Journal of Hazardous MateriaIs 37 (1994) 83-89 83
CLEAN PROCESS FOR THE PRODUCTION OF DEFLUORINATED
DICALCIUhI PHOSPHATE USING PHOSPHATE ROCK
M. GIULIETTI ‘,2
1 IPT- Institute for Technological Research, Chemistry Division, Cidade Universitaria,
CEP 05508, SZo Paulo, Brazil. Fax 55-I l-869-3 13 I.
* UFSCar- Federal University of SZio Carlos, Sao Carlos, Brazil
ABSTRACT
A new process for the production of defluorinated dicalcium phosphate using phosphate
rock was developed. This new process has the main characteristic of being clean, without the
production of any liquid waste. The process has essentially three main steps: phosphate rock
digestion, defluorination and neutralization. In the first step a phosphate rock slurry is batch
digested with concentrated sulfuric acid in the presence of some additives that retain the fluorine
in the solution, as well as enable the adequate growth of the calcium sulfate crystals and maximize
the phosphate dissolution. At the end of the digestion the slurry is flltrated and the mother licquor
is sent to the next step. The second step of the process is a fluorine salt precipitate sedimentation
carried out by the addition of a potassium or sodium salt. The clear solution containing mainly a
weak phosphoric acid is then neutralized, in the third step, in order to produce the final dicalcium
phosphate product. After filtration the dicalcium phosphate is dried and the mother licquor is
recycled to the first step of the process, the phosphate rock digestion. All the liquids of the
process are recycled and no wastes are produced. This new process is economically interesting
because it uses a very cheap raw material, the phosphate rock, instead of the conventional
defluorinated phosphoric acid.
INTRODUCTION
Phosphorus is an essential nutrient for living systems. Its main source in the nature is the
so called apatite, a phosphate mineral, found in most cases in igneous or sedimentary deposits of
fluorapatite, also called natural phosphate. In order to be used as fertilizer or mineral supplement
for animals the mineral fluorapatite must be in a soluble form. There are basically two
soiubilization processes: thermal, that destroy the crystalline structure using heat, and the wet
process, that uses the chemical energy to make this destruction. ABer the solubiliition, other
soluble phosphorus compounds are produced, mainly calcium and ammonium phosphates. For
animal feed use, the most popular product is dicalcium phosphate (DCP) in dihydrate
(CaHP0~.2H20) or anhydrous (CaHPOa) form.
DCP is generally obtained by the neutralization reaction between dethrorinated phosphoric
acid and a calcium source, normally calcium carbonate or calcium hydroxide. The defluorinated
phosphoric acid is more expensive than the normal wet phosphoric acid; it is used at 48-54%
P OS, less than 0.2% F and is produced in large units of more than 100 thousands tpy of P205
[ ?I. Small units are not economically feasible due to the high costs of evaporation to concentrate
the acid. So, the defluorinated phosphoric acid is an expensive raw material and it is available only
near large phosphoric acid producers. On the other hand, phosphate rock is a relatively cheap and
available raw material. In the international markets, the price relation between the P20, from the
phosphoric acid @ 48-54% PzO5 and phosphate rock @, 32-36% P205, varies from 2.5 to 5.0.
0304-3894/94/$07.99 0 1994 Elsevier Science B.V. All rights reserved
SSDZ 0304-3894(93)E0117-K
84
For defluorinated phosphoric acid, obtained by the thermal process or by solvent extraction, this
relation can reach the value of 10.0. On the other hand. for the DCP. the relation between P,O<
L _I
content and detluorinated phosphoric acid P,Oj content’is only 1.3.
The economic interest to produce DCP directly from phosphate rock lies in this price
relation. It is possible to obtain relations between the market price of P,Oj in DCP to the cost of
P,lOj of phosphate rock with values near 3.0. This is very attractive from the economic point of
view
PROCESS DESCRIPTION
The proposed process uses phosphate rock and produces phosphoric acid with the
objective of producing DCP, without the need for concentrating phosphoric acid, which is
normally a complex and energetically expensive step. This makes it possible to produce
phosphoric acid in small scale, with nominal capacities of less than 20 thousand tpy of PzOj and
even in batch processing.
Figure I shows the proposed process diagram. Phosphate rock is fed in the digestion
reactor with the recycle water that comes from the gas washing column and with some additives
that promote good crystal growth of gypsum, CaS0+2H,O, and aids the fluorine removal.
Concentrated sult%ric acid is slowly added to the digestion reactor. The heat released by the
reaction elevates the reaction temperature, thus favoring the phosphate rock attack. The gases
liberated by the reaction are absorbed in a washing column that utilizes the water that comes from
the DCP filter. The washed gases are discharged to atmosphere. The main reaction occurring in
the digestion reactor is:
Calo(POJ),$2 + IOH2SO4 + lOH20 + 6H3PO~flOCaSOJ.2H20 + 2I-E (1)
After completion of the reaction the slurry is discharged to a filter where the gypsum and
insoluble impurities are seoarated from the ohosohoric acid solution. After the filtration. the cake
is washed with fresh water to remove the detained phosphoric acid. The amount of water used
must be nearly the same as the water retained by the cake. In this way the total water balance is
conserved. The gypsum produced has about 30% humidity: a small amount of unreacted
phosphate rock and can be used in this form in agriculture, as sotI conditioner[2].
The next step of the process, is the defluorination of the diluted phosphoric acid produced.
This is made by the addition of an Inorganic salt, such as sodium chloride in a settling chamber.
After decantation the clear and defluorinated phosphoric acid solution is removed by slowly
siphoning, in order to prevent the removal of tluossilicate salt. The main reaction involved in the
defluorination step is, for the NaCl case:
H2SiF6 + 2NaCI -_) Na2SiF6 + 2HCI (2)
If the amount of silica in the phosphate rock is not enough to complete the fluorine
removal, a supplement of active silica must be added in the reaction step. The precipitated
fluossilicate salt can be removed from the decanter, washed and dried to be used as a fluorine
supplement in water treatment systems, or as a fluorine source in the pharmaceutical industry.
The detluorinated phosphoric acid solution is then sent to the neutralization step, where
the DCP is produced. The clear acid solution must be free of sulfate ion. This can be done bv
correcting the pH in the digestion reaction step. The sulfate free acid solution is then neutralized
with a calcium source, such as calcium hydroxide or calcium carbonate. This step is controlled by
the pH value of the solution. Temperature defines the form of the precipitated DCP crystals. For
85
temperatures above 9Occ, anhydrous DCP will preferentially be formed. During the
neutralization, the main reaction occurring in the system is, for the Ca(OH)z case:
Ca(OH)2 + HIPOJ + CaHPO.+ nH?O + (2-n)HzO (3)
with n=2 for the dihydrate form or n=O for the anhydrous form.
To obtain a DCP product with a good crystallinity and with well-formed crystals it is
necessary to control the supersaturation during the neutralization. This can be done by slow
addition of the calcium source. In this way the formation of insoluble phosphates can also be
avoided. AtIer reaction completion, about 2 hours, the slurry is sent to a filter for separation of
the DCP and the mother licquor. The DCP must be washed with fresh water in the same amount
as the contained humiditv (about 40%). in order to keep control of the total water balance. The
mother licquor receives ihe DCP war&g water and is’sent to the washing gas column, closing
the cycle. In this way, no effluents are produced.
The DCP produced is then dried in a convective way and sent to the storage system. The
DCP can be produced at different concentrations, which gives good flexibility to the process.
EXPERIMENTAL
Several experiments were conduced for all process steps. The experiments were done in a
1 liter closed flask with control of the temperature by means of a thermostated bath. The flask has
also a gas washing system that eliminares the fluorine in the effluent gas and prevents the water
from evaporating, in order to keep the global water balance. This system is also used for the other
process steps, i. e., fluorine removal and diluted phosphoric acid neutralization. For all the solid-
liquid separations vacuum Biichner filter is used.
Three kinds of phosphate rock raw materials were used: Morocco (sedimentary), Catallo
(igneous) and Patos de Minas (sedimentary). The chemical analyses of these phosphate rocks are
presented in Table 1. In all experiments the phosphate rocks were ground to -200 mesh. The
sulfuric acid used was of commercial grade. Calcium hydroxide,. as well calcium carbonate, was of
technical grade. Normally these present good quality and a low Impurity content.
In all cases only the batch way of processing was used, taking in account however the
mother licquor recycle aspect. In this way all experiments were conduced recycling five times the
mother liquor to the digestion reactor.
The parameters studied for each process step is described below. For the phosphate rock
digestion reaction step, the parameters studied were as follows:
l digestion reaction time: 2.4 and 6 hours.
l digestion reaction temperature: 50 and 75oC.
l speed rotation of the reactor impeller: 90 and 300 rpm.
l free sulfate level in the reactor: 1. 2 and 5%.
l surface active agent in the reactor: 0 and 0.5%.
l addition of active silica: 0 and 50% of the stoichiometric needs
l addition of gypsum crystal growth conditioner: 0 and 1%.
l final pH: 0.2 to 1S.
l time of sulfuric acid addition: 20 to 70% of the total reaction time.
For the defluorination step, the parameters studied were as follows:
l sedimentation time: 6, 12 and 24 hours.
l addition of fluossilicate salt seeds: 0,O. 1 and 2%
LDOITIUES
’ TO
nln.
1
SULFURIC
-_)
MID
DIGESTION GAS
+
PlKlSPHPlE REACTION WASHING
RGCX
,
ISALT
MITER NEUTRA-
+FILTER - P SETTLING ---_)
LIZATION
1
I
GYPSM 10 FLlKkXIlICIIE ClCIla
ffiRICULlURE SlLT HIDROYlDE
Fig& Schematic diagram for DCP proposed process
Table 1. Phosphate rocks chemical analysis
PHOSPHATE ROCK
PATOS DE MINAS
l sedimentation pH: 0.2 to 1.5.
l defluorination salts: NaCI, Na(S0,) , KCI, K,SO+ NaNO,, KNO,, Car&,
NaHCO;, hlgC T?, NaCl KCI.
l salt:F molar ratio: 0.8 to 2 ofthe stolchiometric.
l addition of extra active silica: 0 and 5%.
Finally, for the neutralization step, the parameters studied were the follws
l neutralization time: 0.5, I, 2 and 3 hours.
l neutralization temperature: 40, 60 and 95oC.
l calcium excess: 0, 5 and 20% of the final stoi chiometric value.
l initial pH: 0.2 to I.5
l final pH: 2.8 to 6.0
l time of calcium source addition: 10 and 70% of the total neutralization time
In order to adequately evaluate the effects of changing the above described parameters,
the following variables, for each step, were measured.
For the phosphate rock digestion reaction step:
l filterability of the gypsum cake
0 P105 conversion.
l fluorine removal.
l gypsum crystal size distribution,
l sulfate and calcium in the dried cake
l X ray diffraction of the dried cake
For the defluorination step:
l fluorine removal.
l X ray diffraction of the precipitate
For the neutralization step:
l filterability of the cake (DCP product).
l P205 conversion.
l calcium, total and soluble (in neutral ammonium citrate solution) phosphorus , sulfate
and fluorine analysis.
l X ray diffraction of the final dried product.
In this way, almost all process variables were evaluated and the results are discussed
below.
RESULTS AND DISCUSSION
In the phosphate rock digestion step it is possible to state that the conversion,
expressed as the ratio of PzOs in the leaving gypsum to the PzOs fed, is not affected by the
rotation speed, in the range studied. For the three phosphates the best results for the P,Os
conversion are:
l digestion reaction time: 4 hours.
l digestion reaction temperature: SOoC.
l free sulfate level in the reactor: 2%.
l surface active agent in the reactor: 0.5%.
l addition of active silica: 50% of the stoichiometric.
l addition of gypsum crystals growth conditioner: 1%.
88
l final pH. I. I.
l time of sulfuric acid addition: 60% of the total reaction time.
In order to obtain good gypsum crystals and consequently good cake filterability it is
necessary to create adequate crystal growth conditions, High reaction temperatures cause a high
number of nuclei and tend to give bad cake filterability in spite of increase of the PzOj
conversion. Long reaction times tend to give big crystals, but reduce the production rate. The
range of gypsum crystal size for almost all experiments was from 30 to 80 microns as the mean
diameter.
In the phosphate rock reaction step, the fluorine removal was from 40 to 60% for all
phosphate rocks studied. in the best operational conditions described above.
For the conditions studied, the cake was formed essentially of calcium sulfate dihydrate
with small amounts of calcium sulfate hemihydrate, besides unreacted phosphate rock, silica and
fluossilicate salts. The P,Oj conversion in the phosphate rock digestion step was always between
85 to 96%.
In the fluorine removal step, it is possible to state that the higher the sedimentation time
the more efficient the fluorine removal. Best results were obtained when the sedimentation was
seeded with tuossilicate salt crystals.
The fluorine removal tends to be higher at elevated pH. Increasing the salt : F molar ratio
the fluorine removal also increases but not significantly. Addition of extra active silica does not
affect the veld of fluorine removal. Best results were observed when ootassium salts were used.
The analysis of the precipitate by X ray diffraction always shows the presence of the
corresponding fluossilicate salt and a small amount of gypsum. The fluorine removal in this step
varied from 60 to 85%. which represents globally, 70 to 95% removal.
In the neutralization step good results were observed for all phosphate rocks studied. The
best operational conditions observed for the neutralization were as follows:
l neutralization time: 1 hour.
l neutralization temperature: 95OC.
l calcium excess: 5% of the final stoichiometric value.
l initial pH: 1.2
l final pH: 3.5
l time of calcium source addition: 50% of the total neutralization time.
In terms of P205 conversion, it is possible to say that best results were observed at high
temperatures and long neutralization times. For higher temperatures the P 05 content increases
due to the formation of anhvdrous DCP. instead the dihvdrate DCP. The t2tlterabilitv of the final
DCP cake product is bette; when the neutralization is conducted at low tempera&es because
then the number of nuclei formed is lower and the crystal size higher. High final pH values result
in higher amounts of “insoluble” phosphates, like hydroxiapatite or fluorapatite, which results in
low quality of the final DCP product. When a sulfate salt is used in the defluorination step, a small
amount of calcium sulfate is present at the final product.
The X ray diffraction analysis always shows the presence of anhydrous DCP for
temperatures below 7OoC. The amount of P,Os precipitated at the neutralization step was always
above 90%. In this way, the global P,Os conversion is in the order of 90%.
The final F:P mass ratio was above 100 for the better operational conditions, which agree
with most of standards for feed grade phosphates.
89
A new, clean process for the production of DCP is presented, which uses phosphate rock
as raw material Three ditferent rocks were studied and the process shows good results with all of
them
Figure 2 shows the schematic tlow diagram of a DCP production batch unit. For a nominal
capacity of IO thousand tpy of DCP product an investment of about SO0 thousand US$ dollars
can be expected For a typical unit like that, a pay-out time of 8 months can be attained for a
stable production market
Fig2: Flow diagram for a DCP production batch unit
REFERENCES
I) UNIDO - United Nations Industrial Development Organization. Fertilizer Manual. New York,
UNIDO, Development and Transfer of Technology Series, 13, 1980.
2) RAIJ, B. V : “Gesso Agricola na Melhoria do Ambiente Radicular no Subsolo”. Associaqlo
National para DilLGo de Adubos e Corretivos Agricolas, S%o Paula, 1988.
