepared for testing as the
reductant in the rotary tube furnace tests. The coal was crushed and
pulverized to approximately 70 and 80 pct minus 100 mesh. Analyses of the
phosphogypsum and coal were used to calculate the carbon-to-sulfur mole
ratios.
122
TABLE III. - Analyses of coals used in this study
I Moisture Analvsis. pet
1 Ash 1 Volatiles 1 Fixed ; Btu/lb I I I I carbon I
; 11,819 Sample 1:
As received I 6.04 7.81 73.12 Dried at 105" C 1 NAP 8.31 77.82
Sample 2: I As received I 2.76 18.54 74.38 Dried at 105" C I NAP 19.07 76.49
Sample 3: As received I 1.78 19.31 74.22 Dried at 105" C I NAP 19.66 75.57
Sample 4: As received I 1.31 20.35 70.44 Dried at 105" C 1 NAP 20.62 71.38
Sample 5: As received I 13.73 20.32 57.53 Dried at 105" C I NAP 23.55 66.69
Sample 6: As received I 1.98 23.53 67.42 Dried at 105" C I NAP 24.01 68.78
Sample 7: As received ! 2.71 33.21 61.03 Dried at 105" C I NAP 34.13 62.73
Sample 8: As received I 2.32 38.59 57.00 Dried at 105" C I NAP I 2.14 I 39.50 ! 58.36 1 14.890
i I 1 13.03 1 1 13.87 I
I 1 4.32 ; 1 4.44 1
; 4.69 ; 1 4.77 1
I 1 7.90 I 1 8.00 1
; 8.42 ; 1 9.76 1
I 7.07 I 1 7.21 1
; 3.05 I 1 3.14 1
1 12,579
; 14,605 1 15,020
; 14,730 1 14,997
I 1 14,220 1 14,409
i 12,076 1 13,998
1 14,191 1 14,478 I 1 14,212 1 14,608
1 14,545
TABLE IV. - Chemical analysis of air dried phosphogypsum waste, percent
I Sample A I Sample B CaS04 ............. 1 61.1 1 51.4 H20 ............... I 16.6 14.3 Si02 .............. I 18.9 I 27.9 CaO ............... 1 -8 I 1.9 Other ............. I 2.6 ! 4.5
Total ........... I 100.0 ! 100.0
EQUIPMENT AND PROCEDURES
Initial investigations were conducted in a single-zone, hinged-type
tube furnace fitted with a 1-in-diam by 30-in-long mullite tube. Sample
preparation consisted of weighing the appropriate amounts of each
123
constituent into plastic containers and shaking the combined materials
until they were thoroughly mixed. One- to seven-gram samples of the
mixtures were weighed into ceramic boats and heated in the tube furnace.
The system was purged with nitrogen prior to heating, and measured amounts
of nitrogen were flowed through the tube during the initial heating,
reduction, and cooling steps of the experiment. The residues were allowed
to cool to 100° C or below in the furnace and then to room temperature in
a desiccator before weighing. The chemical analyses and weights of the
residues were used to calculate the yields of calcium sulfide.
Continuous testing was conducted in a rotary tube furnace, which
consists of a 3-in-diam by 6-ft-long mullite tube horizontally mounted in
a rotation apparatus which supports the tube through the 36-in heated zone
of the furnace. The arrangement, which is illustrated in figure 1,
includes a screw feeder and a water-cooled discharge chamber for.
collecting the reaction product. The unit can be tilted to various angles
of inclination and rotated at different speeds, which allows the retention
time to be varied. Phosphogypsum and coal are weighed in amounts
equivalent to the desired carbon-to-sulfur mole ratio and placed in a
baffled rotary mixer for 10 min prior to being pressed into briquettes.
The roll briquetter produces pellets approximately 3/4 in long by 3/8 in
wide. These briquettes are placed in the hot zone of the furnace by a
screw feeder while nitrogen flows countercurrently through the furnace
tube. The product is discharged into a water-cooled, nitrogen-filled
collection chamber, which is emptied at intervals into another
nitrogen-filled chamber so that samples can be removed without allowing
air into the system.
124
FIGURE 1.- Phosphogypsum reduction reactor
RESULTS AND DISCUSSION
LABORATORY-SCALE TESTS
Preliminary data were collected from the l-in tube furnace tests to
provide guidelines regarding the time and temperature required to obtain
conversion of calcium sulfate to calcium sulfide. Initial reduction tests
were conducted on anhydrite and phosphogypsum waste using carbon monoxide
as the reductant. It was determined that anhydrite could be converted to
calcium sulfide in 2 h at 850° C using 5 pct CO, but a temperature of
900° C was required to convert the calcium sulfate in phosphogypsum to
calcium sulfide under the same conditions. Experiments using coal
containing 23.55 pct volatiles as reductant indicated that phosphogypsum
could be converted to calcium sulfide in 2 h at 1,000° C (Smelley, et al.,
1985). A series of tests using this coal was conducted to evaluate the
effects of the carbon-to-sulfur mole ratio and the addition of magnetite
catalyst on the reduction of phosphogypsum at 1,000° C. A nitrogen flow
rate of 250 mL/min was used, and the reaction time was 2 h. Results of
these experiments are given in table V.
These results indicate that conversion was increased by the addition of
the catalyst and also by increasing the carbon-to-sulfur mole ratio.
To evaluate the effect of coal volatile content on conversion, a series
of experiments was conducted using coals with volatile contents ranging
from 8.31 to 34.13 pct. A reaction time of 2 h was used with a nitrogen
flow rate of 250 mL/min. The results, shown in table VI, indicate that
conversion varies directly with volatile content between approximately 20
and 34 pct.
Since the high-volatile coal produced much higher conversion at 1,000° C,
a series of tests was performed to determine if this coal would affect
conversion at temperatures below 1,000° C. Results of these experiments,
which were conducted with and without the addition of catalyst, are shown
in table VII. In both of these series, the mole ratio of fixed carbon to
sulfur was 2.2 and a nitrogen flow rate of 5 mL/min was used.
127
These results indicate that reduction of phosphogypsum can be accomplished
at temperatures lower than 1,000° C with yields in the high 90's. The
results also show that the addition of the chromium-doped iron oxide
catalyst had little effect on the conversion of phosphogypsum to calcium
sulfide when the reductant was high-volatile coal.
Varying the nitrogen flow rate from 5 to 250 mL/min and using a
carbon-to-sulfur mole ratio of 2.2 and high-volatile coal, the conversion
increased slightly as the flow rate decreased. The results of experiments
conducted at nitrogen flow rates of 5 and 250 mL/min and at various
temperatures are shown in table VIII.
128
ROTARY TUBE FURNACE TESTS
Experiments were conducted in the 3-in-diam rotary tube furnace to
generate additional process information on a larger scale and establish
ranges of operating conditions. In some of the initial tests the
phosphogypsum and coal particles had a tendency to adhere to the inside of
the furnace tube on contact with heat. It was determined that briquetting
the feed would alleviate this problem. The rotary tube furnace was
operated at various temperatures and retention times using coal sample 8,
which contains 39 pct volatile content. Tests were conducted at feed
rates between 25 and 30 g/min, and the angle of inclination was varied
from 0.4° to 3.1° at tube rotation speeds of 4 and 8 r/min. In these
tests the carbon-to-sulfur mole ratio used was 2.2. Results of these
experiments are shown in figure 2 and table IX. Figure 2 illustrates the
relationship between reaction temperature and calcium sulfide yield. For
the lower rotational speed and the lowest two angles of inclination, at
temperatures of 850° and 900° C, yields in the high 90's were obtained.
As the angle of inclination and/or speed of rotation of the tube was
increased, the retention time decreased. The retention times associated
with the yields produced at 900° C are given in table IX. Conversions of
99, 98, and 90 pct were accomplished at 900° C with retention times of 32,
21, and 11 min, respectively.
CONCLUSIONS
The conversion of calcium sulfate in phosphogypsum to calcium sulfide
was accomplished using carbon monoxide as reductant, at a temperature of
900° C in the 1-in tube furnace. The required temperature was reduced to
800° to 850° C using high-volatile coal as the reductant. In the 3-in
rotary tube furnace 98-pct conversion to calcium sulfide was obtained in
approximately 20 min at a temperature of 900° C when high-volatile coal
and a carbon-to-sulfur mole ratio of 2.2 were used. Decreasing the
retention time to approximately 10 min produced 90-pct conversion to
calcium sulfide. Iron oxide catalyst enhanced the production of calcium
sulfide from phosphogypsum when low- or medium-volatile coals were used.
However, when high-volatile coals were used the catalyst had little or no
effect.
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance and advice of Dr. David
P. Borris, Executive Director of FIPR, and Mr. Mike Lloyd, Research
Director of FIPR. The voluntary cooperation of the following Florida
phosphate companies in offering valuable advice and suggestions on the
research efforts is also gratefully acknowledged: Gardinier, Inc., Agrico
Chemical Co., Farmland Industries, Royster Co., W. R. Grace & Co.,
Occidental Chemical Co., Conserv, Inc., International Minerals and
Chemical Corp., and CF Industries.
REFERENCES
1. May, A., and J. W. Sweeney. Assessment of Environmental Impacts
Associated With Phosphogypsum in Florida. BuMines RI 8639, 1982, 19 pp.
131
2. Zellars-Williams, Inc. Evaluation of Potential Commercial
Processes for the Production of Sulfuric Acid From Phosphogypsum. Florida
Institute of Phosphate Research Report 01-002-001, Oct. 1981, 62 pp.
3. Martin, D. A., F. E. Brantley, and D. M. Yergensen. Decomposition
of Gypsum in a Fluidized-Bed Reactor. BuMines RI 6286, 1963, 15 pp.
4. George, D. R., and J. M. Riley. Process for Recovery of Sulfur
From Gypsum. U.S. Pat. 3,591,332, July 6, 1971.
5. Marchant, W. N., S. L. May, W. W. Simpson, J. K. Winter, and
H. R. Beard. Analytical Chemistry of the Citrate Process for Flue Gas
Desulfurization. BuMines IC 8819, 1980, 20 pp.
6. PEI Associates, Inc. Supplement to Region IV's Areawide EIS
Central Florida Phosphate Industry. Ongoing EPA contract 68-03-3197 PN
3617-7; for information contact Mr. Ed Mullin, Project Manager,
Cincinnati, OH.
7. Miller, R. L., and H. Sutcliffe, Jr. Water Quality and
Hydrogeologic Data for Three Phosphate Industry Waste-Disposal Sites in
Central Florida, 1979-1980. U.S. Geol. Surv. WRI 82-045, Apr. 1982,
85 pp.
8. Austin, R. D. Sulfur From Gypsum. Proceedings, International
Symposium on Phosphogypsum, Lake Buena Vista; FL, Nov. 5-7, 1980,
pp. 355-375.
9 . Zellars-Williams, Inc. Evaluation of Fluid Bed Decomposition of
Phosphogypsum. Florida Institute of Phosphate Research Report No.
0l-002A-002, Aug. 1982, 124 pp.
132
10. Kendron, T. J., and G. M. Lloyd. Phosphogypsum, A Source of
Sulfur Dioxide: New Signs of Life? Phosphorus and Potassium, No. 137,
May-June 1985, pp. 33-36.
11. Smelley, A. G., M. M. Ragin, D. R. Brooks, and B. J. Scheiner.
Conversion of Phosphogypsum to Sulfur. Part 1. - Reduction of Calcium
Sulfate to Calcium Sulfide. Proceedings, Third Seminar on Phosphogypsum,
Tampa, FL, Dec. 5-7, 1985.
133
RECOVERY OF SULFUR FROM PHOSPHOGYPSUM.PART 2. - CONVERSION OF CALCIUM SULFIDETO SULFUR
By
Alexander MayResearch Chemist
David A, RiceChemical Engineer
Olice C. Carter, Jr.Environmental Engineer
Tuscaloosa Research Center
For Publication in the Proceedings of the Second InternationalSymposium on Phosphogypsum held in Miami, Florida, December 10-12, 1986
BUREAU OF MINES
U.S. DEPARTMENT OF THE INTERIOR
This report is based upon work done under an agreement between theUniversity of Alabama and the Bureau of Mines.
134
UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT
135
RECOVERY OF SULFUR FROM PHOSPHOGYPSUM
PART 2. - CONVERSION OF CALCIUM SULFIDE TO SULFUR
By Alexander May, 1 David A. Rice, 2 and Olice C. Carter, Jr.3
*** ABSTRACT
As part of a cooperative effort between the U.S. Bureau of Mines and
the Florida Institute of Phosphate Research (FIPR), with input from the
phosphate industry, the conversion of phosphogypsum to sulfur is being
investigated. The proposed process incorporates the thermal reduction
of phosphogypsum to calcium sulfide and a hydrometallurgical treatment
to convert calcium sulfide to sulfur. The research described herein is
focused on the latter half of the process in which the calcium sulfide
is converted to ammonium bisulfide and subsequently oxidized, using air
in the presence of a catalyst, to produce elemental sulfur, which is
adsorbed on the catalyst. The removal of sulfur from the catalyst is a
key area and is vital to the development of a successful sulfur
recovery process. Two approaches for recovering the sulfur are being
investigated, 1) thermal treatment to volatilize sulfur, and 2)
leaching the sulfur from the catalyst with liquid ammonia. This paper
summarizes the results of the ongoing laboratory studies using
activated carbon which was loaded with sulfur through oxidation of
ammonium bisulfide solution. The work has shown that recovery of
sulfur from carbon using vacuum distillation at temperatures below
140° C will be difficult. Ammonia leaching of the sulfur appears to be
technically feasible, with at least 70 pct of the sulfur present on the
carbon being recoverable as an elemental sulfur end product.
1 Research chemist.23Chemical engineer.Environmental engineer.
Tuscaloosa Research Center, Bureau of Mines, U.S. Department of theInterior, Tuscaloosa, AL.
137
*** INTRODUCTION
This investigation is being cofunded by the U.S. Bureau of Mines and
the Florida Institute of Phosphate Research (FIPR) since its purpose
addresses the missions of both agencies. The overall objective of the
program is to develop a process to convert phosphogypsum to elemental
sulfur.
Forecasts of sulfur supply between 1990 and 2000 predict a depletion
in primary sulfur reserves and an increase in sulfur demands and prices
on a worldwide scale (Boyd and Phillips, 1985). The fertilizer
industry in the United States consumes about 65 pct of the Nation's
sulfur (Morse, 1985). This sulfur is used to make sulfuric acid, which
is mixed with phosphate rock to make phosphoric acid, the major
ingredient used to make phosphate fertilizers. The byproduct of the
phosphoric acid production is the calcium sulfate dihydrate,
phosphogypsum. The sulfur, although consumed, is not lost, but is
converted to phosphogypsum, from which it is potentially available. By
the year 2000 the estimated quantity of stockpiled phosphogypsum in
Florida will be about 1 billion tons (May and Sweeney, 1982). Its
available sulfur would help meet the increased demand for sulfur in the
United States.
Research is being conducted by the Bureau of Mines into two process
steps: (1) thermal catalytic reduction of phosphogypsum to calcium
sulfide, the subject of Part 1 (Ragin and Brooks, 1986) of this paper;
and (2) a hydrometallurgical process to convert calcium sulfide to
elemental sulfur. The initial phase of the Bureau's efforts in this
investigation was reported on at the Third Seminar on Phosphogypsum, in
138
1985 (Smelley, et al., 1985 and May and Sweeney, 1985). The
hydrometallurgical approach for conversion was selected because
laboratory experience and the literature (Lefers, et al., 1978 and Chen
and Morris, 1972) indicated that CaS can be oxidized to elemental
sulfur in aqueous systems. It was demonstrated that CaS could be
hydrometallurgically converted to NH4SH. It was also demonstrated
that NH4SH could be converted to elemental sulfur by reaction with
air on a carbon catalyst, in slurry form, followed by distillation of
the NH3 and some of the H2S. Lastly, an improved concept was
developed and laboratory tested. That work demonstrated that the
conversion of NH4SH to sulfur could be accomplished by passing the
NH4SH solution and air through a column containing carbon-coated
mullite beads. The carbon acted both as a catalyst and as an adsorbent
for the elemental sulfur that was formed in situ. The basic reactions
for the overall process are summarized in equations 1, 2, and 3.
The present work is an extension of the previous investigation and
has been primarily directed toward laboratory evaluation of potential
practical methods for recovering elemental sulfur as a product from
carbon catalysts that had been previously loaded with sulfur from
ammonium bisulfide solutions. The two specific techniques that are
being evaluated as to their technical feasibility are 1) thermal
treatment and 2) leaching.
139
Work has been carried out to develop a conceptual flowsheet that uses
liquid anhydrous ammonia as a solvent for leaching sulfur from carbon
and to provide information required for a preliminary economic
evaluation of the process.
Work in the following areas is discussed separately under major topic
headings in this report:
0 Recovery of Sulfur From Carbon by Thermal Treatment
0 Recovery of Sulfur From Carbon by Ammonia Leaching
0 Carbon Recyclability
0 Flowsheet Development
*** RECOVERY OF SULFUR FROM CARBON BY THERMAL TREATMENT
A series of sublimation-distillation tests with pure sulfur or
mixtures of carbon and sulfur was completed. The purpose of these
tests was to evaluate the feasibility of volatilizing elemental sulfur
from carbon after the carbon had been loaded with sulfur through
oxidation of NH4SH.
EXPERIMENTAL PROCEDURE
Glassware was assembled to form a sublimation or closed distillation
system consisting of a reaction flask, side arm, cold trap, and
receiving flask. A laboratory two-stage vacuum pump and ionization
gauge were connected at a point in the system just past the cold trap.
The reaction flask was placed inside a laboratory oven with only the
cold trap and receiving flask outside the heated zone. Icewater or dry
ice-antifreeze mixtures were used to cool the cold trap.
140
Small samples, 1 to 20 g of sulfur-laden carbon and/or sulfur (Fisher
Scientific5 laboratory grade sublimed sulfur), were placed in the
5Reference to trade names does not imply endorsement by the Bureau
reaction flask. The cold trap and vacuum pump were attached, and
heating of the sample under evacuation was initiated. The system was
operated at 120° to 140° C, at pressures of 10 to 30 microns. With
mixtures of carbon and sulfur, samples were extracted with CS2 after
testing to remove sulfur. The extent of sulfur removal was determined
by difference using a head analysis and the post test extraction data.
RESULTS AND DISCUSSION
A series of tests was conducted in which the samples containing
sulfur were slowly heated up to, and beyond, the melting point of
sulfur (114.5° C) in an effort to at first sublime and then distill the
sulfur under the maximal vacuum achievable by the two-stage laboratory
vacuum pump. The results of testing under various conditions of
temperature and pressure are summarized in table I.
TABLE I. - Sublimation-distillation tests
of Mines
Test 1 was conducted using pure flowers of sulfur, and it was shown
that pure sulfur could be sublimed or distilled completely in the
system. Only small portions of samples prepared by mixing pure flowers
of sulfur and activated carbon were sublimed. A mixture of flowers of
sulfur and acid-washed carbon (10 g sulfur and 100 g carbon), after
testing, still contained 84 pct of the original sulfur present when
vacuum distilled for 1 h at approximately 30 microns pressure and a
temperature of 133° C. Only a trace of condensed sulfur was evident in
the cold arm of the apparatus (test 2).
A sample of carbon with adsorbed sulfur was prepared by oxidation of
an ammonium bisulfide solution. Analysis by extraction with CS2
indicated a sulfur loading of 1.99 g of sulfur per 100 g of carbon in
the mixture. This mixture was vacuum distilled for 4 h at temperatures
and pressures ranging from 95° to 130° C at 22 to 29 microns (test 3)
and from 108° to 130° C at 13 to 16 microns (test 4). In these tests,
none of the sulfur originally present was removed by the procedure, and
there was no condensation of sulfur on the cold side of the apparatus.
Based on the above test work, sublimation of sulfur in the presence
of carbon, whether adsorbed on its surface or as a simple physical
mixture, does not occur at, or near, the same P-T conditions as it does
with sulfur alone. The sulfur vapor apparently has a high affinity for
the activated carbon surface. Because of this and the difficulties in
pumping down a system containing activated carbon, vacuum distillation
under the conditions examined does not appear to be a promising route
for sulfur recovery. However, additional work is continuing to
investigate higher temperatures and other conditions.
142
*** RECOVERY OF SULFUR FROM CARBON BY LEACHING
Potential methods for leaching sulfur from activated carbon were
examined. Based upon practical and environmental considerations,
liquid anhydrous ammonia was selected as a potential candidate for
extracting sulfur from carbon. The solubility of pure sulfur in liquid
ammonia as a function of temperature (Ruff and Hecht, 1911) is given in
table II.
Several preliminary attempts were made to dissolve pure flowers of
sulfur in liquid ammonia at ambient pressure and -33.4° C (boiling
point of NH3 at 1 atm). These attempts proved to be entirely futile.
Ruff and Hecht (1911) delineated a most important aspect of sulfur
dissolution in ammonia. That is, the temperature must be greater than
-11.5° C for rhombic sulfur to initially dissolve, and once dissolved,
it will remain in solution below -11.5° C. At the boiling point of
liquid ammonia, -33° C, the rate of dissolution of rhombic sulfur is
nil. Consequently, dissolution must be carried out under elevated
pressures and at temperatures above -11.5° C. At present, the sulfur
species that occur on the carbon surface are not known; consequently,
the optimal leaching temperature is also unknown.
The following section describes laboratory work to evaluate the
potential of using liquid anhydrous ammonia for stripping sulfur from
activated carbon. Ten-gram batch leaching tests were carried out as
screening tests to determine if anhydrous ammonia had any potential for
leaching sulfur. A 60-g leaching test was done to obtain a sulfur mass
balance on the leaching process and to determine the maximum degree of
sulfur extraction that could be expected using repeated leaching with
liquid ammonia.
MATERIALS AND PROCEDURES
Materials
The following materials were used in the laboratory studies described
in this report:
6Analysis of specific carbon used in tests by Bureau of Mines
Anhydrous Ammonia - Liquid anhydrous ammonia was prepared by
condensing commercial (Linde) ammonia using a stainless steel
condensing coil which was immersed in a dry ice-acetone bath.
144
The liquified ammonia was prepared immediately before use and
kept in a covered beaker.
Sulfur-Loaded Carbon - Prepared using Filtrasorb 400 activated
carbon and ammonium bisulfide by techniques previously described
(May and Sweeney, 1985). In brief, the carbon was placed in a
small laboratory column; shown in figure l, and loaded with
sulfur by cocurrently flowing NH4SH solution and air downward
through the column.
Two sulfur-loaded carbon samples were prepared under similar, but
different, test conditions. These samples analyzed 7.8 pct and
13.4 pct total sulfur and were used for 10-g batch leaching tests
and the 60-g batch test, respectively.
Apparatus
Leaching experiments were carried out using a 3.81-cm-diam by
34.93-cm-long laboratory-scale reactor in which liquid ammonia could be
contacted with carbon that had been previously loaded with sulfur. The
reactor was originally constructed from a carbon steel pipe (10-g batch
tests) and subsequently changed to a 304 stainless steel pipe (60-g
batch test).
Procedure
The general test procedure was as follows:
1. A 10- to 11-g charge of sulfur-loaded carbon was placed in the
pipe reactor, which was then sealed and prechilled in a dry ice-acetone
bath (about -50° C).
2. The reactor was removed from the bath and opened, and a
predetermined volume of liquid anhydrous ammonia (25 to 250 mL) was
poured into the reactor. The reactor was quickly resealed, shaken, and
then placed horizontally in a water-ice bath at 0° C.
145
146
3. The reactor was agitated vigorously 30 s every 5 min for a
predetermined period (10 to 60 min). During this time period, the
cylinder was allowed to warm up from -50° to near 0° C, thus passing
through the -11.5° C temperature required for dissolution of rhombic
sulfur.
4. The reactor was removed from the ice bath and then chilled in
the dry ice-acetone bath to reduce the ammonia vapor pressure to less
than atmospheric.
5. The reactor was positioned vertically, and the discharge valve
was opened slowly, allowing the sulfur-laden liquid ammonia to flow
into a beaker. The volume of ammonia recovered was noted to be about
25 mL less than the feed volume for each test.
6. After discharge of the ammonia from the reactor, with resulting
pressure letdown, the leached carbon was removed from the reactor for
subsequent analyses.
7. The sulfur-laden ammonia was allowed to evaporate in a beaker
under ambient conditions to produce a sulfur product which was then
refined using carbon disulfide to determine the actual amount of sulfur
extracted from the carbon.
RESULTS AND DISCUSSION
Ten-Gram Batch Leaching Tests
The results of the 10-g tests, summarized in table III, showed that
sulfur could indeed be leached from a sulfur-loaded carbon that
analyzed 7.8 pct total sulfur by LECO. A recovery of 73 pct was
obtained with a leaching time of 20 min. The volume of ammonia used
for leaching had little influence on overall sulfur recovery, probably
147
because the sulfur is so highly soluble in ammonia, 30 to 40 wt pct,
according to Ruff and Hecht (1911).
NOTE.--Sulfur-load carbon feed was 7.8 pct total sulfur. Recovery isbased on carbon disulfide extraction of the sulfur residue remainingafter evaporation of the sulfur-ammonia leach liquor. Recoveries havebeen corrected for hold-up losses in the reactor and for minordifferences in the weight of carbon used in several tests (11.0 vs10.0 g).
Sixty-Gram Batch Test
This test used 60 g of activated carbon that had been loaded with
sulfur under somewhat different conditions than the carbon used for the
10 g tests (a larger loading column was used and the bisulfide was
stage-added). The total sulfur content by LECO analysis of this carbon
was 13.4 pct. Since the inherent sulfur in the as-received carbon was
0.8 pct sulfur (by LECO), the actual sulfur loading, or "potentially
recoverable sulfur," on the carbon due to NH4SH conversion was 13.4 -
0.8, or 12.6 pct sulfur.
Sixty grams of activated carbon (136 mL) was placed in the stainless
steel pipe reactor. The reactor was then prechilled to -50" C in the
148
acetone-dry ice bath for several minutes. The reactor was then opened,
and 250 mL of liquid NH3 was added. The reactor was closed, placed
in an ice-water bath for 20 min, and shaken vigorously every 5 min. At
the end of the leach the reactor was again chilled to -50" C, and the
sulfur-containing ammonia solution was discharged into a 250-mL
volumetric flask and weighed. The reactor was rechilled to -50" C,
250 mL of liquid NH3 was added, and the carbon was releached for
another 20 min in the water bath. At the end of this second leach, the
second sulfur-containing ammonia solution was collected, and a third
releaching was carried out using the same procedure. Each of the three
sulfur-ammonia solutions was allowed to evaporate, and the amount of
sulfur contained in each evaporite was determined by carbon disulfide
extraction. Also, the ammonia gas produced during evaporation of the
first sulfur-containing ammonia solution was passed through a gas
scrubbing train containing ZnC12 to determine if H2S gas was being
evolved. No H2S gas was detected. Lastly, the carbon that had
undergone three leaching steps with ammonia was analyzed by LECO for
total sulfur, and by carbon disulfide extraction for elemental sulfur.
Results
The results of repeatedly leaching the 60-g (136-mL) carbon sample
three times with liquid ammonia (250 mL per stage) are summarized in
table IV.
149
Closure of the sulfur mass balance was excellent. From the data it
is seen that 7.26 g of sulfur from the 60 g of sulfur-loaded carbon
reported to the ammonia leachate and leached carbon products, out of a
total of 7.56 g of sulfur (from 60-g carbon at 12.6 pct sulfur = 7.6 g
sulfur). Thus, about 96 pct of the sulfur was accounted for. About 47
pct of the total sulfur was recovered in the first ammonia leaching
stage, and the total sulfur recovery after three leaching stages was
68.2 pct. The results of this leaching test showed that essentially
all of the sulfur could be accounted for in the leachate and leached
carbon products.
*** CARBON RECYCLABILITY
The purpose of this screening test was to determine if the carbon had
the potential for recycling to the sulfur-loading stage after sulfur
had been leached from it using liquid ammonia.
Leached carbon products from several of the 10-g ammonia leaching
tests (7.8 pct total sulfur by LECO before leaching) were combined into
one sample analyzing 3.6 pct sulfur. This sample was then releached
150
with 250 mL liquid ammonia for 20 min. A 40.0-g split of the releached
carbon was used for a sulfur reloading test using the carbon column
apparatus; the remainder was analyzed for sulfur using both carbon
disulfide extraction and LECO total sulfur analysis. The carbon column
reloading test was done using the same basic technique used in the
original loading. The results are summarized in table V.
The carbon that had been subjected twice to leaching by anhydrous
ammonia, was reloaded to a total sulfur content of 7.4 pct sulfur, only
slightly less than the 7.8 pct sulfur loading of the original
sulfur-loaded carbon used in the 10-g leach tests. Also the comparison
of the CS2 extraction analysis before leaching (6.15 pct sulfur) and
after leaching and reloading (6.6 pct sulfur) is favorable. Thus, it
appears that the carbon responds well for at least one complete duty
cycle. Obviously, additional cycling tests must be done in the future
to determine the effects of repeated cycling of the carbon. Lastly, it
should be mentioned that during the ammonia leaching test some
decrepitation of carbon occurred (appearing to be minor), which should
be monitored in future tests.
151
*** FLOWSHEET DEVELOPMENT
Based on the present research, a preliminary conceptual flowsheet has
been prepared for the overall conversion of phosphogypsum to sulfur,
incorporating ammonia leaching of the carbon to produce an elemental
sulfur end product.
Figure 2, prepared by T. Phillips (1986) of the Bureau of Mines,
depicts several stirred reactors in which the calcium sulfide is
reacted with ammonium bicarbonate produced by scrubbing the C02- rich
CaS kiln off-gas with recycled ammonia. This ammonium bicarbonate
could be used in place of CO2 and NH40H reagents. The reactors
will probably require cooling since the reaction is exothermic. The
precipitated calcium carbonate product is washed counter-currently
prior to disposal, with the wash water being sent to the CaS
dissolution step. It is expected that the cake will wash easily, and
washing should be possible using pan or horizontal belt filters. The
final washed carbonate cake should contain very little ammonium
bisulfide solution but may have to be heated to drive off residual
bisulfide in the form of H2S, prior to disposal or sale. The residual
H2S could be recycled back to the stirred reactors. Some of the
pregnant ammonium bisulfide liquor may be recycled to keep the pulp
density of the slurry in the reactors at a level where stirring will
not be a problem. Information has not yet been developed regarding
number of reactors required, operating pressure, pulp density, cake
washability, etc.
Figure 2 shows the possible route for converting the ammonium
bisulfide liquor to a final sulfur product. In this flowsheet the'
ammonium bisulfide would be oxidized to sulfur, in the presence of air,
152
using a column loaded with a catalyst such as activated carbon. The
effluent from the column may have to be recycled upon itself. Work is
presently in progress to understand the mechanism of conversion since
this may help in selecting more effective catalysts. In any event, it
is believed that the sulfur that is loaded onto the catalyst must be
removed in a separate cycle, possibly after drying of the catalyst.
Stripping is envisioned to be accomplished in situ, similar to the
loading-stripping cycle done in ion exchange. The preliminary
laboratory work presented in this paper has indicated that anhydrous
ammonia will solubilize and strip off the sulfur.
The stripping of sulfur from the loaded column would be done with
anhydrous ammonia at nominal pressures and temperatures of 60 psi and
0" C, respectively. Predrying of the carbon with steam may be
required. Considerable work will have to be done to determine the most
economic operating conditions. The data also show that the dissolution
rate should be rather fast, with 10 to 20 min probably being
sufficient. Also, while excess quantities of ammonia have probably
been used in the present test work, it is reasonable to believe that
several volume equivalents (two to three volumes of ammonia per volume
of carbon) may be satisfactory to strip most of the sulfur from the
carbon. The degree of stripping required for reuse of the carbon in
the loading cycle, however, will have to be experimentally determined.
Following stripping, the sulfur-containing liquid ammonia would be
concentrated by evaporation to produce a saturated sulfur-ammonia
solution. Further evaporation and melting of the sulfur would be done
in the next stage. All ammonia gas products would be condensed and
recycled for the next stripping stage.
154
It is believed that the catalyst loading and stripping is technically
feasible and indications are good that the catalyst can be recycled, at
least to some extent. Future work must be directed toward process
development to improve sulfur recovery, on both catalyst loading and
stripping. Also, processing using liquid ammonia can be quite energy
intensive, and the key to its economic success will be in the design of
a good heat recuperation system.
Based upon this preliminary flowsheet, and assuming high yields in
each step of the process, the Bureau has developed operating cost
estimates. While the figures are expected to be continually refined,
the present estimates for the overall process of converting
phosphogypsum to elemental sulfur, using coal ($50/ton) as a reductant,
are summarized in table VI. The cost for reducing phosphogypsum to
calcium sulfide is estimated to be about $94/ton of sulfur produced.
The subsequent processing of calcium sulfide to produce NH4SH
solution is about $12/ton, and oxidation of the bisulfide using a
carbon catalyst is about $24/ton. Stripping of the sulfur from the
catalyst, using anhydrous ammonia, is estimated to be about $22/ton.
Thus, the stripping step is roughly 15 pct of the total cost of
$152/ton of sulfur produced from phosphogypsum.
155
Coal cost estimated at $50/ton for reduction stage.
*** CONCLUSIONS
The results of laboratory studies using activated carbon that had
been loaded with sulfur through oxidation of ammonium bisulfide
solutions have shown that -
1. Recovery of an elemental sulfur product from the carbon using
vacuum distillation at temperatures below about 150" C is difficult.
2. Anhydrous liquid ammonia can extract about 70 pct of the sulfur
from the carbon to produce an elemental sulfur end product. Leach
times are on the order of 20 min.
3. The conditions for dissolution of rhombic sulfur using ammonia
are critical, with temperatures above -11.50'C being required. Exact
conditions, from a practical and economic viewpoint, for the sulfur
species existing on the carbon remain to be determined.
156
4. Good closure can be obtained on a sulfur mass balance on the
operation of leaching sulfur from carbon using ammonia. Ninety-six pct
of the sulfur has been accounted for either as a sulfur product after
ammonia evaporation of the sulfur-laden ammonia leachate, or on the
carbon after leaching. Very little sulfur, if any, is evolved as
H2S.
5. Based on one complete cycle, prospects for recycling carbon after
ammonia leaching appear to be promising.
6. A conceptual flowsheet for the total process of converting
phosphogypsum to elemental sulfur, incorporating ammonia‘ leaching of
the carbon, has been developed.
7. The Bureau has estimated that the ammonia leaching step for
sulfur recovery from carbon would cost about $22/ton of sulfur
produced The total operating cost for a plant producing an elemental
sulfur product from a phosphogypsum feed would be about $152/ton of
sulfur produced.
*** ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance and advice of Dr.
David P. Borris, Executive Director of FIPR, and Mr. Mike Lloyd,
Research Director of FIPR. The voluntary cooperation of the following
Florida phosphate companies in offering valuable advice and suggestions
on the research efforts is also gratefully acknowledged: Agrico
Chemical Co., CF Industries, Conserve, Farmland Industries, Gardinier,
Inc., International Minerals and Chemical Corp., Occidental Chemical
co., Royster Co., and W. R. Grace.
157
*** REFERENCES
Boyd, R., and T. D. Phillips. 1985. Economics of World SulfurSupply to the End of the Century. Proceedings of the FertilizerSociety, No. 234, London, 36 pp.
Morse, D. E. Sulfur. 1985. Section in Bureau of Mines MineralCommodity Summaries, pp. 152-153.
May, A., and J. W. Sweeney. 1982. Assessment of Environmental ImpactsAssociated With Phosphogypsum. BuMines RI 8639, 19 pp.
Ragin, M. M., and D. R. Brooks. Recovery of Sulfur From Phosphogypsum.Part 1. - Conversion of Calcium Sulfate to Calcium Sulfide.Second International Symposium on Phosphogypsum, Miami, FL, Dec.10-12, 1986.
Smelley, A. G., M. M. Ragin, D. R. Brooks, and B. J. Scheiner.Conversion of Phosphogypsum to Sulfur. Part 1. - Reduction ofCalcium Sulfate to Calcium Sulfide. Proceedings, Third Seminar onPhosphogypsum, Tampa, FL, Dec. 5-7, 1985.
May, A., and J. W. Sweeney. Conversion of Phosphogypsum to Sulfur.Part 2. - Oxidation of Calcium Sulfide to Sulfur. Proceedings,Third Seminar on Phosphogypsum, Tampa, FL, Dec. 5-7, 1985.
Lefers, J. B., W. T. Koetsier, and W. P. M. Van Swaaij. 1978. TheOxidation of Sulfide in Aqueous Solutions. The Chem. Eng. J.,v. 15, pp. 111-120.
Chen, K. Y., and J. C. Morris. 1972. Kinetics of Oxidation of AqueousSulfide by 02. Environmental Sci. and Tech., v. 6, No. 6,pp. 529-537.
Ruff, O., and L. Hecht. 1911. Sulfammonium and Its Relationship toNitrogen Sulfide. Z. anorg. Chem., v. 70, pp. 49-69.
Phillips, T. U.S. Bureau of Mines. Preliminary Cost Estimate -presented at the Florida Phosphate Institute Technical AdvisoryCommittee Meeting, Bartow, FL, August 16, 1986.
153
THE BACTERIAL CONVERSION OF PHOSPHOGYPSUM INTO ELEMENTAL SULFUR
Barrie F. TaylorProfessor, Division of Marine and Atmospheric ChemistryRosenstiel School of Marine and Atmospheric Science
University of MiamiMiami, Florida 33149-1098
ABSTRACT
The main commercial use for elemental sulfur is in the production offertilisers. World demand for sulfur will exceed supply in the 1990s andbiotechnological methods for the recovery of sulfur from phosphogypsum shouldthen be economically feasible. Associations of sulfate-reducing andphotosynthetic bacteria generate elemental sulfur in some naturalenvironments. Defined mixed cultures (Desulfovibrio desulfuricans andChromatium vinosum) formed elemental sulfur in laboratory experiments. Theextent of sulfate conversion into sulfur was good but the association wasunstable because of excessive rates of sulfide production by thesulfate-reducer. Conditions that promote the growth or activity of thephotosynthetic bacterium would stabilise the interaction. A biotechnology forsulfur recovery from phosphogypsum, using sulfur-metabolising bacteria, willprobably be based on sulfate-reduction driven by organic wastes (e.g. sewage).A commercial process with photosynthetic bacteria would be technologicallydifficult. However, genes from photosynthetic sulfur bacteria might be used togenetically engineer sulfate-reducers that accumulate sulfur, rather thansulfide, in a single-stage reactor. Sulfide oxidation could also proceed in asecond-stage reactor, either chemically or biochemically. Biochemicaloxidation would use chemolithotrophs whereas chemical conversion might entailprocessing with dimethyl disulfide.
INTRODUCTION
Elemental sulfur, after conversion into sulfuric acid, is mostly used tomanufacture fertilisers; this fact emphasizes the importance of sulfur to thephosphate industry (Boyd and Phillips, 1985). World demand for elementalsulfur will exceed production during the 1990s, according to recent estimates(Boyd and Phillips, 1985; Gilbert, 1985; Urquhart, 1986). Conventionalsources of sulfur will need to be supplemented by new technologies. Oneprobable innovation is the recovery of sulfur from sulfate wastes, such asphosphogypsum. The increasing price of sulfur means that biotechnologicalmethods for its recovery from sulfate wastes may soon be economically viable(Boyd and Phillips, 1985).
Butlin and Postgate in 1954 described a shallow lake in North Africa inwhich sulfur accumulated. Spring waters, issuing through gypsum deposits, fedthe lake. The lake was anoxic with high levels of H S, and the sedimentscontained up to 50% of their dry weight as elemental sulfur. The local peopleharvested sulfur from the lake during the dry season- Red gelatinous masses,with underlying green and black material, were present in the water andsediments. Butlin and Postgate isolated pure cultures of green (Chlorobiumsp.) and red (Chromatium sp.) photosynthetic sulfur bacteria from thegelatinous masses. They also established an "artificial lake" in a laboratoryexperiment. A synthetic medium, resembling the lake water in composition, wasinoculated with gelatinous material and a crude culture of sulfate-reducers.
159
FL). The HPLC system consisted of a Waters UK6 injector and Waters Model 6000Apump (Waters Associates, Milford MA) with a Conductomonitor III Detector andShimadzu Data Processor. The solvent was 0.5 mM H2S04 at a flow of 0.5 ml/min.Typical retention times in minutes for some organic acids were: citrate, 6.5;5-ketogluconate, 6.8; gluconate, 7.7; malate, 10.0; lactate, 10.3; fumarate,10.4; succinate, 12.9; acetate, 13.3.
RESULTS
Initial experiments with mixed cultures showed hydrogen sulfideproduction in accord with the equation (Table I):
2 Lactate + SOElemental sulfur was not sulfate reduction occurred so
= 2 Acetate + CO2 + S2-
rapidly that hydrogen sulfide accumulated to levels that were toxic orinhibitory to the photosynthetic organisms.
Elemental sulfur was formed when the inoculum of the phototrophconsiderably exceeded that of the sulfate-reducer. Sulfide did not accumulateand about 75% of the sulfate used was converted into elemental sulfur in mixedcultures of D. desulfuricans and C. vinosum (Table II). In this experimentacetate accumulation did not correspond with lactate use, probably because thephotosynthetic bacterium grew and consumed some of the acetate.
DISCUSSION
Elemental sulfur production was rapid with D. desulfuricans and C.vinosum but the associations did not survive more than one subculturing. Theresult, however, indicates the possibility of establishing mixed systems ofsulfate-reducers and photosynthetic bacteria that generate elemental sulfurfrom sulfate at rapid rates. The establishment of a stable interactionprobably depends on balancing sulfate reduction and sulfide oxidation rates.This could be achieved either by selecting appropriate strains or by promotingthe growth of the phototrophs, for example with a specific organic compound(e.g. acetate).
Sulfur production by associations of sulfate-reducer(s) photosynthetic bacteria requires an organic supplement because the system isnot thermodynamically self-supporting (Taylor, 1986). On a large scale thesupplement could be biomass or a waste such as sewage. Even with a cheaporganic supplement, commercial systems using photosynthetic organisms areprobably not economically or technologically realistic because of the need forlight.
Reactors fed with sewage and sulfate (phosphogypsum) are possible forgenerating H2S on a commercial scale (Butlin, et al., 1956, 1960). Asingle-stage process that accumulates elemental sulfur rather than H2S isconceivable but it requires a new type of sulfate-reducing bacterium (Taylor,1986). Elemental sulfur is not an intermediate in sulfate reduction butsulfate-reducers that re-oxidize sulfide to elemental sulfur as a survivalmechanism might exist in high H2S environments. Alternatively the desiredorganism could be genetically engineering using the genes, for the oxidationof sulfide to elemental sulfur, from sulfur lithotrophs (Taylor, 1986).
"Two-stage processes would use separate reactors the oxidation theH2S to elemental sulfur; either chemical or even biochemical- with dark,aerobic oxidation by thiobacilli (Sokolova, 1961). A recent commercial-scale
process dimethyl disulfide recover and generate from sour gaswells (Pennwalt, 1986) might also be
CONCLUSIONS AND RECOMENDATIONS
1. The imminent shortfall in world 'sulfur production, and projected increases for this prime industrial
recovering sulfur from sulfate wastes be investigated.2. Bacterial sulfate-reduction driven by organic wastes provides the
basis of a biotechnology for recovering sulfur from phosphogypsum.3. Several methods 'are possible for converting the H2S, generated from
sulfate-reduction, into elemental sulfur. Sulfate-reducing bacteria thataccumulate elemental sulfur rather than sulfide may allow a single-stageprocess. The other choices are two-stage systems with either biochemical orchemical oxidation of the H2S.
ACKNOWLEDGEMENTS
Financial support supplied by a grant (OCE-8516020) from the NationalScience Foundation and a contract (84-01-39) from the Florida Institute ofPhosphate Research.
162
REFERENCES
Boyd, R. and T.D. Phillips. 1985. Economics of World Sulphur Supply to the Endof the Century, pp. l-41. The Fertiliser Society, The British SulphurCorporation Limited, London.
Butlin, K.R. and J.R. Postgate. 1954. The Microbiological Formation of Sulphurin Cyrenaican Lakes, pp. 112-122. In: J. Cloudsley-Thompson (ed.),Biology of Deserts, Institute of Biology, London.
Butlin, K.R., S.C. Selwyn and D.S. Wakerley. 1956. Sulfide Production fromSulfate-Enriched Sewage Sludge. J. Appl. Bacteriol. 19: 3-15.
Butlin, K.R., S.C. Selwyn and D.S. Wakerley. 1960. Microbial SulfideProduction from Sulfate-Enriched Sewage Sludge. J. Appl. Bacteriol. 23:158-168.
Gilbert, K. 1985. Sulfur Market Tightening Worldwide. Chem. Eng. News 63:14-15.
Guerrero, R., E. Montesinos, C. Pedros-Alio, I. Estene, J. Mas, H. VanGemerden, P.A.G. Hofman and J.F. Bakker. 1985. Phototrophic sulfurbacteria in two Spanish lakes: vertical distribution and limitingfactors. Limnol. Oceanogr. 30: 919-931.
Pennwalt Industries. 1986. Dimethyl disulfide for sour gas wells. Organotopics4: 4.
Pfennig, N. and H.G. Truper. 1981. Isolation of members of the familiesChromatiaceae and Chlorobiaceae, pp. 279-289. In: M.P. Starr, H. Stolp,H.G. Truper, A. Balows and H.G. Schlegel (ed.), The Prokaryotes, Vol. I.Springer-Verlag, New York.
Postgate, J.R. 1984. Genus Desulfovibrio, pp. 666-672. In: N.R. Krieg (ed.),Bergey's Manual of Systematic Bacteriology, Vol. l. and Wilkins,Baltimore.
Sokolova, G.A. 1961. Microbiological production of sulfur from sulfide seamwaters. Microbiology 29: 638-641.
Taylor,. B.F. 1986. The microbiological recovery of sulfur from phosphogypsum,
pp. 79-91. In Proceedings of the Third Workshop on By-Products of thePhosphate Industries, Florida Institute of Phosphate Research, Bartow,Florida.
Taylor, B.F., T.A. Hood and L.A. Pope. 1987. Determination of elemental sulfurin cultures and sediments by HPLC after reaction with triphenylphosphine.Ann. Meet. Amer. Soc. Microbial. Q-145: 306.
Urquhart, J. 1986. Sulfur Prices Approach Record Levels-as Supplies Shrink andShortage Looms. Wall Street Journal 207(72): 44.
163
STABILIZATION
STUDIES ON THE FORMATION OF CEMENTITIOUS COMPOUNDS USING PHOSPHOGYPSUM
G.L. Valenti; Professor, Dipartimento di Ingegneria dei Materiali e della Produzione, Universita di Na-poli, Italy
L. Santoro, Professor, Dipartimento di Chimica, Universita di Napoli, ItalyJ. Beretka, Principal Research Scientist, CSIRO Division of Building Research, Melbourne, Australia
ABSTRACT
The formation of calcium sulfoaluminate hydrates has been studied in a number of hydratory sy-stems with the purpose of evaluating the possibility of disposal of phosphogypsum by conversion intoproducts useful as building materials. Lime, aluminium hydroxide, granulated blast-furnace slag, flyash and portland cement have been used as sources of CaO and Al2O3.
The system phosphogypsum-lime-aluminium hydroxide-water has been investigated in the tempe-rature range 25-80°C and for different composition ratios. A temperature of 60°C and a compositionstoichiometric for calcium trisulfoaluminate hydrate (ettringite) formation have been found to give op-timum conversion.
The system phosphogypsum-slag-water has been studied at 25°C in the presence of a small amountof hydrated lime and in the range of phosphogypsum/slag composition ratio up to 20/100.
The system phosphogypsum-fly ash-lime-water has been studied at 25 and 40°C and in the rangeof composition up to 30 parts of phosphogypsum for 100 parts of a 60/40 mixture of fly ash and lime.
In the systems containing slag and fly ash phoshogypsum is equally or more efficacious than puregypsum with regard to the formation of the major hydration products, calcium silicate hydrate andettringite.
The hydraulic behaviour of a system composed of 50, 30, 20 and 15 parts of fly ash, calcined pho-sphogypsum, portland cement and lime, respectively, has been investigated at 21 and 55°C. The fastformation of calcium sulfate dihydrate and its conversion into ettringite contribute to early mechanical -strength and durability, respectively.
INTRODUCTION
The study of hydratory systems containing gypsum, lime and alumina is of interest in the manufac-ture of lightweight building elements. by hydrothermal reactions (Azuma and Ichimaru, 1976; Azumaet al., 1976) and in the evaluation of the hydraulic behaviour of energy-saving binders like blended por-tland cements containing substantial amounts of plaster of Paris (Mehta, 1980). This interest is due tothe formation of calcium sulfoaluminate hydrates, mainly ettringite, a phase characterised by low densi-ty, satisfactory mechanical strength and good water resistance.
A wide range of solid industrial by-products can be used as sources of the reactants. This wouldresult in reduced exploitation of natural resources, energy saving and also in protection of the environment.
Phosphogypsum and other by-product gypsums can be used as the source of calcium sulfate, andblast furnace slag, fly ash or other appropriate industrial by-products can be used as the source of cal-cium and aluminium oxides. Low-calcium fly ash also requires addition of lime.
This paper reports the results of the investigations carried out on the hydration of the systems:a) phosphogypsum-lime-aluminium hydroxide;b) phosphogypsum-granulated blast furnace slag;c) phosphogypsum-fly ash-lime;d) calcined phosphogypsum-fly ash-Portland cement-lime.
167
SYSTEM PHOSPHOGYPSUM-LIME-ALUMINIUM HYDROXIDE
The molar rations between the reactants CaO,Al(OH), and CaSO4
.2H2O were varied betweenthe limits 3: 2: 1 and 3: 2: 3. 6. The water/solidratio was 0.75. Curing was carried out under 100%r.h.; the temperature ranged between 25 and 80%,and aging times between 1 and 32 days (Santoroet al., 1984).
Optimum conditions for the formation of et-tringite and for the consumption of phosphogyp-sum and other reactants were found to be 60°Cand the composition stoichiometric for ettringiteformation.
Figure 1 is referred to optimum conditionsand reports the concentrations of phosphogypsum,AI(OH)3 and ettringite vs. time, evaluated byquantitative thermal analysis.
SYSTEM PHOSPHOGYPSUM-GRANULATED BLAST FURNACE SLAG
Thermal analysis proves that in this system the effect of phosphogypsum is two-fold. Firstly, it reactswith the reactive calcium and aluminium oxides in the slag to form ettringite, and secondly it enhancesthe formation of calcium silicate hydrate, which account for the well known hydraulic properties of theslag in strongly alkaline media. Ettringite and calcium silicate hydrate are responsible for the develop-ment of mechanical strength of these binders at early and later ages, respectively.
Three compositions containing 5, 10 and 20 parts of phosphogypsum per 100 parts of slag respecti-vely, have been examined. They will be referred to as SP5, SPlO and SP20, respectively. For the sake
of comparison, the behaviour of a system without CaSO4.2H2O (denoted by S) and three compositions,
denoted by SG5, SGl0 and SG20, in which pure gypsum replaced phosphogypsum in sufficient quanti-ties to ensure the same slag/calcium sulfate ratios as for the compositions SP5, SP10 and SP20 werealso investigated. In all cases 0.5 parts of Ca(OH)2 per 100 parts of slag were added to each composi-tion as the alkaline activator. The water/solid ratio was 0.5, the curing temperature was 25°C and agingtimes were up to 182 days (Valenti et al., 1984). The amounts of unreacted slag and gypsum were deter-mined according to the extraction procedures reported by Kondo and Ohsawa (1968).
Figures 2 and 3 show the percentages of reacted gypsum and slag converted to calcium silicate hy-drate (CSH), as function of aging time. The systems containing phosphogypsum exhibited greater reac-tivity with regard to both ettringite and CSH formation.
SYSTEM PHOSPHOGYPSUM-FLY ASH-LIME
Similarly to the previous system, the two major hydration products were ettringite and calcium sili-cate hydrate.
Three compositions, namely 10, 20 and 30 parts of phosphogypsum per 100 parts of a 60: 40 flyash-Ca(OH)2 mixture, have been examined. They will be referred to as ALPl0, ALP20 and ALP30,respectively. As for the previous system, the behaviour of a system without gypsum (denoted by AL)and three compositions, denoted by ALGl0, ALG20 and ALG30, in which pure gypsum replaced pho-sphogypsum, was also investigated. The water/solid ratio was 0.6, and aging times were up to 182 daysat 25°C and up to 56 days at 40°C (Santoro et al., 1986). The amount of unreacted lime was determinedby extraction following the Franke method (Shebl and Ludwig, 1978).
Table I shows the aging times at which calcium sulfate dihydrate is no longer detected by differen-tial thermal analysis. The kinetics of calcium sulfate consumption is favourably influenced by increasingtemperature. Phosphogypsum proves to be more reactive than pure gypsum at the highest level of addition..
Figures 4 and 5 show the amount of reacted lime as a function of aging time at 25 and 40°C respec-tively. The following considerations can be made: (a) adding phospho- or pure gypsum increases theamount of reacted lime; (b) due to a higher hydration rate, the lime fractional conversions at, 40°C and56 days are roughly the same as those reached at 25°C after as many as 182 days.
At 25°C lime conversion increases with increasing the initial sulphate concentration, and no majordifference is observed between phospho- and pure gypsum. On the other hand, at 40°C lime conversiondoes not seem to be influenced by the amount of pure gypsum added, while it reaches higher valueson adding increasing amounts of phosphogypsum.
Thermal analysis has shown that in the presence of added gypsum, only ettringite and CSH form.At the times at which gypsum is completely converted (Table I), by means of simple stoichiometric cal-culations, it is possible to divide the total amount of reacted lime into the two parts which are chemicallycombined in ettringite and CSH. Table II shows, for any of the compositions with added gypsum, theamount of lime converted to CSH at the longest ageing times, i.e., 182 days at 25°C and 56 days at40°C. Temperature and type of gypsum being the same, the amount of lime converted to CSH decreasesas the amount of added gypsum increases. However, the extent to which this happens is greater in thecase of pure gypsum.
169
parts of lime to 100 parts of a 50: 30: 20 mixture of fly ash, calcined phosphogypsum and portlandcement, has been studied.
The. chemical composition of the raw materials is shown in Table III.
Compressive strengths of pastes (w/s ratio = 0.6) cast into cube moulds with 25 mm edges andcured at both 21 and 55°C at 100% r.h. were measured. The results are shown in Table IV.
Differential thermal analysis has revealed that gypsum is almost completely converted after 90 daysat 21°C while it is consumed within only 4 days at 55°C.
CONCLUSIONS
Phosphogypsum can be advantageously employed in hydratory mixtures containing lime and alu-mina to give products, based on calcium sulfoaluminate hydrates, which are of potential use as buildingmaterials.
Other industrial by-products like blast-furnace slag and fly ash (eventually mixed with lime) canbe utilized as sources of CaO and Al2O3. The reactive silica present in slag and fly ash generates cal-cium silicate hydrates which add mechanical strength to the products.
in these systems, when compared to pure gypsum, phosphogypsum not only appears to be equallyor more reactive with regard to the formation of calcium trisulfoaluminate hydrate, but also stimulatesmore effectively the formation of calcium silicate hydrate.
Mechanical tests on pastes of a binder composed of calcined phosphogypsum-fly ash-Portland cement-lime have shown that phosphogypsum can be effectively used in energy-saving blended portland cementsin which the setting of CaSO4
. l/2 H2O and portland cement give mechanical strength at early ages,while later contributions to strength and durability arise from the formation of calcium trisulfoalumina-te and silicate hydrates.
REFERENCES
Azuma, T. and K. Ichimaru. 1976. Calcium aluminate sulfate-based inorganic hardened product. JapanKokai 76 62, 826.
Azuma, T., K. Ichimaru, T. Murakami and K. Tateno. 1976. Calcium aluminate monosulfate hydrate.Ger. Offen. 2, 551, 310.
Kondo, R. and S. Ohsawa. 1968. Studies on a method to determine the amount of granulated blastfur-nace slag and the rate of hydration of slag in cements. Proceedings of the fifth International Sym-posium on the Chemistry of Cements. 4:255-262.
Mehta, P.K. 1980. Investigation of energy-saving cements. World Cement Technology. 11 (4): 166-177.
Santoro, L., G.L. Valenti and G. Volpicelli. 1984. Application of differential scanning calorimetry tothe study of the system phosphogypsum-lime-aluminium hydroxide-water. Thermochimica Acta.74: 35-44.
Santoro, L., I. Aletta, and G.L. Valenti. 1986. Hydration of mixtures containing fly ash, lime and pho-sphogypsum. Termochimica Acta. 98: 7l-80.
Shebl, F.A. and U. Ludwig. 1978. Investigations relating to the determination of calcium hydroxideby the Franke method. Zement-Kalk-Gips. 10: 510-515.
Valenti, G.L., L. Santoro, and G. Volpicelli. 1984. Hydration of granulated blast-furnace slag in thepresence of phosphogypsum. Termochimica Acta. 78: 101-112.
172
FLYASH-GYPSUM: GRANULES FOR THE PRODUCTION OF
PORTLAND POZZOLANA CEMENT
N. BHANUMATHIDAS, D IRECTOR (TECHNICAL)N. KALIDAS, PROJECT EXECUTIVE
Coromandel Gypsum Pvt. Ltd. Visakhapatnam (India)
ABSTRACT
Out of its many uses, flyash is well established for its role
as a pozzolana in the production of Portland Pozzolana Cement (PPC) .
However, the method of blending flyash with clinker is technically
a critical aspect because of the difference in specific gravities
of these two products to result in a non-homogenous mix.
One of the practices of blending is the intergrinding of clinker
and flyash, added with gypsum for its role of retardation, and
this method is highly prevalent for its superior techno-economic
virtues, over the other practice of inter-mixing ground cement
and flyash.
The fines of flyash and phosphogypsum often create a snag
by choking the feeding mechanism as well as posing handling problems.
The only solution to overcome these problems is to granulate the
flyash and gypsum together in such a ratio as to be suitable in
the right proportion to PPC. Such flyash-gypsum granules would
serve the cement industry with one product of two constituents
for easy dosing, resulting in better operative conditions and cost
effectiveness.
173
Taking the advantage of fact that hemihydrate sets in minutes
and converts to dihydrate, flyash powder has been intermixed with
hemihydrate, duly calcined out of refined phosphogypsum, added
with water and made into a thick paste to produce granules. These
granules have shown good strength development within 10-15 minutes
for stacking and, on subsequent drying, to withstand the stresses
during transportation.
For optimum economic feasibility, the calcined gypsum has
to be maintained at a low content. Hence, in order to study this
aspect flyash-dihydrate gypsum-calcined gypsum were mixed in various
proportions and the granules, thus produced, were studied with
reference to their strength developments.
The overall observations leave the authors with a conclusion
that the admixing of calcined gypsum has yielded acceptable strength.
towards flyash granulation.
INTRODUCTION
Flyash is a byproduct of thermal power -plants, being generated
at a rate of 25-40% of coal combusted, having accumulated to millions
of tons, particularly in India. It involves several hectares of
land for dumping and enormous costs for disposal in other means.
As the saying goes, “necessity is the mother of invention",
the problem posed by flyash has initiated innovations contributing
to convert this ash into cash. Research studies were undertaken
to find out the means and deeds for the gainful and effective
174
utilization of flyash especially in the construction activity.
It is another important sector, calls for the everlasting attention
of the country, which can consume any waste material of calcareous
and argillaceous nature to give fillip to the theme, 'Wealth from
waste’. Outcome of such studies was very encouraging, confirming
its suitability in the. multi-pronged fronts as follows:
a. As an argillaceous raw material in the manufacture of Portland
Cement;
b.
C.
d.
As a pozzolana in the manufacture of Portland Pozzolana Cement;
In making different types of bricks and other building units;
In the manufacture of sintered aggregates for making light
weight concrete;
e.
f.
As one of the ingredients in the concrete mix at site;
As an aggregate in road construction and
As a filler material, specially in the reclamation of land
and in mining;
It is wishful to study the physical, chemical and mineralogical
characteristics of flyash that have made it suitable towards above
uses.
PHYSICAL CHARACTERISTICS:
The definition of pozzolana states, 'A pozzolana is a siliceous
material which, while in itself, possessing no cementitious value
will, in finely divided form and in the presence of moisture,
175
chemically reacts with lime or cement to form compounds of cementi-
tious proper ties’ . It confirms that fineness is a single important
physical characteristic which influences the activity of flyash
more than any other physical factor. In general, the Blaine fineness
of flyash falls in the range of 4000-6000 cm2/g which indicates
higher degree of fineness than that of Portland Cement.
Though no established data is available to show that lime
reactivity is in direct relation to the fineness of flyash, in
general, it can be inferred that lime reactivity of flyash increases
with the fineness.
Coming to the physical constitution of flyash, it consists
of solid or hollow spherical particles of siliceous and aluminous
glass with small proportions of thin-walled, multi-faced polyhedrons
called Cenospheres. The existence of these spheroids is an added
advantage for the established use of flyash as a pozzolana and as
an aggregate in the concrete.
CHEMICAL CHARACTERISTICS :
The Chemical analyses of some of the Indian flyashes are given
in Table I. The Si02 of flyash reacts readily with CaO of lime.
or cement forms the different calcium silicate hydrates, the cementi-
tious compounds. Out of the hydrous calcium silicates formed,
the presence of monocalcium silicate is prominent, which has the
characteristic of relatively low solubility, contribution to water
tightness as well as to strength.
176
The chemical composition of the flyash varies depending on
the geological and geographical factors of the coal deposits, combus-
tion conditions of coal and collecting techniques incorporated
in the thermal plants.
MINERALOGICAL CHARACTERISTICS :
On instantaneous combustion, the mineralogical constituents
of pulverised coal react chemically and undergo thermal changes
to yield flyash with the following mineral phases :
FLYASH AS AN ACCEPTED MATERIAL :
The multi-directional study and analyses of flyash has inferred
that it is an appropriate material to be used in the construction
activity multi-purposely . Two serious problems, disposal of accumu-
lated flyash and production of building material to cater the needs
of growing construction industry, have been mutually solved, in
a single frame, when flyash was put into use.
Many advantages have been recorded by using flyash towards
each purpose :
1. The smooth discrete spherical particles of flyash lend mobility ,
through the ball bearing action, to the concrete thereby increas-
ing the workability of the mix. Because of this, consistency
factor is decreased resulting in a dense matrix. This compact-
ness reduces the bleeding and saggregation of concrete and
also the permeability of concrete.
2. In the manufacture of Portland Pozzolana Cement, flyash is
replacing, up to 15-25% of the clinker that is produced with
high inputs of power and fuel. Hence, the cost of production
of PPC is getting reduced to that extent and also resulting
in the savings of energy.
3. Use of flyash in concretes, as a replacement to Portland Cement
to certain percentages reduces the heat of hydration, thereby
making it suitable for mass concrete applications.
178
4. Sulphate resistance of PPC is higher than that of OPC owing
to the comparatively low contents of C3A.
5. Abundant availability of flyash renders a long range guarantee
to its consumers to meet the schedules of their production
programs without fail.
6. Flyash is more ideal than any other pozzolana, as it emerges
out from a process involving high temperatures thereby making
itself ready for use with its suitable chemical composition.
7. Use of flyash in concrete and precast building components
increases their thermo-insular behaviour in view of the existence
of hollow glass spheres.
The utilisation of flyash has some other contributions such
as:
a. Abatement of environmental pollution;
b. Conservation of non-renewable mineral resources.
Inspite of these many advantages, the powderous constitution
of flyash poses problems of handling & dosing, there by reducing
its scope of application. These problems are mainly faced by the
cement manufacturers who produce Portland Pozzolana Cement, since
flyash, as a pozzolana, is more conformable technically than any
of its other applications. Hence, use of flyash in this angle
can be increased provided it is supplied in some agglomerated form
such as granules, briquettes etc., making transportation and handling
tasks easy. This paper discusses such scope with technical virtues
179
and suitability of the product.
It is an universally established fact that gypsum, CaSO4.2H2O
is added to cement clinker, to control the initial reactions so
as to prevent flash set. As a general trend, mineral gypsum was
being used for this purpose. But the output of phosphogypsum,
as a by-product in the phosphoric acid industries at a rate of
4-5 tons for every ton of P2O5 produced, has’ posed similar problems
of dumping and disposal as those of flyash.The countries, where
mineral gypsum is scarcely available, have resorted to use phospho-
gypsum, after due refining, as a retarding agent in cement, thereby
establishing its suitability . After observing it, even the countries
having mineral gypsum sources, are snifting to phosphogypsum, which
results in the conservation of non-renewable mineral resources.
As a matter of fact, purity of CaSO4.2H2O in phosphogypsum is high
at 92-94% consistently throughout the source, as it emerges out
from a chemical process of constant process parameters. The case
is different with mineral gypsum,the purity of which is comparatively
low and varies from mine to mine, and at times within the same
source.
As phosphogypsum is also powderous posing handling problems,
the manufacturers of PPC have to face similar type of problems
with two ingredients . Especially, these fine powder. cause choking
in the hoppers, while feeding, to cause production problems.Because
of the fines, even the grinding efficiency of the ball mills is
effected. Hence, it is proposed and tried to produce granules
180
out of these two materials, that can serve the cement industry
with a single product of two important ingredients.
THE PRODUCT :
Dihydrate gypsum, CasO4.2H20, a non-cementitious material,
attains its cementitious properties on calcination, yielding four
grades of calcined gypsum. depending- on the calcination temperature.
Out of the four grades, hemihydrate is most unstable form, which
on hydration, sets within minutes and converts to dihydrate gypsum.
at 4% on a typical sample- of pozzolana cement. As per this, the
proportion of flyash to dihydrate gypsum has been resulted as 85 :15.
Taking the weight of hydration for hemihydrate into consideration,
it is computed to add 12.6 parts of hemihydrate to achieve 15 parts
of dihydrate
Another point of importance is the cost aspect of the product,
in view of the involvement of calcined gypsum. The success of
the product lies in its economic feasibility . Hence, d i f f e r e n t
combinations were tried and studied, replacing hemihydrate partially
with dihydrate gypsum and the results are given in Table II
The amount of flyash to be admixed with clinker depends on
the quality of clinker also, apart from the suitability of flyash.
Hence, the same granular product can not be accepted by all the
cement plants. Hence, another proportion with high flyash and
low gypsum. content was also analysed and the results obtained
are given in Table III,
CHEMISTRY OF GRANULATION :
In view of the low content of CaO available at 1.5 to 2.5%
normally, there was no opportunity for the mineralogical constituents
to form as in that of OPC. In other words to say, flyash is an
182
inert material in the absence of lime.
When the granules were produced, it was expected of some stoichi-
ometry to take place between gypsum and aluminates of flyash but
nothing of that sort was observed, as there was no evidence of
gain in strength attributable to such stoichiome tric reactions.
This is to say that, while the flyash in association with refined
Dihydrate gypsum has not at all shown any strength development,
the granules produced with 6% and 12% hemihydrate have resulted
with 0.15 kg and 0.45 kg green strengths respectively per cm dia
of granule, thereby giving credence to granule strength to the
setting behaviour and hardness of hemihydrate only than any other
factor.
However, all the above cited three types of granules, after
drying, while dropped into water have bursted as bubbles, ultimately
becoming soft and incoherent. This could be correlated with the
rapid formation of ettringite, in the abundant availability of
water, leading to the abnormal expansions within the mass. However,
under atmospheric conditions, the formation of ettringite can not
be ruled out by absorption of atmospheric humidity. But such ettrin-
gite undergoes further stoichiometric reactions, on addition of
granules with clinker, during hydration, to contribute to the cementi-
tious proper ties.
The hemihydrate renders cohesive bonding to flyash particles
during hydration, contributing for the formation of granules and
their strength, in the process of which it converts to dihydrate
183
to play its further role as retarder, on addition with clinker,
in the PPC.
1.
2.
3.
4.
5.
CONCLUSIONS
The at tempt to add dihydrate gypsum also, as a partial replace-
ment to the required portion of calcined gypsum, could not
be fulfilled because of the weak strengths resulted.
In view of the confinement on the proportion of hemihydrate,
to suit its quantity as retarder in the ultimate cement, it
is constrained to add not more than 12.6 parts, to result
in the commensurate moderate strength. It is desirable to
increase the strength by other means and deeds to befit the
granules for transportation, to longer distances, or for storage
in higher stacks.
For this purpose, taking the advantage of thixotropic behaviour
being rendered by flyash-gypsum paste, it can be explored
to minimise the water content, thereby subjecting the denser
mass to extrusion, in order to produce a more strengtheir
cylindrical granules.
Because of the danger for the granules to disintegrate through
rapid formation of ettringite on exposure to water, i t i s
advisable for the product to be stored under cover.
Though PPC is a more virtuous cement over OPC, the handling
and transportation of flyash stood as a major stumbling-block
in the rapid production of PPC and its popularisation. With
184
the solution offered by the authors, commercial production
of flyash granules can be positively resorted to, in order
to solve the need of granulating phosphogypsum also, to achieve
the procurement of two ingredients in one product for its
attractive techno-economic factors of feasibility.
6. With the growing awareness in environmental protection, the
use of both flyash and phosphogypsum can be made extensive,
only when the problems caused in their handling are put to
an end. Both handling and dosing become easy, in granular
form, encouraging the higher rate of PPC producti.on.
ACKNOWLEDGEMENTS
It is the Salzgitter's granulation technology to manufacture
granular retarder out of phosphogypsum that has inspired the authors
to conceive this work. The authors owe their allegiance to M/s
Salzgitter Industribau GmbH, West Germany, and their Head of Gypsum
Division, Mr. Gunther Erlenstaedt, for the time to time guidance.
185
RECOMMENDED PROCEDURE
FOR
SAMPLE PREPARATION AND TESTING STABILIZED GYPSUM MIXTURES
Donald Saylak, P.E. Ramzi TahaProfessor-Materials Graduate Research
Science Assistant
Dallas N. LittleProfessor-Materials
Engineering
Civil Engineering DepartmentTexas A&M UniversityCollege Station, Texas
ABSTRACT
Because of the current increase in research activity associated with
structural integrity assessments of stabilized gypsum mixtures, it has
become imperative that a unified set of specimen preparation and testing
procedures be established. Existing standard test procedures which were
developed for either natural gypsum or stabilized soils do not neces-
sarily apply to by-product gypsum systems. The literature shows that a
wide range of strengths can be generated as sample configurations and
compaction modes vary. Since these strengths are ultimately used to
design the thickness of a pavement layer or supporting structure it
follows that the values passed on to and used by the designer must be
obtained under test conditions which best simulate those the material
will experience during construction and service.
This paper presents a laboratory procedure for the determination of
engineering properties which have been used to qualify fly ash and cement
stabilized gypsum mixtures for road construction in Texas. In its
development, problems associated with the selection of sample size, pH,
optimum moisture content, type and level of compactive effort, storage
conditions and test methods will be discussed. A set of recommended test
and materials selection specifications are also provided. Comparison of
laboratory-generated results with those obtained during the construction
of one job site in Texas will also be presented.
187
BACKGROUND
Research activity at Texas A&M University began in 1982 (1,2) with an
attempt to exploit the phosphogypsum stockpiles at the Mobil Chemical
Company's Pasadena, Texas plant for use in road base construction. Two
significant findings were reported:
(1) The strength achieved following stabilization using either
a high-lime fly ash or portland cement was considerably
higher in a 7 year old, inactive pile (Pile 2) than in one
that was currently active (Pile 3) - see Figure 1.
(2) At that time, the factor which appeared to be the primary
influence on strength was stockpile acidity as reflected
by PH. The active pile (pH ~ 2.5) developed little or no
strength whereas the aged pile (pH ~ 5.5) showed a higher
degree of stabilization.
(3) The particle size distribution and shape of the material
in the older pile was more varied than that found in Pile
3. These features lend themselves to achieving higher
densities and stability through better compaction.
A new study was initiated under the sponsorship of the Bureau of
Mines to verify the influence of pH on strength (3,4). The acid level
was varied by neutralization and washing with water. The results of the
washing portion of the study are shown in Figure 2. In an attempt to
artificially raise the pH of Pile 3 gypsum to that of Pile 2 only about
50 percent of the anticipated strength increase was realized. It was
later shown that improved compaction capability provided by the more
dense size distribution in the older pile would account for the remaining
strength.
Mobil's phosphogypsum taken from Pile 2 was stabilized with cement
and a high-lime (Class C) fly ash and used to construct a series of
asphalt-covered road bases for the City of LaPorte, Texas (5). These
streets are still in operation at this writing and at least 400 more
projects using cement-stabilized phosphogypsum have been constructed in
188
189
the Houston, Texas area. Another project on Texas State Highway 146 is
currently under construction with two more planned for the 1987-88
construction season.
More recently a new form of dihydrate gypsum has begun to be
inventoried in Texas. This material is produced from sulfur recovery
operations mandated for power plants burning lignite or sulfur coals.
Carrying the designation "synthetic gypsum", this material is currently
being produced by 18 plants in Texas at a rate of 1 million tons per year
with 9 new plants expected to become operational by 1993.
Studies now under way at Texas A&M (6) using synthetic gypsum pro-
duced by Texas Utilities Generating Company (TUGCO) at its Martin Lake,
Texas plant show a pronounced similarity to the Mobil material except the
pH and particle size distributions, as produced, are already satisfactory
for strength development. One of the differences between the as-produced
(new) and aged (old) synthetic gypsum is reflected in their CaSO3contents - see Table 1. The conversion of CaSO3 to CaSO4 requires only a
relatively short time and therefore does not present any significant
aging factor. Figure 3 shows a comparison of 7-day compressive strength
between a new and aged gypsum. The difference in age of the gypsum used
in the two mixtures was about 2 weeks. Additional natural aging showed
no further decrease in strength.
190
Figure 3. 7-Day strength for unaged and aged cement stabilized gypsumfrom Texas Utilities Generating Company stockpiles.
191
An increasing number of applications for stabilized, by-product
gypsum mixtures appear to be finding their way into the public sector.
This suggests an immediate need for standardizing the methodology for
assessing the performance characteristics of these mixtures. The dis-
cussions which follow are directed primarily, to the dihydrate form of
by-product or synthetic gypsum. However, except for the manner in which
it is stabilized, much of the recommended procedures could also be applied
to the hemihydrate form of gypsum as well.
RECOMMENDED PROCEDURES
Material Selection
Gypsum: Three primary factors required for preliminary screening of
stockpiled gypsum are pH, particle size distribution and moisture con-
tent. For good stabilization, only gypsums with pH values above 5.0 are
recommended. Particle gradations should reflect a distribution of sizes
free from agglomerates capable of being retained on the No. 4 sieve.
In the determination of moisture content the temperature at which the
uncombined water is removed is critical and should not exceed 104OF
(4OOC). This is lower than 113oF (45oC) as suggested for gypsum con-
cretes in ASTM C 472 and 140°F (6OoC) as suggested in ASTM D 1557 for
determining moisture-density relationships. Due to the inherent acidity
of phosphogypsum, the higher temperatures facilitate decomposition to the
hemihydrate. When heated above 212OF (100°C) the gypsum particles take
on the appearance of a delaminated ash - see Figure 4.
Stabilizer: Until recently the selection of the portland cement did
not require any restrictive specifications. Deussner (7) showed that
tricalcium aluminate [C3A] in the presence of large amounts of sulfate
promotes the generation of Ettringite crystals which can lead to expan-
sive deformations and cracking in cement stabilized materials. Papers
presented at this symposium suggest that cements with a 7 to 8 percent
192
193
maximum allowable [C3A] content should be used for gypsum stabilizations.
Type II (i.e., sulfate resistant) cements can also be used if the slower
rate of strength development can be accommodated. On the other hand Type
III (i.e., high early strength) cements can also be used, when cure rate
is of concern, provided the [C3A] content is not excessive.
When fly ash is used for stabilization, the strengths developed in
the stabilized mixtures will be directly related to the CaO content. So-
called "high-lime" fly ashes have CaO contents in excess of 16 percent.
Gypsum stockpiles which maintain high levels of acidity may require
washing or chemical neutralization to achieve pH > 5. Washing, as men-
tioned above, can remove a considerable amount of fines and produce a
stabilized mixture lower in strength. Fly ash particles can, in addition
to acting as a pozzolanic stabilizer, replenish the fines lost by washing
and further enhance strength development (4).
Sands: Sometimes there is a need to further improve the strength in
stabilized gypsum mixtures to meet allowable construction specifications.
This can be accomplished by adding a graded field sand or concrete sand,
The addition of sand tends to reduce the optimum moisture content.
Consequently, higher dry densities and higher compressive strengths may
be obtained with lower amounts of cement. Tests run on two local Texas
sands indicated that strength was improved by using the coarser sand (6).
Testing
This section will be devoted to sample selection, preparation,
storage and testing of stabilized gypsum mixtures. Primary emphasis was
given to the use of existing standard test methods, where possible. An
attempt is also made to caution against the use of ambiguous terminology
in order to set the stage for the development of consistent, generally
acceptable procedures for use in evaluating stabilized by-product and
synthetic gypsum mixtures.
194
Sample Selection: There have been a wide variety of specimen geo-
metries employed in assessing cement-stabilized materials (8). They
range from cylinders of varying L/D ratios and cubes. Structurally, the
most desirable configuration is a cylinder with L/D ratio of 2. In
situations where other ratios are used correction factors have been
developed (see Figure 5) so that reported strengths will reflect the con-
ditions of the preferred sample geometry (9). Many have utilized a 2-in
diameter by 4-in long design (10), a 2.8 diameter x 5.6-in long specimen
or the 4-in diameter x 4584-in long specimens as set forth in ASTM D 1633
(see Table 2). A correction factor of 1.10 is used with the data taken
on the 4-in specimens to convert to a strength for a specimen with an L/D
of 2 (11c). It should be noted that the 4-in diameter sample is the
predominant configuration for all tests except that used by the Texas
Highway Department (11i).
Method of Compaction: Were it not for the direct influence that
degree and method of compaction have on compressive strength it would be
expedient to simply ignore sample configuration as a test parameter and
rely solely on correction factors to convert the test results. It has
been well established that any mix design procedure which effects the
compactive energy delivered to a specimen will inevitably have a signi-
ficant effect on its cured strength. Investigators have shown that a
direct and near linear relationship appears to exist between energy of
compaction and compressive strength (10). It may require significantly
more stabilizer for a mixture compacted by Standard Proctor (llg) to
achieve the same strength of one compacted using the Modified Proctor
method (lla). Hence, the amount of energy which can be delivered in the
field should ultimately dictate the level of compaction which should be
used to prepare samples in the laboratory.
195
Figure 5. Curves for obtaining correction factors for converting cubes and cylinders to strengths in4-in cubes and 6-in diameters by 12-in long (L/D = 2) cylinders (Ref 9).
At the same time there exist a considerable amount of ambiguity
associated with specifying compactive effort. For example, the use of
the term "Modified" Proctor has frequently been used to denote that 32.6
ft-lb/in3 of energy was delivered to a specimen regardless of the speci-
men's geometry and the manner in which this energy was induced. This can
be grossly misleading. In a general sense, any compaction procedure
which departs from ASTM D 698, which is commonly referred to as
"Standard" Proctor can be considered "Modified". Several states such as
Texas have adopted such a departure by requiring only 13.5 ft-lb/in3
(11i). In reality, Modified Proctor compaction, as defined (although not
specifically so stated) by ASTM D 1557, not only sets forth specific
specimen configurations, but also how they should be densified.
In general, ASTM D 1557 (Method A) calls for a 4 diameter specimen
4.584-in long, filled with 5 layers of stabilized material, with each
layer receiving 25 free-fall blows from a 10 lb weight falling from a
height of 18-inches. The impactor is a 2-in diameter plate requiring
that the specimen be rotated to distribute the load during compaction.
This rotation sets up a kneading action which effects the distribution of
the particles in the mixture. The use of a 2-in diameter specimen with a
2-in diameter compaction plate will not induce this kneading action
because the same column of material will recieve the direct impact of all
the blows delivered.
The major consideration in any mix design or quality assurance pro-
cedure is that the conditions under which the materials were prepared and
evaluated in the laboratory be consistent with the capability of modern-
day construction equipment. It would be inappropriate to use a design
compressive strength obtained under Modified Proctor compaction condi-
tions if that degree of densification were not achievable in the field.
Table 3 shows a comparison of laboratory densities with those obtained
during the construction of the LaPorte, Texas project in 1983 (5).
Except for the 5 percent cement and 25 percent fly ash mixtures one
should expect to achieve field densities at least 95 percent of those
obtained by ASTM D 1557. It should be noted that no attempt was made on
this project to maximize field densities.
ASTM D 1632 also calls for the method of "making and curing"
3x3x111/4-in. flexural test specimens. Compaction of these specimens
should produce the same densities achieved in the samples for compression
testing. It is recommended that flexure specimens be compacted in one
inch layers. These specimens can also be used for flexural fatigue as
well as static tests.
198
Compaction in the field is normally achieved by sheepsfoot, steel
drum or pneumatic rollers each of which create a kneading action. There-
fore, ASTM D 1557 appears the best method for simulating this action.
The authors have performed some exploratory tests using a standard Hveem
kneading compacter similar to that used for asphalt concretes (12). The
results of these tests have shown that a wide range of compacted densi-
ties and strengths are achievable with little or no apparent damage to
the individual gypsum particles as can be induced by excessive impact
blows. Perhaps better simulation of field compaction in the laboratory
should be an area for future study.
Storage During Cure: The method of curing specimens prior to testing
can have a pronounced effect on their compressive and tensile strength.
Therefore, it is important that the curing method represents a conser-
vative simulation of the in-situ environment that the mixture will exper-
ience during its service life. Lin (10) investigated this influence
using three curing methods: Air drying, sealing and soaking.
1) Air-Dry: indicating that the specimen had been removed
after 7-days in the curing environment (sealed or 100
percent humidity room) and maintained in open-shelf,
laboratory conditions for 2 days. The resultant moisture
content of all specimens was below 3 percent.
199
2) Sealed: indicating that the specimen was tested with the
moisture content resulting from sealed curing. Sealed
curing was obtained by tightly wrapping the specimen with
a plastic membrane after removal from the mold. The
28-day moisture content was approximately 2.5 percent
lower than the moisture content of the mix at the time of
compaction.
3) Soaked: indicating that the specimen was removed from the
curing environment (sealed or 100 percent humidity room) ,
and submerged in water for a 2-day period.
The resulting 28-day compressive strengths for each of the three
methods is shown in Figure 6. Although the air-dry condition permits the
greatest strength to be achieved it does not reflect the mixture's
moisture susceptibility. This is addressed by test methods such as
ASTM D 1633 which calls for curing in a moist room (75oF and 100 percent
humidity). The authors found that such exposure can have an adverse
affect on sample integrity during the first 3 days of exposure when
curing is beginning to develop. Where such exposure may actually benefit
hydration in cement-treated soils there is a definite irreparable damage
incurred in cement-gypsum systems which can in-turn compromise the
results generated in the strength tests. Sealed bags still provide a 100
percent humidity enclosure without subjecting the sample abnormally high,
direct moisture content.
Texas Method 117-E (11h) subjects soils and base materials to 10 days
of capillary absorption by storing specimens over porous stones partially
submerged in a water bath. This procedure usually produces a reduction
in strength by revealing the moisture susceptibility of the gypsum filled
mixture (see Figure 7). Although this test may tend to simulate water
intrusion to the base from the surface or through the subgrade it tends
to be overly conservative. On the other hand, this type of storage may
be of value in assessing the relative effects of additives designed to
reduce moisture susceptibility. This is an area which requires addi-
tional study. For the present, sealing in plastic bags appears to offer
the best condition for storage during cure.
200
Unconfined Compressive Test: The primary test used for assessing
strength in cement-stabilized materials is the unconfined compression
test, ASTM D 1633. This method prescribes either a 4-in diameter by
4.584-in specimen - Method A (L/D = 1.15) or a 2.8-in diameter by 5.6-in
specimen - Method B (L/D = 2). For the latter, the specimens are com-
pacted by rodding followed by static compaction. Obviously, this is
inconsistent with Proctor compaction methodology, as discussed earlier.
Therefore, it is recommended that the compression strength test procedure
set forth in ASTM D 1633 be used but that sample size and compaction
procedures in ASTM D 1557 be utilized in sample preparation. This
requires that the resulting data-be converted to reflect an L/D of 2
through pre-established correction factors.
Design and Quality Control Criteria
Field experience has clearly demonstrated that stabilized gypsum mix-
tures, if properly designed, can successfully perform the primary mission
of a road base (i.e., to protect the vulnerable and variable subgrade
from overstressing). Subgrade overstressing is the primary cause of most
major road base failures which eventually leads to the failure of the
surface course. To perform successfully, the base must be of acceptable
quality and structural design. Conditions for acceptability can vary
from region to region depending on the environmental factors and traffic
loads which dictate the design.
Durability: Wetting-Drying and Freeze-Thaw durability tests are
specified by ASTM D 599 and D 560, respectively. These tests also use
the 4-in diameter x 4.584-in long specimen. In addition to providing a
means for classifying soils, these tests also have been used to establish
allowable unconfined compression strengths for design purposes. Figure 8
vides a 95 percent sample surv
specified by these two methods
tests would be considered over
design strength.
shows that a soil-cement mixture with a 650 psi compressive strength pro-
ivability under the severe environments
(13). For many geographic areas these
ly conservative as a basis for specifying a
202
Figure 8. Compressive strength vs durability using wet-dry(z) andfreeze-thaw(11f) tests.
The acceptable strength requirements in a state like Florida which
experiences little or no freezing would be expected to be different than
those required in northern states. In states like Texas both conditions
must be considered. Preliminary studies at Texas A&M (14) have shown
that when freezing and thawing is not a factor the strength required
could be as much as 50 percent below the 650 psi normally targeted by the
Texas Highway Department. If true, a considerable reduction in stabi-
lizer content and cost can be realized.
203
Flexure and Fatigue: Preliminary Texas A&M research has demonstrated
how shear strength in the pavement, o APP, flexural strength, o FLEX, and
flexural fatigue life, Nf, can be used to provide reliable acceptance
criteria for the design of stabilized gypsum bases. In this study,
stress-amplitude, flexure fatigue tests were used to develop a relation-
ship between unconfined compression, fatigue strength and resistance to
low temperature cracking. The results involving a few mix design are
shown in Figure 9a and b. It can be seen that when the fatigue data are
plotted as Applied Stress vs Nf (Figure 9a) a curve exists for each mix
design. However, when plotted as the ratio o APP/ o FLEX as in Figure 9b
all data can be represented by one curve. This ratio equals 0.5 at
Nf = 105 cycles. Furthermore, it has been shown that the flexural
strength is approximately 0.2 times the unconfined compression strength,
UCS,(14,15).
0.5 = o APP = o APPo FLEX 0.2 UCS
orUCS = 10 o APP
This indicates that the compressive strength is of the order of 10
times the flexural stress to cause failure in the road base. In a well
designed pavement system this flexural stress can be between 30 and
40 psi. On this basis, the design strength would be around 350 psi
instead of 650 psi which is normally required. A demonstrated test
fatigue life in excess of 105 cycles would be satisfactory for roads with
low traffic volume whereas 108 cycles would suffice for high traffic
volumes. The latter would also demonstrate good resistance to low tem-
perature as well as freeze-thaw cracking.
In summary the recommended test methods and specification limits for
use of stabilized gypsum mixtures in road construction are shown in
Tables 4 and 5, respectively. It should be stated that this is only a
starting point. As more data are made available from the field these
suggestions may require alteration or the development of new procedures.
204
Figure 9. Flexural fatigue curves for various cement stabilizedgypsum mixtures (a) applied stress vs Nf and (b) stressratio vs Nf (14).
205
206
CONCLUSIONS
A summary of procedures for material selection, sample preparation,
curing and testing have been presented to provide design and quality con-
trol criteria for stabilized-gypsum mixtures. An attempt was made to
deal with inconsistencies and ambiguities in reporting test results.
Recommendations were made for establishing separate design strengths for
geographic locations concerned with freezing and thawing as well as those
which are not. It is hoped that the recommendations proposed in this
paper could lay the ground work for ASTM standardization of test method-
ology for use with all industrially generated gypsums.
REFERENCES
1. Gregory, C. A., "Enhancement of Phosphogypsum with High Lime FlyAsh," Master's Thesis, Texas A&M University, May 1983.
2. Saylak, D., Gregory, C. A. and Ledbetter, W. B., "LaboratoryInvestigations of Phosphogypsum from Mobil Pile 2", Final Report forthe Mobil Chemical Company, Pasadena, Texas, August 1985.
3. Saylak, D. and Gadalla, A. M. M., "Strength Development in By ProductPhosphogypsum for Road Bases and Subbases", Final Report on Grant No.GO145052 for the Bureau of Mines, Tuscaloosa Research Laboratory bythe Texas Transportation Institute, Texas A&M University, October1985.
4. Saylak, D., Gadalla, A. M. M. and Yung, C., "Neutralization andStabilization of Phosphogypsum for Road Construction", Proceedings ofthe 3rd Workshop on By-Products of Phosphate Industries, Tampa,Florida, pp. 315-338, November 1986.
5. Gregory, D., Saylak, D. and Ledbetter, W. B., "The Use of By-ProductPhosphogypsum for Road Bases and Subbases", presented at theTransportation Research Board Meeting, Washington D.C., January 1984.
6. , Project No. 1690 to the Texas A&M Development Foundationfrom Texas Utilities Generating Company (TUGCO), 1985.
7. Deussner, M., Neinz, D. and Ludwig, V., "Mechanism of SubsequentEttringite Formation in Mortars and Concretes after Heat Treatment,"American Ceramic Society 87th Annual Meeting Proceedings, Cincinnati,Ohio, May 1985.
207
8. Symons, L. F., "The Effect of Size and Shape of Specimen Upon theUnconfined Compressive Strength of Cement - Stabilized Materials",Magazine of Concrete Research, Volume 22, No. 70, March 1970.
9. , Concrete Manual, U.S. Bureau of Reclamation, pp. 574-575,
10. Lin, Kuo-Ting, Figueroa, J. L. and Chang, W. F., "EngineeringProperties of Phosphogypsum", Proceedings of the 2nd Workshop onBy-Products of Phosphate Industries, Miami, Florida, p. 55, May 1985.
11. ASTM and Other Test Methods Cited(a) ASTM D 1557: "Test for Moisture-Density Relations Soils and
Soil Aggregate Mixtures Using 10-lb (4.5 Kg) Rammer and 18-in(457-mm) Drops".
(b) ASTM D 1632: "Making and Curing Soil-Cement Compression andFlexure Test Specimens in the Laboratory".
(c) ASTM D 1633: "Compressive Strength of Molded Soil-CementCylinders".
(d.) ASTM D 558: "Test for Moisture-Density Relations of Soil-CementMixtures".
(e) ASTM D 559: "Wetting and Drying Tests of Compacted Soil-CementMixtures".
(f) ASTM D 560: "Freezing and Thawing Tests of CompactedSoil-Cement Mixtures".
(g) ASTM D 698: "Moisture-Density Relations of Soils andSoil-Aggregate Mixtures Using 5.5 lb Rammer and 12-in Drop".
(h) "Determination of Moisture-Density Relations of Soils and BaseMaterials", Texas State Department of Highways and PublicTransportation, Method TEX 113-E-SDHPT Testing Procedures-Volume 1.
(i) "Triaxial Compression Tests for Disturbed Soils and BaseMaterials", Texas State Department of Highways and PublicTransportation, Method TEX 117-E-SDHPT Testing Procedures-Volume 1.
12. ASTM D 1560, "Resistance to Deformation and Cohesion of BituminousMixtures by Means of Hveem Apparatus.
13. Soil Cement Laboratory Handbook, Chapter 5, "Esta-blishment of Cement Factors for Construction", Portland CementAssociation.
14. Little, D., "Characterization of Performance Criteria ofCement-Stabilized Phosphogypsum, presented at 2nd InternationalSymposium on Phosphogypsum (unpublished), December 1985.
15. Lin, K. T., Nanni, A. and Chang, W. F., "Engineering Properties ofDihydrate Phosphogypsum, Portland Cement and Fine AggregateMixtures", Proceedings of the 3rd Workshop on By-Products ofPhosphate Industries, Tampa, Florida, December 1985.
208
OIL ASH STABILIZATION USING PHOSPHOGYPSUM
E. H. Kalajian, Ph.D., P.E. R. J. Boueri, B.S., M.S.
Professor and Department Head
Department of Civil Engineering
Florida Institute of Technology
Melbourne, Florida 32901
Civil Engineer
John Zimmerman Associates
Boca Raton
Florida 33432
INTRODUCTION AND OBJECTIVE
Oil ash waste is the solid residue remaining after the combustion
of oil at electric power utilities. Present techniques for disposal
of oil ash in solid land fills are of environmental and economic
concern. Duedall and Kalajian have recommended the stabilization of
oil ash into blocks for use in artificial reef construction. The
stabilized oil ash blocks must have a design strength of 300 psi to
remain intact during fabrication, handling, curing, transport to the
site and reef construction. The blocks once placed must remain
chemically stable over the life of the reef and be environmentally
acceptable.
Oil ash stabilized with the addition of coal ash, lime, cement,
and sodium carbonate has been recommended for reef construction based
on engineering properties and seawater tank leaching studies (Duedall
and Kalajian). The recommended mix design is expensive and
alternatives to the use of lime and fly ash are sought. Phosphogypsum
is a material which has properties which make it a less expensive
replacement for the stabilization of oil ash.
209
The objective of the research investigation was to stabilize oil
ash for offshore artificial reef construction using mixtures of
phosphogypsum and Portland cement.
SCOPE
The oil ash was obtained from the Cape Canaveral Plant of Florida
Power & Light Company. The phosphogypsum was collected from the
stacks and from the production line of Occidental Chemical Company,
White Springs, F1. Preliminary mix designs using phosphogypsum, oil
ash and cement yielded compressive strengths of 793 psi after 18 days
curing at ambient temperatures for a mix of 5 parts oil ash, 5 parts
phosphogypsum and 2 parts Portland Type II cement.
A testing program was defined to evaluate the performance of
phosphogypsum stabilized oil ash. The engineering properties of
phosphogypsum stabilized oil ash using off-stack and off-line
phosphogypsum were evaluated and compared to a mix design using fly
ash and lime. Stabilized oil ash samples were fabricated using two
different methods. A kneading compaction technique using the Harvard
Miniature apparatus was used with a modified mold (1.4 in. diameter and
a 2.8 in. length). The second fabrication method used a vibrator
operating at a frequency of 3200 rpm with the same size sample. A
superplasticizer was added to increase the flow characteristics of the
mix during fabrication and to minimize voids. Mix designs are
summarized in Table 1 with the following legend: (C) compaction, (V)
vibration, (1) lime-fly ash, (2) off-stack, (3) off-line. A
population of 52 samples of the ratio (5:5:2) was used for each mix
design.
210
The unconfined compressive strength of the phosphogypsum
stabilized oil ash samples was determined at time periods up to 108
days in a cured, air dry and seawater saturated state to simulate
artificial reef placement. Density and porosity were determined for
the seawater saturated samples.
RESULTS
The physical and engineering properties of oil ash and
phosphogypsum are shown in Table 2. The specific gravity of the
phosphogypsum compares with Florida DOT results for dihydrate
phosphogypsum. The maximum dry unit weight of the off-line material
is low compared to the off-stack material. The process at White
Springs produces hemihydrate phosphogypsum waste which converts to
dihydrate phosphogypsum on the stack. The percentages of dihydrate in
the samples was not determined.
The unconfined compressive strength of the cured, air dry and
saturated samples are presented in Table 3. All mixes exceed the
design strength of 300 psi after 28 days.
For the cured samples, phosphogypsum mixes (C2, C3, V2, V3)
showed slightly higher strengths than the lime-fly ash (VI, Cl) at 28
days. After 108 days the samples were very comparable in strength.
Vibrated phosphogypsum samples had higher 28 day strengths than the
kneading compacted samples which may be due to the superplasticizer.
All mixes substantially increased in strength with air drying
with the phosphogypsum samples having the greatest strength increases.
Chang et al. report similar increases for sand phosphogypsum mixtures,
due to a moisture decrease.
212
Samples saturated in seawater showed strength increases with time
of saturation. Vibration compaction mixtures yielded slightly higher
strength increases with time than kneading compacted mixtures.The
stabilized phosphogypsum using off-stack material showed higher
strength after 80 days of saturation when compared to off-line
material or lime fly ash stabilized mixes.
Dry unit weight and porosity of the mixes are presented in Table
4. The dry unit weight did not significantly change after 80 days in
seawater suggesting no leaching of calcium from the samples.
CONCLUSIONS
Oil ash can be stabilized with the addition of off-stack
phosphogypsum and cement to provide a material which can be placed as
an artificial offshore reef.The compressive strength of stabilized
oil ash can be as high as 600 psi after 28 days of curing and
continues to gain strength with time in either an air dry or saturated
environment. The use of vibration compaction with a superplasticizer
is recommended to achieve strength increases as compared to a kneading
compaction technique.Chemical and biological evaluation of this
material will need to be conducted prior to reef placement.
REFERENCES
Chang, W.F.,Figueroa, J.L. and Velarde, H.D., "Engineering Properties ofPhosphogypsum, Fly Ash, and Lime Mixtures,"Proceedings of the SecondWorkshop on By-Products of Phosphate Industries, FIPR, Bartow, FL,1985.
Duedal1, I.W. and Kalajian, E,H."Stabilization of Oil Ash Sludge forArtificial Reef Construction," F.P.& L Contract Report, Sept.1985.
213
COMPACTION CONCRETE UTILIZING DIHYDRATE PHOSPHOGYPSUM
Nader Ghafoori, Assistant ProfessorDepartment of Civil Engineering
Saint Martins College, Olympia, WA
Wen F. Chang, Professor and DirectorPhosphate Research Institute
Department of Civil & Architectural EngineeringUniversity of Miami, Coral Gables, FL
ABSTRACT
The objective of this paper is to present an in-depthinvestigation on the basic engineering properties Ofphosphogypsum-based concrete mixtures. In particular, theattention is given to strength characteristics and its practicalapplication in the form of roller compaction concrete.
Data presentation and discussion focus on:
a)
b)
c)
d)
relationship between mix proportions and strengthcharacteristics when varying environmental testingconditions, as well as specimen age and sizes;
relationship between compressive, splitting-tensile andbending-tensile strength under similar testingconditions;
presentation of field compaction pavement slabs,including a full scale phosphogypsum-based concreteramp and the comparison of field and laboratorysamples; and
comparison of different consolidation methods on thebasis of compressive strength criterion.
The strength results obtained from laboratory and fieldsamples show that dihydrate phosphogypsum mixtures can achieveexcellent compressive strength under compaction. The laboratorycompressive strength approaching 4,600 psi becomes accessible forspecimen containing 7.5% cement under air-dry conditions. Thisstrength considered sufficient for application such as roadconstruction. Furthermore, the results indicate that strengthproperties of concrete mixtures containing phosphogypsum are notonly influenced by the matrix constituents and proportions, butCO a great extent, by the selected consolidation technique.Both impact and static consolidation provide better hardened mixproperties.
215
1. INTRODUCTION
Extensive research over past few years has successfullydisclosed the feasibility of using phosphogypsum, by-product ofwet phosphoric acid industry, as a viable aggregate in cement-based mixtures. Under laboratory compaction, both plain andcement-based phosphogypsum mixes have shown higher compressivestrength than the similar specimens consolidated usingconventional vibration placement technique. The improvement of
strength under compaction consolidation is attributed to thephosphogypsum's nature and its extremely fine gradation, and alsoas a consequence of the reduced amount of mixing water requiredby compaction method.
The scope of this paper is to demonstrate that concretemixes containing dihydrate phosphogypsum possess good, andacceptable engineering properties and it can be superior toconventional concrete when appropriate proportioning andconsolidation technique are used. The advantages are sizeableparticularly for the mixtures of low cement content.
The core of experimental work presented in this paperconsists of data collected from the complete range ofphosphogypsum-based concrete mixtures that were prepared at their
corresponding optimum moisture content, compacted in accordancewith the Modified Procter method, and tested under sealed, air-dried or soaked testing conditions. Engineering parameters suchas compressive strength, splitting-tensile strength, bendingtensile strength, modulus of elasticity and relationship amongthem were investigated.
In addition the effects of factors, such as environmentaltesting conditions, specimen size curing time and curing types,on strength properties were also determined.. Finally, thestrength results were compared with data obtained from similarmixtures consolidated. by high - frequency vibration, staticcompaction and on-site roller compactors.
2. MATERIALS
The material used for the matrix preparation consisted ofPortland Cement Type I (ASTM.C-150), dihydrate phosphogypsum(specific gravity = 2.396), finely crushed limerock (finenessmodulus of 2.6 and specific gravity of 2.7) and limerock peagravel with a maximum nominal size of 3/8 inches (specificgravity of 2.69). More description of the composition of thesematerials can be found in reference 1.
216
3. SPECIMEN PREPARATION
The dry component of the matrix, in order of cement,limerock sand and dihydrate phosphogypsum were first mixed in aconventional pan-type mixer for a period of ten minutes; then theappropriate amount of water was added. Finally the adequateamount of pea gravel was placed in the mixer until a homogeniousmix was obtained.
Two types of specimens, cylinder (2 x 4 inches) and beam(2 x 2 x 7 inches) were used during this investigation. Specimenswere compacted, using an impact compacter, in accordance with theModified Proctor compaction method (ASTM D-1557, accumulatedenergy of approximately 56,500 foot pound per cubic foot), unlessotherwise stated.
After consolidation, the samples were wrapped in a plasticsheet and cured at room temperature (77-F, R. II. 60%) for thedifferent curing period. Specimens were then tested under sealed,soaked, or air-dried conditions.
4. EXPERIMENTAL RESULTS
4.1 Compressive strength:
4.1.1 Influence of Mix Proportion: One of the majorobjective of this study was to investigate the compressionbehavior of the compacted phosphogypsum concrete mixtures whenvarying the mix proportions. Figures 1 and 2 present the 28-daycompressive strength as a function of sand content for themixtures of 20 and 40% gravel content, respectively. Bothfigures contain five groups of Portland cement Type I varyingfrom 5 to 25% by dry weight. From these test results, severalobservations are noticed:
i) All compacted mixtures have shown considerablecompressive strengths considering the percentage levelof cement in the mixture. Sealed compressivestrengths above 3,000 and 4,500 psi can readily beobtained by using, respectively, 7.5 and 10% cementcontent.
ii) The compressive strength greatly depends on the mixproportions and the cement content in the mixtures.From results obtained it can be seen that, as thecement content increases the compressive strengthincreases. It is also noticed, that whenphosphogypsum is replaced by sand, the compressive
217
strength continuously increases for thehaving
mixturescement content of 15% or more,
value iswhereas a peak
reached for the cement content of 10% orless. These peak curves reflect thecontribution
strengthof dihydrate phosphogypsum to low cement
content mixtures. Furthermore for the same low cementcontent mixes, the limited amount ofphosphogypsum
use of onlyis beneficial for strength
characteristic of phosphogypsum-based concrete.
4.1.2 Effect of Testing Conditions: Past research studies haveclearly suggested that the free moisture held in concrete at thetime of testing has an adverse effect on its strengthKeeping
behavior.this in mind,
mixturesthe compressive strength of concrete
containing dihydrate phosphogypsum under sealed; air-dried and soaked conditions were investigated.
Test results, as shown in Figure 3, indicate that thephosphogypsum-basedstrength in
concrete mixtures exhibit higher compressiveair-dry as opposed to soaked
Thistesting condition.
improvement of strength after drying, thanthe one reported for conventional mixes.
is much higherMoreover test results
indicate the sensitivity of phosphogypsum concretemoisture at time of testing,
mixture, to
-content of the mix increases.reduces as the percentage of cementNo significant changes in strength
were noticed when testing conditions changed from sealed to thesoaked.
4.1.3. Effect of Curing Test resultsobtained
Age and Curingfrom
Type:strength versus time diagram, as illustrated in
Figure 4, indicate that the concrete mixturesdihydrate phosphogypsum
containingcontinues to gain strength with curing
age, if proper and steadily curing conditions are provided. Thistrend follows the general pattern of conventional concrete. The?-day to 28-day compressive strength ratio ranges from 0,549 to0.732 and 0.561 to . 812 for the mixtures of 20 and 40%content, respectively.
gravelThese values are practically identical to
the one valid for conventional concrete.
Figures 5 and 6 present the soaked compressive strength asa function conditions(sealed
of sand content for two types ofand wet).
curingThe comparison of results lead to the
conclusion that both curing conditions have equivalent effect onstrength development and behavior. This observation is valid forall mixtures of low and high cement content.
4.1.4. Influence of Specimen Size: Anmade to obtain a series of results by
experimental attempt wastesting 2 x 4 inches and
4 x 8 inches phosphogypsum concreteproduced in
cylindricalthe laboratory by impact
specimenscompaction. Such test
results for the 28-day compressive strength are shown in Figure 7as a function of sand content. As expected, the 2 x 4 inch
220
221
224
specimens offered higher strength than those of 4 x 8 inchcylindrical samples.2-inch cylinder
The compressive strength ratio of 4-inch toincreased,
limerock sand.as phosphogypsum was replaced by
This result leads to the conclusion that the sizeeffect is more significant in the mixture of highcontent. The
phosphogypsumaverage compressive
cylinders was approximately 82%strength of the 4-inch
of its equivalent 2-inch diameterspecimens for the phosphogypsum-based concrete mixes. This valuei s slightly lower than the one predicted for conventionalconcrete mixtures.
4.2. Splitting and Bending Tensile Strengths: This portion of theinvestigation was aimed at finding a mathematical solution torelate compression and tension strength of concrete mixescontaining dihydrate phosphogypsum. To follow up on this matter,two standardized methods, splitting resistance (split-tensionstrength, ASTM.C-496) and tensile strength in bending (modulus ofrupture, ASTM-C-78) were adopted. The least square statisticalanalyses based on material law of the exponential type, similarto the one suggested for conventional concrete, were performed.As is shown in Figures 8 and 9, the distribution of theexperimental results yield in the following numericalexpressions for the 28-day sealed strengths:
F =s.t.
0.2013 (Fc).9473 (splitting-tensile strength)
F s.t.= 0.758 (Fc).845
(bending-tensile strength)
4.3 Static Modulus of Elasticity: The secant modulus ofelasticity corresponding to forty-five percent of ultimatestrength, as suggested by The American Concrete Institute (ACI) ,was selected to determine the modulus of elasticity of concretemixtures containing dihydrate phosphogypsum. Currently "AC I "uses equation of the form:
Ec = a(w) 3/2 (Fc)1/2
" a " is determined by considering the ratio,Using the method of
was found to beleast square, the
23.4 (the distributionof results is shown in Figure 10). This is lower than the onesuggested by the American Concrete Institute for conventionalconcrete (a = 33) Moreover, the resulting equation suggests thatinclusion of dihydrate phosphogypsum gives additional flexibilityand elasticity into concrete structures.
4.4. Comparison Study in Laboratory Consolidation: Theeffectivenessnamely:
of at least three different consolidation methods,impact compaction (Modified Proctor method), static
226
227
229
compaction (applied load at a deformation rate of 0.1 inch/min.without any energy release) and external high-frequency vibration(electrically drivenrpm) ,
vibrating table with a frequency of 3,600based upon the laboratory compressive strength criterion,
were investigated.
Two classes of cylindrical specimens, having 10 to 25%cement content, consolidated by using externalvibration and
high-frequencythe Modified Proctor impact compaction, are
presented in Figure 11 as a function of sand content. As it canbe seen, the strength curve corresponding to the mixesconsolidated by vibration steadily increases whereas, thestrength curve generated by compaction show a peak value. Thisis to say that, with proper proportioning and consolidationmethod, the use of dihydrate phosphogypsum improves the strengthcharacteristics of the concrete mixes.
Figure 12 illustrates the 28-day compressive strength as afunction of cement content, when the sand and gravel
40% for the threepercentage
is, kept constant at different methods ofconsolidation. The shaded portion of this diagram representsstrength results for mixtures subjected to static compactionunder pressure varying from 750 psi, lower limit, to 3,000 psi,upper limit. It can be noticed that the strength results obtainedby impact compaction (Modified Proctor method) are well insidethe shaded area, whereas the compressive strength data of thespecimens prepared by external vibration are substantially lowerfor the most portion of the curve. Furthermore, the resultsindicate that the lower cement content, the higher advantage ofusing compaction as opposed to vibration (strength isapproximately doubled at the cement content of 7.5%). Moreover,it should be noted that the difference among these consolidationmethods reduces as cement content of the mix increases; in 'factas is shown in Figure 12, at 25% cement content, the strengthresults obtained from external vibration falls within the shadedarea.
Better strength properties obtained by the compactionmethods rather than external vibration can be partiallyexplained by consideration on the mixing water content. Therequired optimum moisture content, used for the preparation ofimpact compaction specimens, together with their equivalentmixing water, used for preparation of mixtures by externalvibration, are shown in Figure 13 as a function of sand content.As it is seen the external vibration requires much more amountsof mixing water, as compared to its equivalent impact compaction.As a result, the corresponding strengths are much lower forspecimens prepared by external vibration consolidation.
4.5. Field Compaction: Field compacted phosphogypsum concretepavement slab, including a full scale plain phosphogypsumconcrete ramp, were constructed by using a field vibratoryroller. The core samples were tested and then compared with thelaboratory results in order to demonstrate the feasibility of
230
232
utilizing phosphogypsum-based concrete mixtures inconstruction projects.
road
Tables 1 through 3 present thethe equivalent
strength results of field andlaboratory specimens. As it is seen, the core
samples extracted from field compacted slabswith those obtained from laboratory tests.
compare favorablyThe strength results
obtained with road construction vibratory roller (6 in.layer,
lift, 110 passes) are equivalent to the mid values obtained in
the laboratory with Modified and Standard Proctor methods.
5. CONCLUSION
Data assembled during the course of this investigation leadsto the following conclusions:
1)
2)
3)
4)
5)
The unique properties of dihydrate phosphogypsum undercompaction consolidation can significantly contributeto the compressive strength of concrete mixes. Itscontribution to strength characteristics isparticularly noticeable in the mixture of low cementcontent.
The strength values of concrete mixtureshigh percentage of phosphogypsum are strongly
containingaffected
by the moisture at the time of testing. Both sealedand wet curing conditions have shown similar effect onstrength development.
The influence of specimen size is slightly higher thanthe one predicted for conventional concrete. Inaddition, the size effect is more significant in themixtures of high phosphogypsum content.
Both impact and static compaction are preferred overexternal vibration consolidation. BY using directforce application, the potential of the materials usedin the mixture can be fully utilized.
On-site applicationconfirms that goodcharacteristics
strengthshown by compacted
concrete in the laboratory,phosphogypsum
are also attainable in thefield.
234
ACKNOWLEDGEMENTS
Grateful acknowledgement is made to the Florida Institute ofPhosphate Research, Bartow, Florida, for funding this project.
REFERENCES
Ghafoori, N., "Phosphogypsum-Based Concrete:Characteristics and Road Construction
EngineeringApplication", Ph.D.
Dissertation, University of Miami, August 1986.
Lin., K. T., Figueroa, J. L., andProperties of Phosphogypsum",
Chang, W. F., "EngineeringProceeding of the Second
Seminar on Phosphogypsum, Florida Institute ofUniversity of Miami,
Research,May 1984, Miami, Florida.
Ghafoori, N., Chang, W. F., "Engineering Characteristics ofDihydrate Phosphogypsum-based Concrete", Proceedings of theThird Workshop on By-ProductsTampa, Florida, Dec. 4-6, 1985.
of Phosphate Industries,
238
DURABILITY OF PORTLAND CEMENT MORTARCONTAINING PHOSPHOGYPSUM
Chengsheng Ouyang, Ph.D. Candidate
Antonio Nanni, Assistant Professor
Wen F. Chang, Professor and DirectorPhosphate Research Institute
Department of Civil & Architectural EngineeringUniversity of Miami, Coral Gables, Florida
ABSTRACT
The use of phosphogypsum in portland cement-basedmixtures poses the question of sulfate attack. Expansion andcompressive strength of portland cement mortars containingphosphogypsum are examined in this study. Effects ofphosphogypsum content, cement content and type, curingconditions, fly ash addition, and different environment onsulfate attack are discussed in details. It is stated thatportland cement mixtures containing phosphogypsum can beused as a construction material without causing significantsulfate attack, provided that low C3A cement is used orphosphogypsum content is limited.
INTRODUCTION
Recent investigations (Lin, et al., 1986) indicatethat mixtures consisting of portland cement, conventionalaggregate and phosphogypsum possess valuable strengthsufficient for use in building and road construction.However, the presence of sulfate in phosphogypsum raisesthe question of durability. The chemical reaction betweenthe tricalcium aluminate (C3A) contained in portlandcement and the sulfur trioxide (S03) of phosphogypsum leadsto the formation of ettringite (3CaO.Al203. 3CaS0432H20),which is accompanied by expansion and strength loss(Mehta, 1967 and Samarai, 1976). This type of chemical
attack can be referred to as sulfate attack. Ultimately, thefeasibility of phosphogypsum based-applications is dependenton the ability of mixtures to resist sulfate attack.
The objective of this paper is to present a study onthe durability of mixtures containing portland cement,crushed limestone passing sieve No.4 (4.76 mm), andphosphogypsum. The variables considered in this study are:phosphogypsum content and origin, cement content and type,method of specimen fabrication, curing conditions, fly ash
addition, and different environments. Six types of portlandcement containing different amounts of C3A were used. Thecement content was varied between 5% and 30% by weight,whereas phosphogypsum content was varied from 0% to 95%.Three specimen fabrication methods, namely: rodconsolidation (ASTM C 452), modified proctor (ASTM D 1632)
239
and static compaction, were used with phosphogypsumobtained from four different plants. The linear expansionand the compressive strength of specimens submerged in freshwater, sea water and sulfate solution (ASTM C 1012),respectively, were recorded for a period up to one year.
EXPERIMENTAL PROCEDURES
The experiment procedures are based on ASTM C 157 andC 452. The standard 1x1x11-1/4 in. (25x25x285 mm) bar (ASTMC 452) was used to measure linear expansion. Compressivestrength was determined on 2x4 in. (51x102 mm) cylinders.Specimens consisted of portland cement, crushed limestoneand phosphogypsum. For the rod consolidated specimens, theamount of water used in each batch was such to produce a110+5% flow (ASTM c 109 and c 157), whereas for thespecimens prepared according to the other two compactionmethods, the optimum moisture content was used (the optimummoisture content is defined as the moisture content yieldingthe maximum dry density for given mix proportions).Specimens were removed from the molds stored in a fog room(100% RH, 24 C) after 47 hours and then immersed in freshwater, sea water or sulfate solution, respectively. Thesulfate solution, containing 4.3% magnesium sulfate and 2.5%sodium sulfate, was prepared according to ASTM C 1012.Thereafter, the linear expansion of bars was recorded andthe compressive strength of cylinders was tested atpredetermined intervals. 'Chemical composition of portlandcements, phosphogypsum and fly ash used in this study aregiven in Tables 1, 2 and 3, respectively.
240
EXPERIMENTAL RESULTS
PHOSPHOGYPSUM CONTENT
Phosphogypsum content strongly affects the expansionand the compressive strength of specimens as indicated inFigure 1. Specimens were prepared by using static compactionwith 3000 psi (20.7 MPa) pressure. Phosphogypsum contentvaried from 50% to 90%, whereas cement content remained 10%.Experimental results show that the increasing amount ofphosphogypsum yields larger expansion and lower compressivestrength. It is noted that specimen with 90% phosphogypsumhas strength loss after three months, but specimens with 50%and 70% phosphogypsum do not indicate strength loss up toone year.
CEMENT TYPE AND CONTENT
The expansion of bars made with 4 types of differentportland cement is presented in Figure 2. Specimens wereprepared by rod consolidation. Bars having high C3A cementshow high expansion, even though all specimens contained10% cement and 16% phosphogypsum.
The compressive strength and expansion of staticallycompacted specimens with 50% phosphogypsum are indicated inFigure 3. Cylinders containing cement with 8.8% C3A showstrength loss after three months. This suggests that low C3Aportland cement should be used for the mixtures containinghigh percentage of phosphogypsum.
Compressive strength ratio and expansion are plotted asfunction of cement content in Figure 4. Results show thatthere is an optimum cement content , which corresponds to theminimum expansion. This optimum cement content isapproximately 17% by weight for a portland cement with 7%C3A. It is noted that a higher than the optimum cementcontent in the mixture yields higher strength but alsohigher expansion.
CURING CONDITION
The formation of ettringite requires the presence ofwater which is provided by the wet curing. As aconsequence, durability is related to the curing condition.Expansion of specimens cured under two different conditions(sealed with a plastic membrane and continuous soaking), is
241
242
243
244
245
shown in Figures 5 and 6 as a function of time. The sealedspecimens show much lower expansion than the ones undercontinuous soaking (Figure 5). However, expansion is notsignificantly different for specimens soaked after 28of sealed curing and specimens continuously soaked
days
6).(Figure
ADDITION OF FLY ASH
An addition of low calcium fly ash can improve thesulfate attack resistance of portland cement mixtures. Theeffects of fly ash on expansion and compressive strength isgiven in Figure 7. The addition of low calciumreduces the expansion
fly ashof mortar bars and increases the
compressive strength of cylinders. This phenomenon can beexpected by the fact that the hydration of fly ash is limeconsuming instead of lime producing and fly ash reduces theporosity of the cement matrix (Mehta, 1986). It should benoted that specimens with and without fly ash have the sameexpansion during thereaction is slow.
first 45 days since the pozzolanic
SEA WATER AND SULFATE SOLUTION
The compressive strength and expansion of mortarsubmerged in fresh water, sea water and sulfate solution,respectively, are shown in Figure 8. Because sea water andsulfate solution contain sulfate ions, specimens in theseenvironments are under both internal and external sulfateattack. Sea water and sulfate solution causeexpansion and slower strength development than in the
largercase
of fresh water immersion. For the specimens containing 10%cement and 24% phosphogypsum, both sulfate solution and seawater immersions reducedof one year.
strength of above 5% up to period
The expansion and compressive strength ratio of bothstatically compacted and rod consolidatedsubmerged in
specimenssea water are indicated in Figure 9. The
static compaction technique reduces expansion andspecimens with a higher strength ratio.
producesThe fact that well
compacted specimens can improve their resistance to seawater attack may be attributed to that thetechnique
compactionreduces porosity so that the penetration of sea
water into the specimen is slowed down.
PHOSPHOGYPSUM ORIGIN
The expansion of mortar bars containing phosphogypsumfrom different manufacturers is given in Figure 10. Mortarbars containing different phosphogypsums have nearly thesame expansion, which seems to indicate that the resistanceof mortar to sulfate attack is independent of thephosphogypsum origin. The reason may be that phosphogypsumproduced by different manufacturers has almost same amountof SO3.
246
247
248
250
Fig. 9 Effect of Compaction on Sea Water Attack
251
Age (month)
Fig. 10 Expansion of Bars Containing Different Phosphogypsum
252
CONCLUSIONS
The use of phosphogypsum in portland cement-basedmixtures raises the question of sulfate attack which mayaffect durability. With respect to the sulfate attack causedby phosphogypsum, the following conclusions can be made:
1. The cement type is the most important single factorcontrolling sulfate attack. High C3A cements result in largeexpansion. Portland cement with C3A not greater than 7% issuggested for mixtures containing high percentage ofphosphogypsum.
2. Since phosphogypsum is the source of sulfate ions,expansion increase with the increasing of phosphogypsumcontent. Based on one-year observations, the use of 50%phosphogypsum in mixtures containing 10% cement with C3A notgreater than 7% does not yield strength loss.
3. Sealed curing results in less expansion than soakedcuring.
4. Low calcium fly ash can improve the sulfate attackresistance of mortar containing phosphogypsum. Addition offly ash reduces expansion and increase strength of mortar.
5. Submersion in sea water and sulfate solutionaccelerates the sulfate attack; however, staticallycompacted specimens show a better resistance to theaggressive environments.
6. Expansion is independent of the origin ofphosphogypsum.
ACKNOWLEDGEMENTS
This research project is funded by a grant from theFlorida Institute of Phosphate Research, Bartow, Florida.Portland cement was provided by Rinker Materials Corp.,Florida.' The phosphogypsum used in this study was obtainedfrom Gardinier Inc., Occidental Chemical Corp., AgricoChemical Corp., and USS Agri-Chemicals Corp.
REFERENCES
Lin, K. T., Nanni, A., and Chang, W. F., 1986. CompressiveStrength of Portland Cement-based Mixtures UsingPhosphogypsum. Proceedings of the Symposium onConsolidation of Concrete. ACI Annual Convention, SanFrancisco, CA.
Mehta, P. K., 1967. Expansion Characteristics of CalciumSulfoaluminate Hydrates. Journal of American CeramicSociety. Vol. 50, No. 4, April, pp.204-208.
Mehta, P. K., 1986. Concrete: Structures, Properties andMaterials. Prentice-Hall Inc., New Jersey.
Samarai, M. A., 1976. The Disintegration of ConcreteContaining Sulfate-Contaminated Aggregates. Magazineof Concrete Research. Vol. 28, No. 96, pp. 130-142.
253
STRENGTH PROPERTIES OF COMPACTEDPHOSPHOGYPSUM-BASED MIXTURES
Kuo-Ting Lin, Ph.D.
Wen F. Chang, Professor and DirectorPhosphate Research Institute
Department of Civil and Architectural EngineeringUniversity of Miami, Coral Gables, Florida
ABSTRACT
Phosphogypsum possesses binding strength when subjected tocompaction. However, phosphogypsum from different producersindicates strength variations. The aim of this paper is to studythe strength variability of plain and cement-based phosphogypsummixtures under compacton. Impact (the Modified Proctor) as wellas static compaction was used for specimens preparation.Phosphogypsum in general indicates good binding strength for roadand precast building applications. The strength can be furtherenhanced by the addition of cement.
INTRODUCTION
Dihydrate phosphogypsum (DPG), a solid by-product of thephosphate industry, possesses good binding property undercompaction and has good potential for use as a constructionmaterial. Substantial strengths are attainable for DPG alone andcement-DPG mixtures subjected to adequate compaction.
The objective of this paper is to investigate the basicstrength properties and their variability of DPG-based mixtures,using DPG from different producers, under impact as well asstatic compaction. Data presentation comprises the followingtwo categories:
1) Plain phosphogypsum. First, the strength properties of plainDPG supplied by 8 phosphate companies in Florida (GardinierInc., Agrico Chemical Co., USS Agri-Chemicals, IMC, Oxy,W.R. Grace, Royster, and Farmland) were studied to examinethe strength variability. Cylindrical specimens of plain DPGwere prepared under impact as well as static compaction andtested under compression. Moisture-density-strengthrelationships were obtained using impact (the ModifiedProctor) compaction, whereas five different pressures: 750,1,500, 3,000, 6,000, and 12,000 psi (1,000 psi= 6.89 MPa)were used for static compaction. The Modified Proctor methodwas employed to simulate the field operational process inroad construction, while static compaction was aimed atdeveloping proper use of DPG in the production of bricks and
255
blocks for building applications. Second, the effect ofhigh-pressure compaction on the strength properties of plainDPG supplied by Gardinier Inc. was further investigated.Static pressures up to 36,000 psi (248.3 MPa) were applied.High-pressure compaction made the material an extremelystrong material and resistant to soaked conditions.
2) Portland cement-based mixtures. The compressive strengthproperties of cement (10%) and DPG (from the eight sources)mixtures, subjected to static compaction pressures from 750to 12,000 psi are investigated and compared. Comparisons ofcompressive strength results are made to indicate thestrength variability and also the different effects of cement(10%) on the strength enhancement of DPG from the 8 sources.
PLAIN PHOSPHOGYPSUM
IMPACT COMPACTION
T h e effect of compaction moisture content on compressivestrength and dry density of plain DPG provided by the eightphosphate companies was studied. Cylindrical specimens werecompacted using the Modified Proctor method. Figures l(a)-l(h)represent the relationship between compressive strength, drydensity, and compaction moisture content of DPG from thesecompanies, respectively. The strength of specimens was testedunder air-dry conditions; when subjected to soaked conditions,the specimens disintegrated for all DPG's from these companies.
It is seen from these figures that the maximum compressivestrength of DPG from these companies generally exists at theoptimum moisture content, i.e. the moisture content at which themaximum dry density is obtained. Thus, it is important tocompact DPG et the optimum moisture content in order to achievemaximum strength and density values.
Also observed from these figures are the variations inmaximum compressive strength. Figure 2 is the comparison of thecompressive strength of different sources of DPG. As can be seen,the maximum compressive strength (under air-dry conditions)varies from 100 psi (0.69 MPa) up to 800 psi (5.52 MPa); althoughthe strength of most DPG ranges between 300 and 500 psi (2.07 end3.45 MPa). Thus, the source of DPG has en important effect onthe compressive strength of DPG under the Modified Proctorc o m p a c t i o n .
STATIC COMPACTION
Table 1 presents the compressive strength under air-dry andsoaked conditions and the dry density of the 8 sources of DPGsubjected to static compaction pressures from 750 to 12,000psi. Specimens were compacted at optimal moisture content.Optimal moisture is defined es the experimentally found maximumwater content at which the water from the mixture was not pressed
256
260
261
out at the given compaction pressure. The optimal moisturecontent of DPG from these sources decreases with increasingcompaction pressure, as shown in Table 2.
From Table 1, the following can be observed:
1) The compressive strength under air-dry conditions and thedensity of all DPG's increase continuously with compactionpressure. It is therefore indicated that DPG in general hasthe material characteristic that the strength is beneficial-ly affected by increased compaction pressure.
2) DPG from all companies except USS has compressive strengthunder soaked conditions when adequate compaction pressuresare applied. At low compaction pressures, such as 750 and1,500 psi (5.17 and 10.34 MPa), either the specimensdisintegrate or the strength is negligible. The strengthbecomes significant when the compaction pressures are higherthan 3,000 psi (20.69 MPa). Further increase of compactionpressure results in strength enhancement under soakedconditions.
3) DPG from different sources indicate varying compressivestrength values, as is the case for DPG under impactcompaction. To see this more clearly, Figures 3 and 4compare the compressive strength under air-dry and soakedconditions, respectively, of these DPG's.
AS Figures 3 and 4 present, the strength variability ofdifferent sources of D P G i s dependent upon the compactionpressure. The variations of strength enlarge as the compactionpressure increases. Although the strength in all casesincreases with the compaction pressure, the rates of increase arevariant for DPG from different sources, as can be seen fromFigures 3 and 4. Nevertheless, significant strength results aregenerally obtained, particularly at high compaction pressures.When compacted at 12,000 psi, DPG from the eight sources havecompressive strengths of 800-2,500 psi (5.52-17.24 MPa) underair-dry conditions, and 250-650 psi (1.72-4.48 MPa) under soakedconditions.
HIGH-PRESSURE COMPACTION
As a result of the excellent strength properties of DPG ingeneral under increased compaction pressure, the effect of high-pressure compaction was further studied on the strengthproperties of DPG supplied by Gardinier Inc. Static pressures upto 36,000 psi were applied. Figure 5 presents the compressivestrength, under both air-dry and soaked conditions, of DPG(provided by Gardinier) as a function of static compactionpressure. As can be observed, high-pressure compactiontremendously strengthens the material under air-dry as well assoaked conditions. Compacted at 36,000 psi, the DPG supplied byGardinier indicates cylindrical compressive strength of 5,500 psi(37.93 MPa) under air-dry conditions and 1,800 psi 112.41 MPa)under soaked conditions.
262
263
264
265
PORTLAND CEMENT-BASED MIXTURES
EFFECT OF CEMENT
Cement has the effect of increasing the strength of DPGunder air-dry and soaked conditions. As an example, Figure 6compares the compressive strength of 10% cement-90% DPG and ofplain DPG mixtures (DPG provided by Gardinier), under both air-dry and soaked conditions, subjected to various static compactionpressures. As seen, the cement (10%) has a positive effect onstrength improvement under both testing conditions for allcompaction pressures indicated in Figure 6. The effect onstrength improvement is not identical, however, for differenttesting conditions as well as for different compaction pressures,To elaborate on this, Table 3 presents the strength ratio(compressive strength of 10% cement-90% DPG to that of plain DPG)in variation with static compaction pressure for the two testingconditions. It is first observed that the strength ratiodecreases as the compaction pressure increases, meaning thatcement has more effects on strength improvement for lowercompaction pressures. Second, the strength ratio for the soakedtesting conditions is consistently higher than that for the air-dry testing conditions, indicating that the use of cementincreases the strength under soaked conditions more significantlythan under air-dry conditions.
STRENGTH VARIABILITY
Table 4 lists the compressive strength under air-dry andsoaked conditions and the dry density of the 10% cement-90% DPGmixtures, using DPG from the 8 sources. Mixtures were preparedunder static compaction at optimal moisture content as presentedin Table 5. From Table 4 it is seen that both the compressivestrength under air-dry and soaked conditions and the dry densityincrease with compaction pressure, as is the case for plain DPGmixtures. Moreover, strength variations can be observed for thesecement-DPG mixtures. This is shown more clearly in Figures 7 and8, where the compressive strength under air-dry and soakedconditions, respectively, is plotted as a function of compactionpressure.
267
268
269
270
271
The strength variations of these cement-DPG mixtures can beattributed to the following two factors:
1) Strength variations of DPG alone.2) Unequal effects of cement on the strength of DPG.
The first factor has been described in the foregoing paragraphs,and the second factor can be seen if strength results of plainDPG and of' cement-DPG mixtures are compared for DPG fromdifferent sources. For example, Figure 9 compares the effects ofcement (10%) on the compressive strength of DPG from two sources(Farmland and Gardinier). While the Gardinier source shows bettercompressive strength than the Farmland source for the case ofplain DPG mixture, the converse is observed for the case ofcement (10%) and DPG mixture. In other words, a higher strengthvalue resulting from plain DPG does not necessarily imply ahigher strength value resulting from a cement and DPG mixture.
TO compare generally the effect of cement (10%) on thestrength of DPG from the 8 sources, Table 6 presents the strengthratios (compressive strength of 10% cement-90% DPG to that ofplain DPG) for mixtures compacted at 12,000 psi and tested underair-dry and soaked conditions. Since the strength results usinghigh-pressure compaction would be of more concern, comparison ofstrength ratios is made only for the compaction pressure of12,000 psi. As it is illustrated, cement (10%) increases thecompressive strength of DPG from the 8 sources both significantlyand differently. By the use of 10% cement, the compressivestrength of DPG from different sources, increases approximately 2to 4 times under air-dry conditions and 3 to 7 times under soakedconditions. This means that the effect of cement on strength isnot constant for different sources of DPG and that the strengthimprovement is generally much greater for specimens tested undersoaked conditions than those tested under air-dry conditions.
272
273
REFERENCES
Ho, R., R. Williams, and W.F. Chang. 1986. Columbia CountyExperimental Road. The Second international Symposium onPhosphogypsum. Florida Institute of Phosphate Research.Bartow. FL.
Kenley, W., and W.F. Chang. 1986. Polk County ExperimentalRoad. The Second International Symposium on Phosphogypsum.Florida Institute of Phosphate Research. Bartow. FL.
275
STABILIZATION OF PHOSPHOGYPSUM AND
SAND MIXTURES WITH SULFUR
Wen Chain ChangVice Chief Engineer
Planning & Designing Institute of Water TransportMinistry of Communication, China
Wen F. ChangProfessor and Director
Phosphate Research InstituteDepartment of Civil & Architectural Engineering
-University of Miami
ABSTRACT
Sulfur concrete is widely used in chemical plants toresist corrosion by acid and salts. In this research,waste sulfur was used to stabilize phosphogypsum and sandmixtures: and the engineering properties of the resultingcomposite were investigated. The effects of sulfurcontent, consolidation pressure and sand content on thecompressive strength, and the splitting tension strengthwere studied. Experimental results showed that the wastesulfur stabilized phosphogypsum and sand mixtures,subjected to static compaction, are a promisingconstruction material.
INTRODUCTION
Sulfur concrete (SC) utilizing a mixture ofchemically modified sulfur and suitable mineral aggregateswas developed by the Bureau of Mines in 1972, to resistcorrosion by acids and salts. This composite isused in metallurgical, chemical,
widelyand fertilizer processing
plants (Mcbee, et al., 1983 ).In this research project waste sulfur from the
fertilizer industryand sand mixtures,
was used to stabilize phosphogypsumand the engineering properties of the
resulting composite were investigated. Compression tests,under both air-dry and soaked conditions, andtension
splitting tests under air-dry conditions were performed on
2x2x2 in. (50.8x50.8x50.8 mm) cubes. Strength propertiesare expressed as a function of the static compactionpressure, sulfur content and sand content. Experimental
277
results showed that the waste sulfur stabilizedphosphogypsum and sand mixtures, subjected to staticcompaction, possess adequate compressive strength for useas a construction material.
SPECIMEN FABRICATION AND TEST
The constituent materials consisting of waste sulfur,silica sand,mixer and
phosphogypsum and water were mixed in a panconsolidated in a steel mold at different
pressures: 500, 3,000, 6,000 and 12,000 psi (3.45, 20.69,41.37 and 82.74 MPa).
After fabrication the specimens were removed from themold, stored for two days in the laboratory, and thencured in a convection. oven to melt the sulfur, fordifferent period of time and different temperatures. Itwas found that, for 2 in. (50.8 mm) cubic specimens underpressures from 500 psi (3.45 MPa) to 6,000 psi
curing at(41.37
MPa) , a temperature of 140 C degrees for twohours obtained the maximum strength. For this reason, twohours of curing time and temperature of 140 C degrees wereused for all subsequent specimens.
Compressionconditions,
tests under both air-dry and soakedand splitting
conditionstension tests under air-dry
were performed to determine the properties ofthe composite. Specimens for soaked strength testing weresoaked in water for two days before testing.
EXPERIMENT RESULTS
The effect of sulfur contentstrength
on the compressiveunder air-dry conditions is presented in Figure
1. The test resultsstrength
indicated that the compressiveof specimens increased with the sulfur content
when subjected to static compaction pressures at 500 psi(3.45 MPa) and at 3,000 psi (20.69 MPa); whereas, when thestatic compactionand at
pressure was at 6000 psi (41.37 mpa)12000 psi (82.74 mpa),
decreasedthe compressive' strength
as sulfur content went above 15% by weight.The decrease in strength is because portion of the sulfurflowed out of the specimens, while they were curing in theoven.
The effect' of sulfur content on the percentage ofsoaked strength to air-dry strength is given in Figure 2.For plain phosphogypsum the strengths under soakedconditions were much lower than air-dry conditionsChang,
(Lin,1986); whereas it can be observed from Figure 2,
increasing the sulfur content increased the percentage ofsoaked strength to air-dry strength. At a staticcompaction pressure of 3000 psi (20.69 mpa) , the
278
Sulfur Content (%)
Fig. 1. Compressive Strength Under Air-Dry Conditions
Vs. Sulfur Content for Different Static
Compaction Pressure.
279
Sulfur Content (%)
Fig. 2. Percentage of Soaked Strength to Air-Dry
Strength Vs. Sulfur Content for Different
Static Compaction Pressure.
280
percentage increased from 47 % to 85%, when the sulfurcontent increased from 5 % to 20% by weight. At a staticcompaction pressure of 6000 psi(41.37 mpa) , when thesulfur content increased from 5% to 20% by weight, thepercentage increased from 50% to 94%.
The effect of static compaction pressure on thecompressive strength under air-dry conditions fordifferent sulfur contents is reported in Figure 3. Resultsshowed that the compressive strength increased as thestatic compaction pressure increased, when sulfur contentswere 5% and 10% by weight. Whereas, at a sulfur content of15% by weight, the strength decreased when the staticcompaction pressure was greater than 6000 psi(41.37 mpa);and at a sulfur content of 20% by weight, the strengthdecreased when the static compaction pressure was greaterthan 3000 psi(20.69 mpa).
The, effect of static compaction pressure on thepercentage of soaked strength to air-dry strength can beobserved in Figure 4. The ratio of soaked strength to air-dry strength increased when the static compaction pressureincreased from 500 psi(3.45 mpa) to 6000 psi(41.37 mpa);whereas, it decreased when the static compaction pressurewas more than 6000 psi(41.37 mpa).
The effect of sand content on the compressive strengthunder air-dry conditions, for a given sulfur content of15% is shown in Figure 5. It shows that, increasing theamount of sand content up to 60 % increased thecompreassive strength. Further increase in sand content,the specimens became unstable at fabrication after theremoval from the mold.
The ratio of splitting tensile strength to compresivestrength, defined as strength ratio, under air-dry.conditions is presented in Table 1. The ratio as obtained,is in the range of 10 to 20 %.
281
Static Compaction Pressure (psi.)
Fig. 3. Compressive Strength Under Air-Dry Conditions
Vs. Static Compaction Pressure for Different
Sulfur Content.
282
Static Compaction Pressure (psi)
Fig. 4. Percentage of Soaked Strength to Air-Dry Strength
Vs. Static Compaction Pressure for Different Sulfur
Content.
283
Fig. 5. Compressive Strength Vs. Sand Content for
Different Static Compaction Pressures.
CONCLUSIONS
Based on the results of the experimental study, thefollowing conclusions can be drawn:
1. Sulfur stabilized phosphogypsum and sand mixturesunder static compaction have higher compressive strengthsthan that of phosphogypsum and sand mixtures.
2. Sulfur stabilization improves the waterresistance of phosphogypsum and sand mixtures, asevidenced by the increase in the ratio of soaked strengthto air-dry strength.
3. Increasing the amount of sand content in themixtures up to 60 % increased the compressive strength.
4. Phosphogypsum mixtures stabilized with wastesulfur as described have the potential for use inconstruction.
ACKNOWLEDGMENT
The authors wish to express their gratitude to theFlorida Institute of Phosphate Research for sponsoringthis research project.
REFERENCES
Lin, K.T. and W.F. Chang. 1986. Strength Properties ofCompacted Phosphogypsum based Mixtures. CondensedPaper of Second International Symposium onPhosphogypsum.
Mcbee, W.C. Sullivan, T.A. and B.W. Jong. 1983. IndustrialEvaluation of Sulfur Concrete in CorrosiveEnvironments. U.S. Department of the Interior, Bureauof Mines, Pittsburgh, P.A. 26962, U.S. Government
285
STABILIZATION OF PHOSPHOGYPSUM WITH EPOXY
Wen Chain ChangVice Chief Engineer
Planning and Design Institute of Water TransportMinistry of Communication, China
Wen F. ChangProfessor and Director
of Phosphate Research instituteDept. of Civil and Architectural Engineering
University of Miami
ABSTRACT
The engineering properties of epoxy stabilizedphosphogypsum have been investigated. Compression testsunder both air-dry and soaked conditions, and splittingtension tests under air-dry conditions were used toevaluate the composite properties.
It was found that epoxy is a good stabilizer and itcan greatly increase the compressive strength ofphosphogypsum under both air-dry and soaked conditions.Increasing the consolidation pressure can reduce theamount of epoxy content. The splitting tension is equal to10-13% of the compressive strength.
1. INTRODUCTION
Phosphogypsum is a waste by-product of phosphorousfertilizer production. Extensive research on the use ofthis waste as a construction material has been conductedat the University of Miami since 1983 ( Nanni, Chang,1984). Phosphogypsum specimens tested under air-dry
conditions can develop compressive strengths varying from250 psi (1.72 MPa) to over 5,500 psi, depending on thedegree of consolidation. However, compressive strengthunder soaked conditions is much lower than that under air-dry conditions (Lin, et a1., 1985)(Lin,Chang,l986). Thewater resistance of phosphogypsum can be improved by theaddition of other binding agents such as portland cementand synthetic resin.
Epoxy is a two component synthetic resin consisting ofthe epoxy resin and the curing agent. Many different epoxyresins are available today, each of these may be used witha number of curing agents as well as with many modifiersand diluents. Epoxy can be used as a binder inconcrete and its major advantages are higher strength,
287
exceptional dimensional stability after curing, excellentadhesion to a wide range of materials, fatigue resistance,creep resistance and high resistance to most chemicalattacks by many acids, alkalis and solvents.
The objective of this research is to investigate theengineering properties of epoxy stabilized phosphogypsum.Compression tests under both air dry and soakedconditions, and splitting tension tests under air-dryconditions were used to evaluate the composite properties.
Three types of epoxy were used in this experiment,all of them are commercial products.of Pettif Paint Company,
El is, the 7085 epoxyE2 and E3 are the 103LV and 9X-S
LP epoxy of Fasco Unlimited of Hialeah Inc., respectively:and the phosphogypsum used in this study was supplied byGardinier Inc., Tampa, Florida. Cubic specimens of 2 in.(50.8 mm) were used in this experiment. The specimens werefabricated under different consolidation pressures andwere air-dried under laboratory conditions aftercompaction. Specimens for soaked strength testing wereimmersed in water for two days before testing.
2. TIME VERSUS COMPRESSIVE STRENGTH RELATIONSHIPS
The compressive strength of "air-dry plain phosphogypsum. specimens increases rapidly in the firstfive days and no additionalobserved .
strength development isafter two weeks of curing
Figure(Lin, et al., 1985).
1 representsstabilized
the compressive strength of epoxy
with age.phosphogypsum under air-dry conditions varying
The pattern of relative strength ratio (strengthat given age to 28-day strength) is presented in Figure 2: at 3 days, the relative strength ratio ranged between 70and 75%, at 7 days, the relative strength ratio ranged
between 80 and 85%; and at 14 days, the relative strengthratio ranged between 95 and 100 %.
3 . EFFECT OF EPOXY CONTENT ON COMPRESSIVE STRENGTH
Test results indicate that the use of epoxy in phosphogypsum 'mixtures can greatly increase compressive
strength under air-dry conditions.increasing the amount of epoxy,
As shown in Figure 3,
compressive increases the composite
strength. Forconsolidated at
p l a i n phosphogypsuma pressure of 3,000 psi (20.69 MPa) the
compressivewith
strength was 1,062an epoxy(E2) content of 10
psi (7.32, MPa); whereasthe
compressive strengthand 20% by -weight,
and'increased to 5,138 psi (35.43
10,579 psi (72.94 MPa),MPa)
consolidationrespectively. At
compressivepressure of 500 psi (3.45 MPa), the
strength of plain phosphogypsum was 478 psi
288
Fig. 1. Age Versus Compressive Strength
289
Fig. 2. Age vs. Percentage of 28-day Strength
290
Epoxy Content (%)
Fig. 3 Relationship Between Epoxy Content and Compressive
Strength Under Air-Dry Conditions for Different
Consolidation Pressure.
291
(3.30 MPa), whereas at an epoxy(E2) content of 5 and 20% by weight, the compressive strength increased to 1,312 psi(9.05 MPa) and 5,912 psi (40.76 MPa), respectively.
4. EFFECT OF CONSOLIDATION
PRESSURE ON COMPRESSIVE STRENGTH
Figure 4 shows that, the compressive strength of epoxy stabilized phosphogypsum specimens under air-dryconditions increased with consolidation pressure.Increasing the pressure from 500 psi (3.45 MPa) to 6,000psi (41.37 MPa) at an epoxy(E2) content of 10%, improvedthe compressive strength from 2,382 psi (16.43 MPa) to6,881 psi (47.44 MPa). At a consolidation pressure of12,000 psi (82.74 MPa) and an epoxy(E2) content of 10%yielded a compressive strength of 11,429 psi (78.80 MPa),which is equivalent to the value obtained forwith 15% epoxy and consolidated at a pressure of( 4 1 . 3 7 M P a ) .
specimens6,000 psi
5. SPECIMENS UNDER SOAKED CONDITIONS
The effect of epoxy content on compressive strengthunder soaked conditions is presented in Figure 5. Itillustrates that, soaked strength is directly proportionalto epoxy content. For a consolidation pressure of 6,000psi(41.37 mpa), increasing the epoxy(E3) content from 5 to15% by weight, increased the soaked strength from 929 psi(6.41 mpa) to 10,063 psi(69.36 mpa).
The effect of consolidation pressure on compressivestrength, under soaked conditions, is presented in Figure6. It indicated that soaked strength is directlyproportional to consolidation pressure. At an epoxy(E3)content of 10%, increasing consolidation pressure from 500psi (3.45 mpa) to 12,000 psi(82.74 mpa), increased thesoaked strength from 1,255 psi(8.66 mpa) to 10,121 psi(69.78 mpa).
6. COMPARISON BETWEEN THE AIR-DRY AND SOAKED STRENGTH
Compressive strength under both air-dry and soakedconditions with differentFigure 7.
epoxy contents is shown in'It can be observed that, the percentage of
soaked strength to air-dry strength increased with epoxycontent.
Comparison between the air-dry and soaked strengthwith different consolidation pressure is as shown in
292
8000'
6000
0 500
Consolidation Pressure (psi)
Fig. 4. Relationship Between Consolidation Pressure
and Compressive Strength Under Air-Dry
Conditions.
293
12000
Epoxy Content (%)
Fig. 5 Relationship Between Epoxy Content and
Compressive Strength Under Soaked Conditions
for Different Consolidation Pressure.
294
295
Epoxy Content
Fig. 7. Percentage of Soaked Strength to Air-Dry
Strength Versus Epoxy Content
296
Figure 8. The percentage of soaked strength to air-drystrength increased with the consolidation pressure.
It may be noted that, when epoxy(E3) content of 15%,and a consolidation pressure of 6,000 psi(41.37 mpa), thesoaked compressive strength was 96.4% of the air-drycompressive strength, and for an epoxy(E2 and E3) contentof 10% and a consolidation pressure of 12,000 psi(82.74mpa), the soaked compressive strengths were 98.0% and 98.6%of the air-dry compressive strength, respectively.
7. RELATIONSHIP BETWEEN
SPLITTING TENSION AND COMPRESSIVE STRENGTH
The ratio of splitting tension to compressivestrength versus compressive strength is plotted in Figure9. It is shown that, the strength ratio (splitting tensionto compression) decreases approximately from 13 to 10%as the compressive strength increases from 1312 psi(9.05mpa) to 11495 psi(79.26 mpa), a relationship which issimilar to that of conventional concrete.
Based oncan be drawn:
a) Epoxybinder in a
8. CONCLUSIONS
the test results the following conclusions
is a good stabilizer. Use of epoxy as aphosphogypsum matrix can greatly increase
compressive strength under both air dry and soakedconditions.
b) Increasing the consolidation pressure can raisethe compressive strength under both air dry and soakedconditions. Therefore, it can reduce the epoxy content.When a consolidation pressure higher than 12,000 psi(82.74 MPa) is applied and the epoxy content is over 10%,the compressive strength under both air-dried and soakedconditions are almost the same.
c) The strength ratio (splitting tension tocompression) of phosphogypsum epoxy specimen decreasesslightly as compressive strength increases, similar tothat of conventional concrete.
ACKNOWLEDGEMENT
The project was funded by a grant from the FloridaInstitute of Phosphate Research, Bartow, Florida.
297
Consolidation Pressure (psi)
Fig. 8 Percentage of Soaked Strengthto Air-Dry
Strength Versus Consolidation Pressure.
298
299
REFERENCES
Lin, K.T. Figuroa, J.L. and W.F. Chang. 1985.Engineering Properties of Phosphogypsum.of The Second Workshop on
ProceedingsBy-Products of the
Phosphate Industry. pp. 49-59.Lin, K.T. and W.F. Chang. 1986. Strength Properties of
Compacted Phosphogypsum-Based Mixtures. Proceedingsof the Condensed Papers of The Second. InternationalSymposium on Phosphogypsum. pp. 181-186.
Nanni, A. and W.F. Chang. 1985. Phosphogypsumfor The Construction Industry.
ProductsProceedings of
the Second Workshop on By-Products of ThePhosphate Industry. pp. 153-167.
STRENGTH PROPERTIES OF PORTLAND CEMENT-BASED MIXTURES USINGPHOSPHOGYPSUM SUBJECTED TO STATIC COMPACTION
Kuo-Ting Lin, Ph.D.
C. Ouyang, Ph.D. Candidate
Wen F. Chang, Professor and DirectorPhosphate Research Institute
Department of Civil and Architectural EngineeringUniversity of Miami, Coral Gables, Florida
ABSTRACT
Excellent strength results of plain phosphogypsum understatic compaction suggest that similar behavior can be observedfor portland cement-based mixtures using phosphogypsum. Thisstudy reports the effects of compaction moisture content,compaction pressure, mix proportions, and extended soaking periodon the strength properties of cement, DPG, and fine aggregate(sand) mixtures under static compaction. Beneficial strengthresults, particularly when using high-pressure compaction,indicate that these mixtures have good potential for the precaststructural products.
INTRODUCTION
Under direct applied pressures, dihydrate phosphogypsum(DPG) particles bind together and develop certain compressivestrengths [Lin, et al., l985]. When the applied pressures aresubstantial, DPG particles lose their original particle sizes andshapes, binding into a rock-like body [Wirsching, 19841.Compressive strengths of different magnitudes, depending on theintensity of the applied compaction pressure, have been obtainedfor plain DPG under static compaction [Lin and Chang, 19861.
The objective of this paper is to investigate the strengthproperties of portland cement, sand (crushed limerock), and DPGmixtures under static compaction at various pressures from 750 upto 24,000 psi (5.17 to 165.5 MPa). Laboratory experiments of thisstudy focused on the following factors affecting the strengthproperties of these mixtures:
(a) Compaction moisture content,(b) Compaction pressure, and(c) Mix proportions.
Cylindrical specimens (2x4 in.= 51x102 mm) were compacted ina rigid mold in two equal layers, cured under sealed conditions,and tested under air-dry as well as 'soaked conditions.
301
In addition, the compressive strength of portland cement,sand, and DPG mixtures subjected to extended duration of soakedconditions was further tested to indicate the durability propertyof the mixtures. Effects of compaction pressure and cementcontent in the mixtures were investigated. Test data of specimenssoaked for up to six months are reported in this study.
FACTORS AFFECTING STRENGTH
COMPACTION MOISTURE CONTENT
Table 1 presents the 28-day compressive strength invariation with the moisture content at the time of mixing for agiven DPG-sand-cement mixture subjected to static compaction.Specimens were compacted at 3,000 psi (20.69 MPa), and testedunder air-dry conditions.
As can be seen, the mixing moisture content has an effect onthe strength of the mixture. The highest strength exists at amoisture content of around 10% rather than 8%. Namely, theAbrams' law, which states that for conventional concrete mixes ofgiven materials "the strength depends only on one factor-- theratio of water to cement", does not apply to mixtures understatic compaction.
COMPACTION PRESSURE
The strength properties of DPG-sand-cement mixtures,. understatic compaction can be significantly enhanced by increasing thecompaction pressure. The strength enhancement springs from thefollowing contributing factors:
1) Continuous and significant strength increase of DPG withcompaction pressure [Lin and Chang, 19863,
2) Reduced water/cement ratio and porosity of the mixture withincreased compaction pressure.
It has been observed that the strength of cement paste understatic compaction increases greatly with compaction pressure[Skalny and Bajza, 1970, Lecznar and Barnoff, 19613. Increased
302
compaction pressure reduces the porosity of the paste and thusimproves the strength properties. Figures 1 and 2 present thecompressive and splitting-tensile strengths of a DPG-sand-cementmixture tested under air-dry and soaked conditions as affected bythe compaction presssure. It is explicit that the compressive aswell as the tensile strength of the mixture under both air-dryand soaked conditions improves continuously with the compactionpressures ranging from 750 to 12,000 psi (82.76 MPa). Thestrength improvement is attributed to both DPG and cement, whosebinding capability increases with compaction pressure. Alsoobserved is the effect of testing conditions. The strength ofDPG-sand-cement mixtures under air-dry conditions is consistentlyhigher than that under 'soaked conditions for any given compactionpressure, as can be seen from Figures 1 and 2.
MIX PROPORTIONS
The strength of DPG-sand-cement mixtures under staticcompaction is affected by mix proportions. In this paragraph,the effects of cement content and sand (or DPG) content on thecompressive strength of DPG-sand-cement mixtures prepared atvarious pressures (between 750 and 24,000 psi) are investigated.
Figures 3 and 4 show the compressive strength under air-dryand soaked conditions, respectively, of DPG-sand-cement mixturesat various compaction pressures (up to 12,000 psi), as a functionof cement content. The sand content was held constant at 45%,while the cement and DPG contents varied. From these curves, itis observed that, for all compaction pressures, the compressivestrength increases with the cement content up to 25%. Also, itis noticed that equivalent strength results can be obtained byeither approach of the following:
1) A high cement content at a low compaction pressure,2) A low cement content at a high compaction pressure.
According to Figure 3, for example, to obtain the compressivestrength of 4,000 psi (27.59 MPa) under air-dry conditions, thecement content can be around 6% at a compaction pressure of12,000 psi, or 15% at 3,000 psi, or 25% at 750 psi, when the sandcontent is 45%. Use of low cement content at high compactionpressure is economically more attractive, comparing the greatsaving on cement quantity with the limited energy input for high-pressure compaction.
In addition to cement content, sand (or DPG) content alsoaffects the strength of DPG-sand-cement mixtures. Figures 5 and6 indicate the compressive strength of DPG-sand-cement mixturesunder air-dry and soaked conditions, respectively, as affected bysand content. The cement content was held constant at 7.5%, whilethe sand (or DPG) content varied. From these figures, first, itis seen that for all mixtures herein reported the strengthincreases as compaction pressure increases. Second, the strengthcurves corresponding to different compaction pressures indicate
303
Figure i: Effect of static compaction pressure on compressivestrength of a DPG-sand-cement mixture.
304
Figure 2: Effect of static compaction pressure on tensilestrength of a DE-sand-cement mixture.
305
306
Figure 4: Effect of cement content on compressive strengthof DPG-sand-cement mixtures under static compaction(soaked testing).
307
308
basically two types of strength characteristics.
For compaction pressures no greater than 6,000 psi (41.38MPa), the compressive strength of DPG-cement (zero percent sand)can be increased by replacing DPG with sand (Figure 5).Increase in strength continues until peak strength values arereached at a sand content around 70% (DPG content around 20%).It is interesting to mention that similar strength characteristichas been observed for DPG-sand-cement mixtures under impactcompaction [Lin, et al., 19861. At the same time, use of DPGbelow 70% (sand above about 20%) provides mixtures with highercompressive strengths than when no DPG (sand-cement mixture) isused in the mixture.
On the other hand, when the compaction pressure is higherthan 12,000 psi, another strength characteristic is observed. At12,000 psi, the binding capability of DPG increases to such enextent that replacement of DPG with sand has only a very slighteffect on strength increase. As the compaction pressure reaches24,000 psi, the binding strength of DPG further increases andreplacement of DPG with sand results in strength decrease (Figure5). Thus, it is indicated that mixtures with more DPG havegreater potential for strength improvement with increased highcompaction pressures.
To see more clearly the effect of compaction pressure,Figure 7 compares the compressive strengths (air-dry conditions)of mixtures with 7.5% cement and three sand (or DPG) contents, inrelation to compaction pressure. For compaction pressures lowerthan about 12,000 psi, as can be observed, the higher the sandcontent the higher the compressive strength. It should bereminded here that this relation holds only for sand contentbelow about 70%, where the optimum compressive strengths develop,as presented, previously in Figure 5. Beyond this sand content(70%), the strength decreases with further increase of sandcontent.
Conversely, a's the compaction pressure well exceeds 12,000psi, mixtures with more DPG contents (less sand) indicate bettercompressive strength values, as can be seen in Figure 7. It isrelevant to point out that this relation holds for mixtures with7.5% cement and any DPG (or sand) content compacted at highpressures (well over 12,000 psi).
MIXTURES SUBJECTED TO LONG-TERM SOAKING
GENERAL
The chemical reaction between the tricalcium aluminate (C3A)contained in portland cement and the sulfur trioxide (SO3) inphosphogypsum, with the presence of water, causes the formationof ettringite (3CaO.Al203. 3CaSO4.32H20), which in general isaccompanied by expansion and strength loss. For mixtures of givenproportions, the amount of expansion and strength loss depends on
310
Figure 7: Comparison of compressive strength of mixtureswith three sand contents as related to staticcompaction pressure (air-dry testing).
311
the percentage of C3A in the cement used. The higher the C3Acontent, the more the expansion and strength loss [Ouyang, etal., 19861. In this study, the portland cement used has C3Acontent of 7%. In the following paragraphs, effects of compactionpressure and cement content on the compressive strength of DPG-sand-cement mixtures subjected to long-term soaking arepresented. Test specimens after compaction were first curedunder wet conditions (in a moisture room with 100% relative'humidity) for 48 hours, then soaked in water until testing date.Compression tests were conducted immediately after the removalof specimens from the soaking environment.
EFFECT OF COMPACTION PRESSURE
Figure 8 shows the effect of static compaction pressure onthe compressive strength of a mixture containing 10% cement, 50%DPG, and 40% sand as a function of soaking age. As seen, theextended soaking period has no negative effect on theproperties.
strengthIn fact, the strength increases, depending on the
magnitude of compaction pressure, with the soaking period up tosix months. For a relatively low compaction pressure of 1,000 psit h e increase is gradual and limited, while for higher pressuresthe increases are more significant. Hence, increased compactionpressures indicate improvement in durability in addition tostrength properties.
EFFECT OF CEMENT CONTENT
Figure 9 indicates the effect of cement content (5, 7.5, and10%) on the compressive strength as a function of soaking age ofDPG-sand-cement mixtures having a constant DPG content of 37.5%.Specimens were compacted at a pressure of 3,000 psi. As shown,the strength is not affected by the soaking period for all threecement contents. Strength actually increases as a result ofcontinuous curing. The negative effect of ettringite formationo n strength is not observed due to the low C A content (7%) ofthe cement used and also the low DPG content (37.5%) in the-mixture. Recent study [Ouyang, 19861 indicates that for low C Acement the strength is not affected if the DPG content is lessthan 50%.
CONCLUSIONS
Test results of strength properties of portland cement,sand, and DPG mixtures under static compaction can be summarizedas follows:
1) The Abrams' law does not apply to mixtures subjected tostatic compaction.
2) The strength properties of portland cement, sand, and DPGmixtures can be significantly enhanced by increasing the
312
Figure 8 : Effect of static compaction pressure on compressivestrength as a function of soaking age.
313
Figure 9: Effect of cement content on compressiveas a function of soaking age.
strength
314
compaction pressure. Compressive as well as splittingtensile strength of the mixtures improves continuously withthe compaction pressures ranging from 750 to 12,000 psi.This strength improvement is attributed to the bindingcapability of DPG and cement, which increases withcompaction pressure.
3) To obtain a mixture of desired strength (for a given sandcontent), two approaches can be used. One way is to increasethe cement content while compacting the mixture at a lowpressure. The other is to increase the compaction pressurewhile using a low cement content. The second option is moredesirable, comparing the tremendous saving on cement costwith the limited increase of energy input for high-pressurecompaction. The mixture having 7.5% cement, 45% sand and47.5% DPG developed compressive strengths of over 4,000 psiunder air-dry conditions and 2,000 psi under soakedconditions when compacted at a pressure of 12,000 psi.
4) Mixtures with high DPG content are suitable for high-pressure compaction. At the compaction pressure of 24,000psi, the highest strength is obtained, for 7.5% cementcontent, when the mixture contains only cement and DPG (nosand).
5) Compaction pressure appears a factor affecting thedurability of portland cement-based mixtures usingphosphogypsum. Better durability may be expected with theapplication of high compaction pressures.
6) For mixtures with low C3A cement and appropriate amounts ofDPG, the strength is not affected by the extended soakingperiod (up to six months).
7) Portland cement-based mixtures using DPG have good potentialfor the precast structural products.
ACKNOWLEDGEMENT
This study was funded by a grant from the Florida Instituteof Phosphate Research, Bartow, Florida. The phosphogypsummaterial used in this study was provided by Gardinier Inc.
REFERENCES
Lecznar, F.J. and R.M. Barnoff. 1961. Strength of Neat CementPastes Moulded under Pressure. ACI Journal. Vol. 57. No. 8.3:973-975.
Lin, K.T., J.L. Figueroa, and W.F. Chang. 1985. EngineeringProperties of Phosphogypsum. Proceedings of the SecondWorkshop on By-Products of Phosphate Industries. FloridaInstitute of Phosphate Research. Bartow. FL. 11:49-59.
315
Lin, K.T. and W.F. Chang. 1986. Strength Properties ofCompacted Phosphogypsum-Based Mixtures. The SecondInternational Symposium on Phosphogypsum. Florida Instituteof Phosphate Research. Bartow. FL.
Lin, K.T., A. Nanni, and W.F. Chang. 1986. Compressive Strengthof Compacted Portland Cement-Based Mixtures UsingPhosphogypsum. Consolidation of Concrete. SP-96. AmericanConcrete Institute. 20:57-76.
Ouyang, C., A. Nanni, and W.F. Chang. 1986. Durability ofPortland Cement Mortar Containing Phosphogypsum. The SecondInternational Symposium on Phosphogypsum. Florida Instituteof Phosphate Research. Bartow. FL.
Skalny, J.P. and A. Bajza. 1970. Properties of Cement PastesPrepared by High-Pressure Compaction. ACI Journal. Vol. 67.No. 3. 7:221-227.
Wirsching, F. 1984. Drying and Agglomeration of Flue GasGypsum. Proceedings of the Symposium on the Technology ofGypsum and Gypsum Products. Kuntze, R.A. Editor. ASTMSTP861. 13:160-172.
316
CORROSION OF REINFORCEMENT IN PHOSPHOGYPSUM-BASED CONCRETE
Wen F. ChangProfessor and Director, Phosphate Research Institute
Department of Civil & Architectural EngineeringUniversity of Miami, Coral Gables, Florida
ABSTRACT
Thisbased
study investigated the effectiveness of phosphogypsum-concrete in corrosion protection of reinforcement. Test
results indicated that PH value increased rapidly whenphosphogypsum-mixtures contained a small amount Of cement.Consolidation by static compaction improvedconcrete as well as permeability of specimens.
the quality of
thickness of the coverIncreasing the
also provided an effective way ofprotecting reinforcement against corrosion.
INTRODUCTION
The objective of this theeffectiveness
study is toof phosphogypsum-based
investigateconcrete in corrosion
protection of reinforcement. Accelerated corrosion tests wereused to determine the effectiveness ofconcrete
the phosphogypsum-basedcover for permanent corrosion
reinforcement.protection of
It wasalone, without
found that prescribing the thicknessspecifying the quality of concrete can be
misleading. The compressivespecifying concrete quality,
strength usedis necessary,
conventionally for
requiredbut not sufficiently
for the determination of concrete quality for corrosionprotection. Permeability numbers to liquids can be usedeffectively as indices.
MECHANISM OF CORROSION PROTECTION
The protection afforded by the phosphogypsum-based concretecover against most types of corrosion is the same as theconventional concrete,the
partially due to its high alkalinity. Ifhydrogen ion concentration (PH) is between 9 and 13 in the
absence of oxygen,larger than 12,
no corrosion takes place. If the PH value is
concrete cover.total inhibition of corrosion is afforded by the
In addition to providing the alkaline condition at thesurface of the reinforcing steel, the phosphogypsum-basedconcrete cover also acts as a barrier which minimizes or preventsthe access of chemically aggressive substances to thealkaline film. Furthermore,
protectivethe cover can provide a secondary
317
inhibitionto the
of corrosion by preventing the supply of free"corrosion-cell".
oxygen
The quality of concrete, which depends on numerousparameters, can be considered as one of the morefactors influencing the barrier-effect of
importantthe cover. It is
important to note that the quality of concrete and the concretepermeability to liquids are affected by the same parameters.Since the degree of impermeability of concrete plays an importantpart in prohibiting corrosion of embedded steel, it is evidentthat permeability numbers can be used as an index for quality.
up to a certain degree, the increase of cover thicknessincreases the corrosion protection of the reinforcement.Furthermore, the permeability of the concrete covermarkedly on its thickness.
dependsIn addition to these functions, the
thickness of the cover influences the progress of penetration ofcarbonation.
PARAMETERS INFLUENCING PERMEABILITY
Due to the heterogeneous internal structure, concretecontains pores and invisible microcracks existing at theaggregate-cement paste and steel-cement paste interface, whichdirectly control its, physical and mechanical properties.Permeability is generally low for a well-compacted, well-cured,and high-quality concrete due to low volume of its pore-system.
Factors influencing the permeability of phosphogypsum-basedconcrete investigated, were : (1) cement content, (2)phosphogypsum content, (3) moisture content, (4) thickness ofcover, and (5) consolidation techniques.
, TEST PROCEDURES
One of the objectives of the investigation was to correlatethe quality of concrete and the thickness of cover with thedegree of corrosion protection afforded in a highly corrosiveenvironment.
Compressive strength and permeability numbers were used asthe criteria in determining the quality of phosphogypsum-basedconcrete. Permeability numbers of various types of concretecover were determined by falling head method. The cylindricalspecimens of 3" in diameter and of l/2", 3/4", l-1/2" and 2"thicknesses were sealed at the curve surface with paraffin. Theinlet water as a function of time, was determined by using agraduated plexiglass pipe.
In order to determine the effectiveness of different typesof phosphogypsum-based concrete cover concerning the corrosionprotection of the steel, accelerated corrosion tests were used.The corrosion solution consisting of 3.5 percent of NaCl solution
318
(by weight) was chosen, since chloride is able to destroy theanti-corrosive mechanisms of concrete cover more rapidly thanother environments. Specimens with known mechanical propertieshaving various thicknesses of concrete cover were totallysubmerged for 5 days in the highly corrosive NaCl solution. Thespecimens were exposed to air for one day, followed by a totalsubmergence in the solution for one day.wetting and drying,
This type of periodicwhich had been found by others to be the most
critical in the case of chloride corrosion, was usedfor 65 days by these corrosion tests.
repeatedlyTo measure the degree of
corrosion, as defined by the depth of the corrosion pits, visualobservation was used.
To measure the PH values,was pulverized.
concrete taken from the specimensTwenty-five grams of pulverized concrete was
diluted in 150 ml of distilled water, and the PH values of thephosphogypsum-based concrete in a slurry form was measured usinga PH meter. The first reading was obtained two days after theformation of the slurry, while the final reading was taken fourdays later.small.
The difference between the two readings was very
EXPERIMENTAL RESULTS
PH VALUE
The high alkalinity of the phosphogypsum-based concrete isprimarily responsible for the corrosion protection of thereinforcing steel. The PH values of phosphogypsum and cementmixtures increased from 3 to 9 when 3% of cement was used in themixtures and the increment became very slow when cement in themixtures was over 10% as shown in Figure 1. Similar testresults were obtained for the phosphogypsum sand and cementmixtures as shown in Figure 2. PH value over 9, in general, isconsidered adequate for corrosion protection of reinforcement.
PERMEABILITY
The quality of concrete depends on its pore structure.Thus, measurements of permeability is highly important to revealthe quality of the hardened concrete. Optimum water content forspecimens consolidated by vibration, is generally higher thanthat of the static compaction method. Therefore, consolidationby static compaction improves quality of concrete and decreasespermeability as compared to that of vibration method.
Figure 3 shows the relationship between permeability numberand cement content for a given phosphogypsum content of 25%.Permeability of specimens consolidated by static compactionresulted in much less permeability than that of vibration methodwhen cement content is less than 10%. Permeability of thespecimens decreased as the cement content increased.
319
Mixture Number
Figure 1
320
Content of Phosphogypsum (%)
Figure 2
321
Test results shown in Figure 4, indicate that permeabilitydecreased as phosphogypsum content increased for specimens with3/4" thickness and cement content of 20% and consolidated by thevibration method.
Water-cement ratio is highly important for quality controlof concrete. Sufficient amounts of water must be provided in themixtures for hydration and workability. Excessive amounts ofwater in the mixtures causes high porosity in the hardenedconcrete which deteriorates the quality of concrete. Figure 5shows that permeability of specimen increased as the moisturecontent increased.
CONCRETE COVER
The increase of cover thickness provides better protectionfor reinforcement against corrosion. Figure 6 shows thatpermeability decreased as the thickness of specimens increased.
Accelerated corrosion tests indicate-d that phosphogypsum-based concrete cover of l-1/2" thickness was capable of providingcomplete corrosion protection for all specimens tested.Furthermore, by using a high cement content (20% or more) in themixture, concrete cover of 3/4" had also prevented thereinforcement from corrosion;
ACKNOWLEDGEMENT
This research was funded by a grant from the FloridaInstitute of Phosphate Research, Bartow, Florida.
323
Phosphogypsum Content
Figure 4
324
Moisture Content (%)
Figure 5
325
Thickness of Cover (in)
Figure 6
326
CONSTRUCTION
GYPSUM AGGREGATE - A VIABLE COMMERCIAL VENTURE
NEIL R. ANDERSON
MOBIL MINING AND MINERALS COMPANY
PRESENTED AT:
THE SECOND INTERNATIONAL SYMPOSIUM ON PHOSPHOGYPSUM
DECEMBER 1OTH - 12TH, 1986
329
The Mobil Gypsum Project is a viable commercial venture of Mobil Mining and Mineral's,
Pasadena, Texasplant. A full time staff of sales representatives and production and
field installation personnel conduct the business of mining, blending, and marketing
phosphogypsum for use as a construction material.
Phosphogypsum is a by-product of fertilizer production and until recently has been
unused and a storage problem. About 400 million tons are now stockpiled in the U.S.
alone. Phosphogypsum is basically calcium sulfate dihydrate. Through an extensive R
&D program, Mobil has developed and proven construction uses for the material. Since
the first commercial sale in June of 1984, over 300 projects utilizing 340,000 tons of
gypsum were installed through December, 1986.
Phosphogypsum is sold as Gypsum Aggregate, a construction material consisting of
phosphogypsum, screened to minus 1", blended with 6% by dry weight, Type II Portland
cement. As a stabilized base material, Gypsum Aggregate is blended with cement and
water in Mobil's pugmill. Gypsum Aggregate is then placed on a prepared subgrade,
usually compacted, and in the Houston area, often lime or cement stabilized. The base
is spread and then field compacted with a pneumatic roller to density. Generallyan
asphaltwear surface is thenapplied, such as a surface treatment or hot mix asphalt.
Gypsum Aggregate base material has been sold to all seven counties in the Houston,
Texas area and numerous municipalities. It was approved for a Texas Department of
Highways State project which was awarded in December, 1986. Gypsum Aggregate has
used extensively for private construction. Projects using the material include
330
several Wal-Mart store parking lots, container storage yards, and many other varied
commercial installations. In addition to base applications, Gypsum Aggregate is used
for utility bedding, replacing stabilized sand, as well as for railroad base, stable
fill, embankment construction, and waste stabilization.
SALES PROJECT BACKGROUND AND ECONOMIC FACTORS
Mobil's Pasadena, Texas diammonium phosphate fertilizer plant is located on the
Houston Ship Channel. There are approximately 25,000,000 tons of gypsum in storage at
the site with limited space for future storage. Shortly after acquiring the plant in
1979, Mobil began a research and development program investigating alternative means
of disposal or optimally use of phosphogypsum.
The research and development program quickly focused on the use of gypsum as a
construction material. An extensive program of research on this application was
begun with the basic research being performed at Texas A&M University at College
Station, Texas. The initial work concentrated on the performance characteristics of
phosphogypsum, from Mobil storage, stabilized with fly ash or Portland cement. The
work was performed by Chuck Gregory under the direction and supervision of Dr.'s
William Ledbetter and Don Saylak of the Texas Transportation Institute of Texas A & M
University.
To what extent the material would respond to stabilization was one of the initial
questions, as well as questions regarding its moisture density relationship, strength
development with age, and response to the mount of stabilization. The first major
documentation of the research work was a masters thesis by Gregory, "Enhancement of
Phosphogypsum with High Lime Fly Ash", dated 5/83. The paper confirmed, as have other
331
A&M studies and other independent research, that gypsum does respond well to
stabilization and that it forms a stable and strong matrix when stabilized with
reasonable amounts of cement or fly ash.
No matter how the material might respond to stabilization, without a market driven by
favorable economics and a need for the material, a commercial venture would not be
possible. Market research in the Houston area indicated a special situation in the
aggregate market. The Houston area is aggregate poor with limited local material
supplies. Sand is locally available, but stone aggregates are transported into the
Houston area by unit trains from quarries located far to the west of Houston. This
high cost supply provides a distinct economic advantage for a base material located in
Houston. Prices for crushed limestone for base applications currently range from
approximately $10.50 to $14.00 per ton, delivered in the Houston area.
The market volume
extremely large.
industrial base.
counties. While
extensive State,
base continue to
economics of the
for base and other construction materials in the Houston area is
Houston is the 4th largest city in the nation with a strong, heavy
The greater Houston area includes seven populous and growing
the slump in oil prices has been detrimental to the Houston economy,
Federal, and County road projects and a strong petrochemical industry
provide a large market for aggregates. The volume and favorable
market led Mobil to the decision to pursue commercial development of
the gypsum project.
Field trials of the material were begun in 1983. A short section of haul road and an
industrial storage yard were installed in early 1983. Both of these projects continue
to provide good service. In June of 1983, the first extensive controlled and
monitored field trial was installed. For this project, Mobil provided material on a
332
test basis to the city of La Porte, Texas. Working closely with Texas A & M University
and the Houston area geotechnical firm of McBride-Ratcliff, Inc., Mobil installed test
sections of stabilized gypsum base along with control sections of crushed limestone
limestone sections. The streets are performing very well, and the analytical results
will be discussed in Section IV. Based on the excellent performance of stabilized
gypsum base in the field trials a commercial sales effort was
commercial sales were made in June 1984.
begun and the first
REGULATORY HISTORY/CURRENT STATUS
As part of the research effort involved in the development of Gypsum Aggregate, a
comprehensive environmental assessment was made which included lab and fielddata
collection. The testing included coring and radiological characterization of the
storage pile, radon emanation measurements from stabilized gypsum, leachate analysis,
airborne dust measurements, and field gamma measurements. This data was reviewed and
analyzed by health physicists as the basis of Mobil's environmental assessment.
It was determined early in the project that it would be advisable to seek formal
approval from appropriate regulatory agencies for the sale and use of phosphogypsum,
even though the requirement for regulation was unclear. Accordingly, Mobil made
contact with the Texas Department of Health, Bureau of Radiation Control, and
following some deliberation, they suggested that we apply for a license. The license
application was submitted along with our environmental data and analysis. The BRC
performed an environmental assessment which concluded "that the use of phosphogypsum,
as described in submitted documents and regulated under the TRCR by the proposed
333
license conditions, will not endanger the public health and safety". The license was
granted on 4/26/83. It allows Mobil to sell phosphogypsum for uses including road,
parking lot, and storage yard base, utility bedding, construction of dikes,
construction of embankments, railroad base, and waste stabilization.
The license requires that Mobil provide a quarterly report detailing the location of
installation sites. In addition, Mobil is required to have the purchaser sign a
statement agreeing to cover the material so that it will not be subject to
dispersion. The Bureau of Radiation Control has indicated that an chip and
seal surface is considered sufficient cover to meet the terms of the license. This
analysis and radiological particulate and radon measurments. Note that a gypsum pile
consists not only of phosphogypsum, but of the pond water, or carrier water retained
in the pile. This water is acidic and high in fluoride. Drained, leached,
phosphogypsum is a much more innocuous material. For example, the pH on the Mobil
pile now being reclaimed is 6.0 - 6.5. The conclusions and any follow up regulatory
action drawn from the EPA study will have an effect on commercial utilization of
gypsum. However, any regulation of gypsum stacks including construction practices and
any other mitigation of ground water effects would not necessarily have a direct
regulatory effect on the sales of recovered gypsum. Dealing with a processed product
as opposed to a waste is a different regulatory matter entirely. At any rate, the
initial data analysis of the EPA studies indicates fairly positive results and
extensive regulation would not appear to be justified.
334
MATERIAL PERFORMANCE CHARACTERISTICS/R& D
Gypsum, when stabilized with cement, participates in the cement hydration reaction to
form a stable durable crystal matrix that provides significant strength as a
construction material. Pictures am attached as Fig. 11 of raw gypsum crystals
(1000X) and 7 day old cement/gypsum matrix (1500X). The hydration crystals are
visible in the matrix picture.
A number of studies have been completed on the performance characteristics of Gypsum
Aggregate and Mobil is sponsoring several ongoing research projects. A few of the
studies sponsored to date by Mobil are summarized below:
PERFORMANCE VARIATION WITH PILE DEPTH, REF. 2.
Representative data from this study is shown in Figures 2 and 3. The study involved a
comparison of unconfined compressive strength for phosphogypsum material taken from
three different levels in Mobil's gypsum stack. Samples were taken at the surface,
from 5 to 10' down, and from 10 to 15' down in the pile. Test cylinders with varying
cement contents were made and cured. Samples were broken for compressive strength
analysis at and 28 days. The data in the figures is partial as additional
conditions, including moisture and density variations, were measured. The data
presented is however, illustrative of the fact that there was no significant
performance difference between the three mining levels. There had been some concern
that as we mined deeper into the pile, into areas that had been less weathered, that
we might see degradation of product performance. This was shown not to be the case,
in fact we have mined 25' into the pile in some areas, and find that performance of
the product remains uniformly good.
335
STRENGTH AND UTILITY BEDDING, REF.3
o obtain significant strength Gypsum Aggregate requires compactive effort to bring
he fine gypsum crystals and cement intosufficiently intimatecontact. Frequently,
n utility bedding applications, bedding material is dumped loose into a trench and
does not received significant compaction. In some cases, compaction is not specified
nd in other cases it may be specified but it's simply not done in the field. In ah
ffort to determine if trench placed uncompacted Gypsum Aggregate would develop
ignificant strength, a field trial was made at the Mobil plant. A 16" pipe was
edded with four test sections of Gypsum Aggregate and stabilized sand. In each case,
"of bedding material was placed in the trench, the pipe was placed and approximately
'of stabilized material was placed above the pipe.
he test sectionswere:
section l - Gypsum Aggregate dumped in, not compacted.
ection 2 - Stabilized sand dumped in, not compacted.
ection 3 - Gypsum Aggregate placed with minimal "jumping jack" compaction.
ection 4 - Stabilized sand placed with minimal "jumping jack" compaction.
after two weeks the sections were excavated and CBR's were measured in the field, at
he top of the stabilized material. Then the stable backfill and pipe were removed
and the bedding sections were tested. The results of the tests are presented in
Figure 4, they show that Gypsum Aggregate develops significant strength even when
simply dumped in place, with its lowest CBR's being 8.5 -16. This compares to lime
stabilized, compacted fill which might have a CBR of approximately 10, or a CBR of 4
or stiff to very stiff natural clays.
339
CBR results for compacted gypsum were comparable to the upper layers of the stabilized
sand. THE Gypsum Aggregate numbers were on the whole, however, lower than the
stabilized sand. Since the time of this study, two significant changes have been made
in the product produced by Mobil for this application. First, the cement content has
been increased from 5 to 7 1/2% and second, the cement type has been changed to a Type
resistance and additional long term strength, versus Type I, for phosphogypsum
stabilization.
FIELD STUDY OF MUNICIPAL STREETS - REFS. 6, 7, 8, 9, 10
As previously mentioned, the major controlled field study of the performance of Gypsum
Aggregate involved the installation of several city streets for the City of La Porte,
Texas. Construction of the streets involved 6" lime stabilized subgrade, followed by
8" of baseI and a two course surface treatment wearing surface. There were several
different test sections which included crushed limestone base as control sections,
phosphogypsum sections of 5, 7 l/2, and 10% Portland Cement and 15 and 25% flyash.
The streets have been periodically measured as part of a 4 year follow-up study by The
Texas Transportation Institute of A & M. The effective thicknesses of the base
sections as evaluated from dynaflect measurements based on a composite modulus of
500,000 psi are given in Figure 5 for the various sections. Note that in all cases
the installed thickness was 8". The 2 and 3 year data are given, and as can be seen
from the table, the effective thickness of the Gypsum Aggregate sections significantly
exceeds that of the crushed limestone control sections.
To quote from the A&M follow-up report, Reference 10, "The phosphogypsum sections
have maintained their load spreading capability and are good base materials. No signs
of degradation appear."
Although the acid content of stored phosphogypsum is greatly reduced when it is
drained and weathered, with a 4.5 to 5.8 pH it remains slightly acidic. The addition
of cement, which is quite basic, neutralizes the residual acidity, and the pH of the
material when it is shipped with 6% cement is about 9. There is potential application
for minimally stabilized Gypsum Aggregate which would be incontact with galvanized
steel reinforcing rods. To find out the amount of cement that would be required to
have a corrosion rate comparable to that of cement stabilized sand, the referenced
study was performed.
Some of the data through 6 months of testing is presented in Figures 6, 7, and 8.
Figure 6 shows the corrosion rate in mills per year of galvanized steel for flyash
stabilized and cement stabilized phosphogypsum. The corrosion for flyash is shown to
Figure 7 shows corrosion rates for varying cement contents of phosphogypsum 3, 4, 5,
and 6% cement. As can be seen, the corrosion rate is reduced as the cement content is
increased.
Figure 8, compares the corrosion rates of 5 and 6% cement gypsum with two different
Houston area sands. The corrosion rate of the gypsum is seen to be comparable to the
sand with exposure data up to 192 days.
343
STABILITY OF
Since Gypsum
CONCRETE EXPOSED TO GYPSUM AGGREGATE, REF. 4
Aggregate is primarily calcium sulfate dihydrate the question of
potential sulfate attack of concrete exposed to gypsum was addressed. To definitively
answerthis question, al year study was conductedby Construction Technology
Laboratories, a Division of the Portland Cement Association. Concrete specimens with
Types I, II, V, and Type V with pozzolan were exposed for 12 months to a saturated 5%
cement 95% phosphogypsum bedding mixture and, as a control, a saturated sand cement
mixture. The results are summarized in Figure 9, and the petrographic analysis shows
that Gypsum Aggregate bedding had no effect on the concrete specimens. As an
additional part of the study, a culvert exposed to gypsum former one year under
field conditions at the Mobil plant was cored and subjected to
by CTL. Again, the microscopic examination showed no signs of
The average percent length change and percent weight change of
exposed to sand, and those exposed to gypsum, were found to be
petrographic analysis
sulfate attack.
the concrete cylinders
comparable. The
flexural strength of the concrete exposed to the gypsum actually averaged a little
higher at 1,327 psi vs. 1,120 psi for the concrete cylinders exposed to the sand
bedding.
Certain projects that were completed in late 1985 using Gypsum Aggregate base
developed surface ridges and cracking patterns indicative of expansion or swelling of
the base material. An investigation of the causes of this phenomenon revealed that
the cement used for these projects was significantly higher in tricalcium aluminate
than what had been used in fully successful projects.
347
Several studies were conducted and are representative figure from one of the studies is
attached as Fig. 10. It is data from a parametric study (Ref. 5) evaluating volume
expansion with time of 94% gypsum, 6% cement cylinders for varying cement types. The
cements types, identified in order of increasing C3A content, are IIB, II, I, III,
and III Special. C3A contents varied from the low of about 5.5% for the Type IIB,
up to the high for the Type III Special of 11.8%. As can be seen from the graph,
there is a definite reduction in expansion with a lower C3A content in the cement.
Several different curing modes were evaluated as part of the study. The moist cure
provided the most dramatic difference in expansion as a function of cement type and
this is the data presented in Fig. 10.
COMMERCIAL APPLICATIONS
As previously mentioned, commercial sales of Gypsum Aggregate began in June of 1984.
Developingsales involved identifying markets, identifying users and decisions makers,
and identifying specific sales contacts. Promotional efforts included development of
a brochure and direct mail programs.
A State of Texas, Department of Highways project was put out forbid in December of
1986 which includes Gypsum Aggregate in the specifications. This will be the first
State project to use Gypsum Aggregate. As previously discussed, the material has been
used by many cities, and in projects in seven counties. The five case histories
attached as Fig. 1, are representatives of some of the projects using Gypsum
Aggregate. These projects are some of the earliest commercial projects utilizing
Gypsum Aggregate. They all continue to perform well, and are providing good service.
349
350
Case Historie Figure 1.
Mobil Gypsum Aggregate
LaPorte Street10 months service
Trailer Park9 months service
Location:Application:
Installed:Installed By:
Stabilized With:
Service/Results:
Location:
Application:Installed:
Installed By:Stabilized With:
Service/Results:
City of La Porte, TexasBase for General Residential StreetsJune, 1983B W F General Contractors, Inc.Various sections: 5%, 7.5%, 10%cement and 15%, 25% Flyash.After continuous service, includingflooding by Hurricane Alicia,streets remain in excellent condi-tion and are performing as well orbetter than adjacent limestonebase control sections.
Tradewind Trailer Park, La Porte,TexasHouse Trailer Sales Lot BaseJ u l y , 1 9 8 3B W F General Contractors, Inc.6% CementLot used as display and sales areafor house trailers and as parkingarea for customer vehicles. Trailersare moved on and off pad whenreceived or sold. Lot remains inexcellent condition.
Amburn Roadlocation: City of Texas City, Texas. Application: Base for GeneralResidential Street. Installed: September, 1984. Installed by: City of TexasCity Public Works Department. Stabilized with: 5% Cement.Service-Results: Amburn Road provides access to residential areas southof FM 1764 as well as being the major access to the College of theMainland. The road has an asphalt surface with concrete curb and gutter.
Lazy Lanelocation: City of La Marque, Texas. Application: Base for GeneralResidential Street. Installed: July, 1984. Installed By: City of La MarquePublic Works Department. Stabilized with: 7% Cement. Service-Results:The road has an asphalt surface with concrete curb and gutter. The roadbase has provided equal or superior performance compared to thelimestone or stabilized sand base on adjacent streets.
Wal-Mart Shopping Centerlocation: Fry Road - Houston, Texas. Application: Base for ShoppingCenter Parking Lot. Installed: September, 1984. Installed by: ForceCorporation. Stabilized with: 7% Cement. Service-Results: The parkinglot serves a Wal-Mart store as well as a group of other stores in thesurrounding shopping center. The Gypsum Aggregate base providedsubstantial savings over limestone.
351
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
Mobil Research and Development Corporation, Princeton, NJ, "Cor.and Materials Study-Phosphogypsum", 515-02, 10/22/86.
McBride-R&cliff and Associates, Houston, TX, "Stabilization Re:of Gypsum from Three Mining Levels", 85-034, 3/22/85.
McBride-Ratcliff and Associates, Houston, TX, "Field CBR Test-TBackfill", 85-034, 9/03/85.
Construction Technology Laboratories, Division of The PortlandAssociation, Skokie, IL, "Effects of Phospho-Gypsum/Cement BeddMixtures on Durability of Concrete", 2/86.
McBride-Ratcliff and Associates -Houston, TX, "Cement Type ParStudy- Interim Report", 10/14/86.
Little, D. N., Texas Transportation Institute, Texas A&M Univ7/07/86 Correspondence.
McBride-Ratcliff and Associates, Houston, TX, "Core Samples and CBRTest Results, La Porte Test Section", 9/30/83.
Gregory, C. A., Texas A &M University, "Enhancement of Phospho-GypsumWith High Lime Fly Ash", Masters Thesis, 5/83.
Gregory, C. A., Ledbetter, W. B., and Saylak, D., Texas TranspcInstitute, Texas A&M University, "Construction and InitialPerformance Evaluation of Stabilized Phospho-Gypsum Test SitesLa Porte, Texas", 5/84.
10. Little, D. N., Texas Transportation Institute, Texas A&M Univ"Dynaflect Deflection Testing & Core Strength Data", 10/08/85.
352
POLK COUNTY EXPERIMENTAL ROAD
William C. KenleyCounty Engineer
Division of Public WorksPolk County, Florida
Wen F. ChangProfessor and Director
Phosphate Research InstituteDepartment of Civil and Architectural Engineering
University of MiamiCoral Gables, Florida
ABSTRACT
An experimental secondary road utilizing phosphogypsummixtures was successfully built in Fort Meade, Florida by thePolk County Division of Public Works in October, 1986. The roadas constructed indicates that phosphogypsum mixtures are not onlyeasier to work with but also provides higher stability than thatof clay mixtures.
INTRODUCTION
The experimental project consists of the construction of oneand one-half miles of secondary road in Polk County, Floridausing phosphogypsum mixtures. The Parrish Road located one mileeast Of Fort Meade and South of U S 98 was selected for theexperiment because the road was convenient to the supply ofphosphogypsum and easily accessed from U S 98.
The experimental project is to provide alternate methods ofrebuilding county and other secondary roads in Florida. It isintended to provide comparable or better material to repair orreplace existing roads with the best possible utilization oflocally available aggregates.
Current construction practice in building these secondaryroads consists of mixing generally granular soil subgrade withfine grained soils transported to the site. Granular soil suchas sand is abundant throughout Florida. However, lack ofadequate source of fine-grained soils such as clay, has been amajor concern of builders of such roads. Futhermore, it has beenfound that roads built with clay-sand mixtures tend to be greatlyaffected by changes in moisture regime which tends to be soft andmuddy during the long and rainy. summer sessions. This hasprompted the Polk County Division of Public Works to take theinitiative in finding 'alternate methods of rebuilding theirroads.
353
In Central Florida, over, 450 million tons of phosphogypsumhave been stockpiled, and the total tonnage may approach onebillion tons by the year 2000. Phosphogypsum when subjected tocompaction can be transformed into solid of valuable strength. Itcan also be used effectively as a binder in road construction toreplace limerock and clay, which is in short supply in many partsof the State of Florida.
The design and testing of the road was a collective effortof the University of Miami, the Florida Department ofTransportation and Polk County. The road was built by the PolkCounty Division of Public Works in the Fall of 1986, andphosphogypsum used for the project was supplied by the USS-AgriChemicals. The experimental project calls for a thoroughenvironmental impact investigation which includes the pre andpost construction sampling of air, soil and groundwater includingdrinking water. Environmental monitoring as described has beenconducted by the University of Miami, in cooperation with theFlorida Department of Environmental Regulation and the FloridaDepartment of Health and Rehabilitative Services.
GENERAL CONSTRUCTION PROCEDURE
The existing road surface layer is levelled with a motorgrader and compacted with a steel drum roller. Phosphogypsum atits natural moisture content is delivered by means of dumptrucks. The phosphogypsum is discharged at the construction siteand evenly spread by means of a bull-dozer and a motor grader tomeet the appropriate lift thickness. A pulvimixer is brought tothe 'site to thoroughly mix the phosphogypsum with the subgradematerial; The depth of the mixed layer is selected according to
the mix constituent proportions called for in the designspecifications. The importance of pulverizing the mixtureconstituents and thoroughly blended them should be emphasized.Final results depend on mix uniformity and moisture content.Following the mixing phase, the road cross-section profile isshaped according to the design drawings. Slopping of the roadsurface is important to avoid rainwater ponding on the pavement.
The road is then compacted by a steel roller and apheumatic tire compactor. The higher the compaction effort thebetter; however the weight component is the most importantbecause of the materials fineness,, The pneumatic tire compactorhas the functions of compacting any low spots, not properlycompacted by the steel roller, and of smoothing the top surface.
ROAD DESIGN AND CONSTRUCTION
The experimental road is shown in Figure 1. On September12, 1986, 3" of phosphogypsum was spread on the first section ofthe road, Station 90+00 to 100+00 and-mixed with a pulvimixer toa depth of 12". The mixture was compacted' with a steelwheel
354
roller and a pneumatic tire roller, and opened to traffic forseveral days. A second 3" of phosphogypsum was then spread on thestabilized road base and mixed to a depth of approximately 10".The mixture was again compacted. This is the procedure that wouldhave been used with clay as a subgrade stabilizer.
The next section of the road, Station 66+00 to 90+00, wasconstructed by adding a single application of 3" of phosphogypsumto the existing road and mixing to a depth of 15". The sectionwas compacted using a sheepsfoot roller. This roller was applied until it rolled on the surface without penetrating. Then thepneumatic roller and steel wheel rollers were used to completethe compaction.
The third section, Station 25+00 to 66+00, was constructedby placing a single 6" lift of phosphogypsum on the existing roadand mixing to a depth of approximately 15". The sheepsfoot rollerwas again used for the first part of the compaction effort.Moistures of compaction were maintained at less than 10% and thefinal compaction was done using the steel wheel roller and thepneumatic tire roller.
Cutback asphalt RC-70 was applied at a rate 'Of 0.20qal/sq.yd. on the surface of the entire road. The section,Station 50+00 to l00+00 was then-covered with a 1 " asphaltsurface. The remainder section of the road was spread with alayer of fine sand.
TESTS ON MECHANICAL PROPERTIES
On-site mixtures were taken to the laboratories fordetermining the moisture density relationship and CaliforniaBearing Ratio. Results from tests conducted in accordance withASTM D1557 are shown in Figures 2 and 3.
A Nuclear Moisture Density Meter was used to determine theroad base density achieved after compaction. A Speedy MoistureTester was instead used for obtaining the moisture content. Themaximum density that could possibly be achieved by the availablecompacting equipment, was about 95% Modified Proctor.
The bearing capacity of the road base was also checked inaccordance with the California Bearing Ratio Test. The averageCBR value of the compacted on-site soil is about 17. By mixingwith phosphogypsum the CBR value of the stabilized bases variedfrom 40 to 150 depending on the moisture content in the base. CBRvalue decreased with the increased of moisture content in themixtures.
CONCLUSION
The road has been open to traffic since October, 1986. Theexperimental project has successfully demonstrated the use ofphosphogypsum in road construction. Evaluation of the
356
construction crew is as follows:
(1) Phosphogypsum can be used as a binder for base coursemixture.
(2)
(3)
Phosphogypsum mixtures are easier to work with thanclay mixtures.
Operation cost including equipment time forconstructing phosphogypsum roads are lower than that ofclay roads.
(4) Rain storms during construction did not cause excessivedelays because the compacted mixture did not absorbwater to any great extent.
(5)
(6)
Shrinkage cracks, frequently occurring in clay roads,did not appear in this construction.
The stability of compacted phosphogypsum mixtures issuperior to that of clay mixtures.
ACKNOWLEDGEMENTS
The project is funded by a grant from the Florida Instituteof Phosphate Research.
The cooperation of the Bureau of Materials and ResearchFlorida Department of Transportation is greatly appreciated.
359
Dr. David J. SumanthAssociate Professor and
Director, Productivity Research Group
Manuel. A. Carrillo, B.S.(I.E.)Graduate Student
Siva R. Koganti, B.S. (M.E.)Graduate Student
in theDepartment of Industrial Engineering
University of Miami
361
I. OBJECTIVE
This paper will address the problem of highway planning byincorporating the Total Productivity analysis to determine the economicfeasibility of road construction using phosphogypsum as the material forthe construction of the road base. Only through the application of thisanalysis, the joint impact of all cost elements on the road constructionproject can be accurately determined.
The field work included conducting a l-mile secondary road constructiontest with Phosphogypsum in Polk County, Florida. The two pavement designsas well as the general picture of the road to be constructed are shown inFigures II.1. II.2. II.3 and 11.4. The scope of this research is toeffectively evaluate the cost elements involved in this constructionproject.
The general economic evaluation methods have been studied by Ogelsby[1975], Ruenkrairergsa [1983], Soberman [1965], and Moyer and Lampe[1963]. The main goal of this review is to investigate the state-of-the-art on road construction cost standards and procedures. Some of thecosting methods used in road construction in general, and secondary roadin particular, are as follows:
1. Cost evaluation methods based on the size of rural roads ( Sharma etal. [1985] ).
2. Proportional and incremental methods of costing ( Villarreal-Cavazos,Garcia-Diaz [1985], Fwa and Sinha [1985] ).
3. Life cycle costing methods ( Sandler, Denham, and Trickey [1984] ).
4. Investment evaluation methods (Gomez-Ibanez and Lee [1983], Moyer andLampe [1963], Hammer-man [1980], and Puenkrairergsa [1983] ).
This review of literature has been conveniently summarized in Table . III.1 in the form of a "State-of-the-Art Matrix" (SAM).
Economic justification methods to compare alternatives are very muchthe traditional ones, and their use has been limited at best. What weattempt to accomplish in this project is to develop an alternativeapproach to Economic Justification in road construction. We believe thatsuch an approach will be universally applicable for all types of road
construction. This approach will be based on the concept of totalproductivity and will link economic analysis with productivity analysis.The next section presents a background in this perspective.
362
363
365
IV. METHODOLOGY USED
The theoretical background for our standardprovided by the Total Productivity Model TPM (Sumanthpp. 156-163 ). The main elements of the model are as
Total Productivity, as defined in the OperationalModel OTPM, is given by the following:
Costing method is[1984] , Chapter 8,follows:
Total Productivity
total tangible outputTotal productivity=---------- , where
total tangible input
Total tangible output= Value of finished miles constructed andTotal tangible input= Value of ( human + material + capital +
energy + other expense ) inputsused
The total productivity in period t is therefore in symbols:
Similarly, the Partial productivity is defined as :
where {j}= { Human,H, M, C, E, X }.
Construction workproductivity indexesPartial Productivity
Materials, Capital, Energy, and Other expense } = {
standards were obtained by calculating the partialbased on miles of road constructed, for example, theof labor, given by:
PP = ppH Human
366
Also, an overall road construction cost standard based on the totalproductivity index, is given by:
0 Miles ConstructedTP = -- = --------------------
I IH + IM + IC + IE + IX
Steps of the methodology:
The methodology used in this study is summarized in Fig. IV.1
1. From the discussions held with the Project Director and Projectinvestigators, and from the Flow Process Chart (Fig. IV.2), the OperationAnalysis is performed.
2. A data collection system was designed.
3. The Polk County engineer was communicated about the readiness of thecrews and the data collection personnel.
4. Field observations were made and the actual tasks list was prepared(Table IV.1). Then, using a Data Sheet (_Table IV.2, field data wascollected in a systematic manner through the joint efforts of a graduatestudent assistant, the Contractor Supervisor, and his personnel.
5. A computer program was developed for Economic Evaluation ( based onthe logic shown in Fig. IV.3. The field data was then processed with thiscomputer program. A sample calculation for task #3 is shown in TableIV.3.
v. RESULTS AND CONCLUSIONS
The numerical results of our study are summarized in Table V.1, andFigures V.1, V.2, V.3, V.4, V.5, and V.6.
In this study we have attempted to demonstrate the feasibility ofapplying a non-traditional approach ( Total Productivity Analysis ) toanalyzing the economic justification aspects with Phosphogypsum vs.traditional materials in an experimental road construction project. Thefindings, indicate the definite economic advantage in using Phosphogypsumas a base material in comparison with clay.
This study also showed that Productivity Researchers and Civilengineers have an opportunity to work together to make economicjustification decisions in a more rational and cost effective manner. Wehope, that this new thinking will stimulate further interest in economicanalysis of construction projects of any type.
367
372
3 7 3
This research was funded by a grant from the Florida Institute ofPhosphate Research, Bartow, Florida. Cur thanks are due to Dr. Wen Changof the University of Miami for providing the financial support and to Mr.William Kenley ( Polk County Engineer ) and his staff for their help ingathering the field data.
VII. REFERENCES
Fwa, T.F., and Sinha, K.C. (1985). "Thickness incremental method forallocating pavement construction costs in highway cost-allocationstudy". sportation Research Record, Vol:1009, pp. l-7.
Gomez-Ibanez, J.A., and Lee, D.B. (1983). "Economic evaluation of highwayinvestment needs". Transportation Research Record Vol:940, pp. 21-27.
Moyer, R.A., and Lampe, J.E. (1963). "A study of annual costs of flexibleand rigid payments for State highways in California". Highway ResearchRecord. Vol:77, pp. 133-172.
Ruenkrairergsa, T. (1983). "Economic analysis of soil-cement baseconstruction compared with crushed-rock base". Transportation ResearchRecord Vol:898, pp. 262-267.
Sandler, RD., Denham, E.T., and Trickey, J. (1984).economic
"Comparative.analysis of asphalt and concrete pavements". TransportationResearch Record. Vol:984, pp. 29-39.
Sharma, S.C., Tayebali, A., and Werner, A. (1985). "Cost-effective anduser- oriented sizing of rural roads". Transportation Research Record.
pp. 15-23.
Sober-man, R.M., (1965). "Economic analysisdeveloping countries".
of highway design inHighway Research Record, Vol:115, pp. 44-63.
Sumanth, D. J., (1984). Productivity Engineering and Management. New.York:McGraw-Hill.
Villarreal-Cavazos, A., and Garcia-Dim, A. (1985). "Development and.application of new highway cost-allocation procedures". TransportationResearch Record, Vol:1009, pp. 34-41.
383
BENCH MODEL PRODUCTION AND USES OFPHOSPHOGYPSUM BRICKS
Antonio Nanni, Asst. Prof., Dept. of Civil & Arch. Engrg.,Michael R. Swain, Asst. Prof., Dept. of Mech. Engrg.,
Bijan H. Ahmadi, Res. Asst., Dept. of Civil & Arch. Engrg.,and
Wen F. Chang Prof. and P.R.I. Dir.Dept. of Civil & Arch. Engrg.,
University of Miami, Coral Gables, Florida
ABSTRACT
A f t e r three years of laboratory research on theengineering properties of phosphogypsum-based mixtures, thefavorable results and the experience gained have reached thepoint where serious consideration can be given to commercialapplications of the by-product gypsum. In the buildingindustry, as separate from road construction, the mostimmediate commercial use of phosphogypsum rests in theproduction of masonry units or bricks.
This paper reports on the design and performance of abench model, semiautomatic press having a capacity of 400kips (1779.2 kN) constructed for the production of1.75x3.75x8 in. (45x95x203 mm) solid phosphogypsum bricks.
Because of the adopted construction method, the brickscan be handled immediately after fabrication. Theirdimensional precision and stability 'lend to dry stackingapplications or to the use of joining materials other thanthe conventional hydraulic cement mortar. This paper brieflysurveys some of the possible products which could besuitable for this purpose.
385
INTRODUCTION
Basic research investigations (Lin, et al., 1986 andLin, et al., 1987) have proven that phosphogypsum subjectedto high compaction pressure can attain considerablecompressive strength. These studies have indicated that thecompaction force can be applied dynamically (dropping weightor vibration roller) for road applications or statically(action of a press) for building products.
The industrial production of phosphogypsum elements canexploit the gypsum binding property. Among all applications,bricks seem to be the most promising product because theirlimited dimensions do not require unusual size equipment toobtain the desired degree of consolidation. For thispurpose, a bench model semiautomatic molding press wasdesigned and constructed at the University of Miami. Thebrick produced by this machine are used for basic researchon material properties and for the construction of prototypemasonry structures to study wall erection techniques andlong term behavior.
The aim of this paper is to report on three classes ofbricks produced with the molding press and to brieflydescribe the production process. In addition, a survey ofjoining materials potentially valid to substitute thetraditional cementitious mortar used in masonry constructionis presented.
BENCH MODEL PRESS
The scheme of the completed press is presented in thediagrams of Figures 1 through 4. The following operationsare performed for the production of one brick:
a. From a hopper mounted on the upper portion of thepress, material at the preset grain size distribution,constituent proportions and moisture content, is fedinto the mold by the combined action of gravity and airjets mounted at the side of the mold (Figure 1).The loose material fully occupies the space availablein the mold (volumetric fill). At this' point. the airjets are closed and a horizontal door between hopperand mold slides forward to impede further materialsupply, trim the top surface and provide confinement tothe upper side of the mold (Figure 2).The ram is raised to compact the material in the moldin one cycle at the predetermined maximum pressure(Figure 2) and, immediately after, is slightly droppedto release any pressure from the confined material.The horizontal door continues its forward motion tofree the top surface of the newly formed brick while
b.
C.
d.
386
FIGURE 2: BENCH MODEL PRESS. FEEDING INTERRUPTIONAND MATERIAL COMPACTION
387
maintaining the supply of material shut. At this pointthe ram fully extends to extrude the brick (Figure 3).
e. The horizontal door slides then backwards to push thebrick out of the press and to resume the materialfeeding. The ram returns to its lower position (Figure4) l
f. The air jets mounted on the mold side are started andthe production cycle is repeated.
Experimentation with the partially completed benchmodel press has provided valuable information not only onautomated production technology but also on conditions andproperties of the material to be molded. Various constituentproportions, moisture contents and compaction pressures weretested. Other variables such as cleanness of the mold,repeated loading cycles, duration of load application andmaterial gradation were also investigated.
During the fabrication of the bricks, independently of'the phosphogypsum type, it was found that there is an upperlimit of the free moisture content at which the brick can befabricated. In fact, for moisture values higher than thethreshold, the wide surface of the brick visibly cracks andheaves immediately after the compaction pressure isreleased. This observation is particularly important becauseparallel work on smaller size specimens (Lin, et al., 1987)simply indicates that excessive mixing water seeps out,dry-density is approximately constant and strength drops.This behavior can now be explained by the negative effectthat. the water outward migration has on the materialstructure. In conclusion,' excessive water is detrimental instatic compaction as it is in dynamic compaction.
EXPERIMENTAL PROGRAM
This experimental investigation aimed to provideinformation on the strength characteristics of 100%phosphogypsum bricks. For this purpose, three types ofmaterial were considered. The first two types, produced bythe same dihydrate process plant and referred to as type Iand II, were respectively collected from a fresh pile andfrom an old pile which included sandy deposits. The thirdtype was originated by the hemihydrate process.
Specimens with different moisture contents werefabricated with the bench model press at a 12,000 psi (82.7MPa) compaction pressure, delivered in one cycle andmaintained for a fraction of a second.
The testing program herein reported was limited to twoof the tests proposed by ASTM C67-83 "Sampling and TestingBrick and Structural Clay Tile". These tests measured thecompressive strength (half-brick) and flexural strength
389
(central load on a 7 in.= 178 mm free span), and wereconducted on dried samples as well as immediately afterproduction and under soaked condition (two-day waterimmersion).
TEST RESULTS
The compressive strength and the modulus of rupture(MOR) of bricks made of phosphogypsum type I and testedunder oven-dry condition is presented in Figure 5 as afunction of the free moisture content. It is observed thatthere is an optimum water content range between 2 and 4% forwhich the compressive strength is approximately 5000 psi(34.4 MPa) and the MOR is 1150 psi (7.9 MPa). These valuesare remarkable,, in particular the flexural strength.
The compressive strength of the same phosphogypsum typeI bricks under different testing conditions is reported inFigure 6. Two considerations can be derived from thisfigure, first, the compressive strength under soakedcondition is high, approximately 2500 psi (17.2 MPa), and,second, the performance under soaked condition or with themoisture content at the time of fabrication is practicallyidentical.
The presence of water in the sample strongly affectsthe tensile resistance as indicated in Figure 7, where theMOR is shown under the three testing conditions. Whenmoisture is present in the brick the flexural strength dropsan order of magnitude below the level obtained for theoven-dry state.
The compressive strength of phosphogypsum type IIbricks is shown in Figure 8 as a function of the freemoisture content for two different curing conditions. Thecompressive strength of this type of phosphogypsum isinferior to that of the previous one for the presence ofsandy particles. Subjecting the bricks to one wet and drycycle does not appreciably decrease strengthcharacteristics.
The dry-density of bricks made of both phosphogypsumtype I and II is reported in Figure 9. Type II has higherdensity because of the presence of the finely grained rocks.
To conclude the experimental results, the compressivestrength of phosphogypsum of hemihydrate origin is reportedin Figure 10. The highest compressive strength isapproximately 4,200 psi (28.9 MPa) and, as in the case ofthe dihydrate, there is not a substantial difference betweenspecimens subjected or not to one wet and dry cycle.
390
SURVEY OF JOINING MATERIALS
A survey of potential joining materials which couldsubstitute the conventional hydraulic cement mortar,currently used for masonry construction, includes fiveproducts. Only the types and not the brand names arereported in this paper; however, the products tested arereadily available in hardware stores nationwide. A list isprovided as follows according to the alphabetical order:
1. Acrylic Adhesive (AA)2. Asphalt Cement (AC)3. Concrete Repair Adhesive (CRA)4. Ceramic Tile Cement (CTC)5. Vinyl Adhesive (VA)
A11 products are canned, available in bulk quantities,ready for use, and can be applied with a brush.
Two types of tests were considered:
1. Direct tension test. Two 2 in. (51 mm) cubes made ofportland cement-phosphogypsum-limestone screening werejoined together after spreading a thin film of bondingagent (approximately 0.02 in.= 0.54 mm thick) onto twoadjacent surfaces. The two blocks were then pulledapart after seven days curing under open shelflaboratory conditions (75% RH, 24 C). The tensile loadwas applied through two eye-bolts embedded in the external surfaces of the prism and attached to thegrips of a universal testing machine.
2. Shear test. Three cubes, constructed, joined, and curedas described, in item (1), were subjected to acompressive force applied to the central element whilesupported at the two external ones. Since the loadingcontact zones were continuous over the 2x2 in. (51x51mm) element surfaces, the only relevant state of stresspresent at the joined interfaces was shear.
The results of the two tests are summarized in Table 1.It can be seen that the Concrete Repair Adhesive (CRA) andthe Ceramic Tile Cement (CTC) had the best performance inboth direct tension and shear. These two classes ofmaterials should be more closely investigated; particularly,the former that seems to be attractive from an economicalstandpoint.
CONCLUSIONS
The most relevant conclusion of this paper is the proofthat phosphogypsum-based bricks of good appearance andstrength characteristics can be factory produced. The benchmodel semiautomatic press designed and constructed for this
394
purpose is still a prototype, but it withdifferent degrees of success,
addresses,all the key features of a full
scale operational production line.
ACKNOWLEDGEMENTS
The authors are indebted to the Florida Institute ofPhosphate Research, Bartow, Florida for providing thefinancial support to the project.
REFERENCES
Lin,
Lin,
K.T., A. Nanni, and W.F. Chang. 1986. CompressiveStrength of Compacted Portland Cement Mixtures UsingPhosphogypsum. Proceedings of the Symposium on theConsolidation of Concrete, ACI Annual Convention, SanFrancisco, CA.
K.T., and W.F. Chang. 1987. Strength Properties ofCompacted Phosphogypsum-Based Mixtures. (in theseProceedings).
395
COLUMBIA COUNTY EXPERIMENTAL ROAD
Robert K.H. Hol& R.W. Williams2
L.L. Cogdill3 & W.F. Chang4
Second International Symposium on PhosphogypsumJ.L. Knight International Center, Miami, Florida
December 10-12, 1986
lSoils and Research' Engineer, Florida Department Transportation, Gainesville, Florida
2County Commissioner, Columbia County, Florida3Field Soils Engineer, Florida Department of TransportationGainesville, Florida
4Professor, Department of Civil & Architectural EngineeringUniversity of Miami, Coral Gables, Florida
397
INTRODUCTION
The continued accumulation of phosphogypsum waste stockpiles in
Florida has created an urgent need to find useful application of these
waste by-products and reduce the sizes of these stockpiles. In recent
years, the University of Miami had performed extensive laboratory
research into the engineering properties of phosphogypsum as well as
phosphogypsum mixtures with portland cement, flyash, lime and sand.
Laboratory test results indicate that phosphogypsum has potential
as a road building material. Whether phosphogypsum will be widely
used in highway construction will depend on the favorable results of
field test roads as well as evaluation of its effect on the
environment (air and groundwater).
This paper presents the engineering properties of the
phosphogypsum and sand mixtures as well as the design and construction
of a 2-mile rural road in Columbia County using these phosphogypsum
mixtures. Environmental monitoring of air, soil and groundwater
before and during construction has been conducted by the University of
Miami and the results will be reported in a separate paper.
398
DESCRIPTION OF TEST ROAD
The experimental road section known as White Springs Road
presently unpaved, is located south of SR 136 between I-75 and US41
southwest of White Springs. It is just under two miles in length and
the roadway consists of two 9-foot lanes sloping from the center to
either side of the roadway It is a lightly traveled rural road
consisting of local cars and light trucks.
The phosphogypsum for this project was obtained from Occidental
. Chemical Company in White Springs. Originally it was planned to
construct the test road entirely with hemihydrate and sand mixtures.
However before construction could begin, Occidental changed their
manufacturing process. It was then decided to build the test road
with dihydrate and sand mix as well as hemihydrate
sand mix.
and dihydrate and
The pavement profile will consist of one to 2-inches of type III
asphaltic concrete placed over about a lo-inch compacted base of
phosphogypsum and sand mixtures overlying the compacted subgrade of
existing soil. The base materials will consist of three different
blends of phosphogypsum and sand of approximately 1:2, 1:1, and 2:1
respectively as well as a short section with 100% dihydrate
phosphogypsum. The lengths of these test sections are about 5000,
2100, 1400 and 900 feet respectively.
399
CONSTRUCTION PROCESS
The construction operation was performed by Columbia County Road
Department personnel. Truck loads of dihydrate phosphogypsum were
hauled to the site in November 1986 and spread to an average depth of
5 inches along the first 5000-foot section. It was mixed into the
existing soil (A-3 fine sand according to AASHTO classification) with
a rotomixer to a depth of about 14 inches giving approximately a 1:2
blend of phosphogypsum and sand. A total of three passes of the
rotomixer was made to achieve uniform blending of the mixture.
Samples of native sand and the mixture were brought back to laboratory
for determination of maximum modified Proctor density. Figure 1 shows
the moisture-density relationships of the 1:2 soil mixture and the
native soil taken from the test site.
Before compaction of the mixture could be achieved, continuous
rainfall and wet weather persisted for about two months. It was late
January 1987 when the compaction operation resumed. During the wet
period, drivers encountered slippery surface conditions in the
compacted areas and soft yielding conditions in the uncompacted areas.
When compaction resumed in late January, a sheepsfoot roller was
used for the initial compaction of the lo-inch base. This was
followed by a steel wheeled roller and rubber-tired roller. A finish
grader was used to shape and complete the base.
'400
To seal and protect the finished base surface, a prime coat of
liquid hot asphalt and sand/asphalt screenings were placed and rolled.
The completed wearing surface consists of a 1 to 2 inches of type III
asphalt concrete sloping from the center of the pavement to both
e d g e s .
402
FIELD TESTING AND INSPECTION
Field testing and inspection were performed during the base
construction stage, It consisted of nuclear density and moisture
determinations, Clegg impact tests and visual observations of mixing
and compaction operations. Following the application of the Type III
wearing surface, dynaflect testing was conducted.
Nuclear Density Tests
Moisture-density relationship (AASHTO T-180 test method) for the
phosphogypsum-sand blend resulted in a project design density of 120
pcf at 7.5% moisture (See Fig 1). Initially, compaction requirements
were established at 98% of the laboratory proctor but, this
requirement was later reduced to 95% due to the types of available
compaction equipment, depth of base layer and unanticipated
construction delays.
Nuclear density and speedy moisture tests were conducted at
various locations along the completed roadway base section to verify
compliance with the established density requirements. Table 1 is a
summary of the density and moisture results obtained at the completion
of the compaction operation. Statistical analysis of the eleven tests
showed a mean density and moisture of 115.8 pcf and 8.1% respectively
with corresponding standard deviations of 2.3 pcf and 1.2%.
40.3
404
Clegg Impact Tests
The Clegg Impact Tester was designed and developed by Dr. 8. Clegg
of the University of Western Australia (1). It consists primarily of
a modified AASHTO compaction hammer weighing 1O lbs (4.5kg) with a
piezoelectric accelerometer attached to the hammer. The hammer is
dropped manually from a height of approximately 18 inches (450mm)
through a guide tube. Upon impact with the soil, output from the
accelerometer is fed into a hand held peak level meter. Clegg impact
values (CIV) of the fourth (4th) impact are recorded for analysis.
Testing with the Clegg Impact device was conducted at each of the
nuclear density test locations. CIV were obtained at centerline and
at 6 feet and 9 feet offsets from centerline, both left and right.
Table 2 is a numerical summary of the CIV data.
Figure 2 shows the transverse variation of CIV. Longitudinal
variations of CI V for the center line and 6 feet and 9 feet offsets
line are shown in figures 3 and 4. The data indicatefrom the center
that CIV further
to the center of
CIV had been
away from the center line are lower than those close
pavement.
correlated with laboratory LBR (limerock bearing
ratio) for some Florida soils such as limerock, sand-clay and clayey
and silty sand but not for phosphogypsum and sand mixtures. Generally
the higher the CIV, the higher the LBR values.
405
406
408
Dynaflect Testing
Following the paving of the 5000-foot test section, the dynaflect
was used for the non-destructive measurement of dynamic deflections.
It consists of a dynamic force generator, sensor assembly and digital
control device mounted on a relatively lightweight (2,000 pound)
two-wheel trailer. Two counter-rotating steel weights provide a 1,000
pound dynamic force to the pavement surface through a pair of rigid
wheels. Deflections along the pavement surface away from the rigid
wheels are measured by five geophones spaced at one foot intervals
(figure 5). Electrical signals from each geophone is amplified and
recorded as deflection in milli-inches. The deflections at sensor G2
are then used to estimate soil modulus E of the base using equations
developed from field research studies conducted by the Florida
Department of Transportation (2).
Dynaflect testing performed on the Columbia County project
consisted of 29 test sites in each traffic lane along the outside
wheel path (OWP) plus 29 test sites in the northbound traffic lane
along the inside wheelpath (IWP). Figures 6, 7, and 8 show the
measured deflections at the five geophone locations for northbound
409
inner wheelpath (NBTL-IWP), northbound outer wheelpath (NBTL-OWP) and
southbound outer wheel path (SBTL-OWP) with the nearest geophone (G1)
measuring the maximum deflection. The IWP deflection data are lower
than the OWP measurements indicating better compaction and confirmed
by the CIV data in the previous section. Table 3 gives a comparison
of the compacted
the Clegg Impact
CONCLUSION
E values for the phosphogypsum and sand mixture and
values (CIV) at approximate corresponding locations.
As of this writing (April, 1987), the 5000-foot section appears to
be performing well. Work had started on the remaining three sections
with higher proportions of phosphogypsum to sand mix. Compaction is
in progress. It is anticipated that the three remaining sections
could be paved by early May.
410
REFERENCES
1.
2.
Clegg, B., "An Impact Testing Device For In-Situ Base Course
Evaluation". ARR B Proceedings, Volume 8, 1976.
Miley, W.G., "Structural Layer Coefficients As Determined By The
Dynaflect", Bureau of Materials and 'Research,
FDOT, January 1984.
416