Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~ ~ 335752
--1--
AGGLOMERATED PHOSPHATE FURNACE CHARGE
This invention relates to an agglomerated phosphate
material suitable for charging in an electric-type phos-
phorus furnace. More particularly, the invention is
concerned with a briquetted phosphatic material as a
supplemental furnace feed for increased phosphorus
production.
In the electrothermal manufacture of elemental phos-
phorus, feed streams of phosphatic material such as
calcined phosphate ore, a carbonaceous reductant such as
coke and optionally a fluxing agent such as silica are
charged into an electric furnace. The charge materials
undergo resistive heating which results in the formation
of a molten reaction mass. Reduction of the phosphate
ore to phosphorus produces a gaseous mixture of phos-
phorus vapor, carbon monoxide and particulates. After
freeing of particulates by electrostatic precipitation,
the gaseous stream is water quenched and the condensed
phosphorus recovered and stored under water. The fur-
nace is tapped periodically to remove molten slag and
liquid ferrophosphorus.
In a typical method of preparing the phosphatic feed
material, raw phosphate ore is first formed into aggre-
gates or agglomerates of the requisite size by compact-
ing comminuted phosphate ore with a binder to form
shaped articles such as pellets or briquettes, usually
the latter. These are then calcined to increase their
crush strength and thereby minimize breakage. The
procedure is much used in the processing of phosphate
shales such as are found in the Western areas of the
United States. These shales usually contain clay which
undergoes sintering during calcination thereby acting as
a binder for the phosphate particles to give a high
strength agglomerate. The technique is not, however,
.. , . . - ~ . . . . -. . . . . . .
` -
-2- ~ 335752
applicable to clay deficient shales or to other phos-
phates lacking binding properties. Agglomeration of
these materials requires additional clay or other bind-
er.
Although the manufacture of phosphorus by reduction
of phosphate ore in an electric furnace is an establish-
ed industry, it is not entirely free of operational
problems. For example, the briquettes of calcined phos-
phate shale are subject to a certain amount of abrasion
during handling and while in transport to the furnace.
As a consequence, a certain amount of calcined ore
particulates are generated. Over extended periods of
operation, these fines, commonly referred to as nodule
fines, accumulate in considerable amounts.
Fines build up can be ameliorated to some extent by
blending a stream of recycle nodule fines with fresh
shale ore. However, this approach tends to be self-
defeating owing to the increased susceptibility to
abrasion of shale agglomerates containing recovered
nodule fines. Nodule fines, unlike raw phosphate shale,
cannot be compacted into strong shapes. Calcining
destroys the binding properties of the shale. Hence the
decrease in strength of phosphate aggregates or bri-
quettes cont~;n;ng nodule fines. Clearly, recycling is
not the answer to nodule fines build up.
Another problem that frequently confronts phosphorus
producers involves matter of supply and demand and comes
about in the following manner. From time to time there
is a marked increase in the consumption of phosphorus.
The timing and extent of these increases are not always
predictable although they tend to parallel periods of
heightened business and economic activity. In some
instances, the demand for phosphorus may exceed the
output of a given plant resulting in loss of potential
sales. Of course, additional furnace capacity could be
; . ... ... . . . . . . . . . . . . . ..
_3_ 1 3 3 5 7 ~ 2
installed but then the plant would be underutilized
during intervals of normal or baseline operation. What
is needed is a means of dealing with the wide swings in
phosphorus markets while keeping within the general
bounds of existing plant design and capacity.
Now, therefore, it is a principal object of the
present invention to provide a process of reducing the
buildup of calcined phosphate fines while simultaneously
increasing the yield of phosphorus from an electric-type
phosphorus furnace in which the charge material is cal-
cined phosphate agglomerates, carbon and the requisite
silica flux. Other general objects and advantages of
the invention will become apparent in the more detailed
description as subsequently set forth herein.
Broadly, the objects and advantages of the invention
are realized by the steps of:
(1) mixing calcined phosphate fines, phosphoric
acid, carbon and silica to form a phosphatic feed
material suitable for winning phosphorus in an electric-
type phosphorus furnace, said charge material containing
proportions of phosphoric acid and phosphate fines such
that on reaction with the phosphate fines the phosphoric
acid is converted to phosphate salts, the carbon being
in at least a stoichiometric amount required to reduce
the phosphorus values in the charge material to
elemental values thereof and the quantity of silica
being sufficient to provide the required amount of flux;
(2) forming agglomerates of said phosphatic feed
material;
(3) forming cured agglomerates by heating the green
agglomerates to effect reaction between the phosphate
fines and phosphoric acid whereby the phosphoric acid is
converted into phosphate salts resulting in increased
hardness and strength of the agglomerates, and
(4) introducing the cured agglomerates into the
_4_ t 3 3 5 7 5 2
phosphorus furnace as a supplemental charge stream.
The homogeneous phosphatic feed material of step (1)
is produced by bringing the components together in a
known blending device such as a pug mill or the like
until a homogeneous mixture is obtained. Desirably, the
mixing device is made of acid-resistant material. The
amounts of components in the phosphatic feed material
should satisfy the requisite stoichiometry as explained
further herein while at the same time providing a con-
sistency that is suitable for converting into agglom-
erates.
The agglomerates in step (2) are produced by forming
shaped structures or configurations of the phosphatic
feed material of step (1). Exemplary shaping techniques
include nodulizing, pelletizing or briquetting. Bri-
quetting is most commonly employed and is carried out in
commercially available equipment such as a roll briquet-
ting press.
The green (uncured) agglomerates are much weaker
than their uncured counterparts and consequently require
more careful handling to prevent undue crumbling. Green
strength can, however, be improved by employing high
phosphoric acid binder levels and/or by precompacting
the blended components prior to agglomeration. As
carried out herein, precompaction comprises the follow-
ing sequence of steps:
(1) compacting a mixture of the calcined phosphate
fines, carbon, silica and acid binder;
(2) granulating the compacted mixture;
(3) sieving the granulated material through a half
inch screen, and
(4) recompacting the screen material.
The precompacting can be performed using a roll
briquetting press of the standard type. Improved
strength derives from the fact that the effective
.
.
_5_ 1 3 3 5 7 S 2
pressure exerted on the material in the second compact-
ing step is significantly greater than in the first
step.
The freshly formed or green agglomerates are cured
in order to increase their crush strength and render
them more resistant to abrasion. In general, curing
consists in heating the green agglomerates at tempera-
tures ranging from about 100C to 500C. Heating
periods can range from about 0.3 hours at the higher
temperature to about 2.0 hours at the lower temperature.
Preferred curing conditions are from about 150C to
200C for one to two hours.
During the curing step, the calcined phosphate fines
and phosphoric acid react to produce complex acid phos-
phate salts, examples of which are thought to include
CaHP04 and Ca(H2P04)2. It is believed that such salt
formation is responsible for the increase in strength
and abrasion resistance of the cured agglomerates.
Curing is also advantageous in that it promotes
expulsion of water, both free and hydrated, as well as
other volatile materials which would be detrimental if
released in the furnace.
As previously pointed out, during curing the phos-
phate fines and phosphoric acid react to give complex
phosphate salts. Such reaction is necessary since any
free phosphoric acid in the agglomerates would be
expelled as P2O5 in the furnace and thus unavailable for
conversion into elemental phosphorus.
The phosphoric acid binder used in practicing the
invention need not be pure or highly refined. In fact,
a low cost technical grade of phosphoric acid, such as
commercially available green wet process acid (WPA) is
entirely satisfactory and even preferred. Acids con-
t~;n;ng by weight from about 30% up to about 100% H3P04
constitute suitable binder materials. The amount of
~ 1 335752
phosphoric acid binder in the phosphatic feed material
can be expressed as follows:
% (by weight) Binder = (gms as X% H3PO~) flOO)
gms total solids + gms as X% H3P04
Green acid assay at about 70% (X=70% in the formula) is
preferred. Binder levels ranging from about 7~ to about
15% are satisfactory.
The percentage of carbon in the phosphatic feed
material should be sufficient to reduce the phosphate
values to phosphorus as well as other reducible sub-
stances present in the calcined phosphate fines to their
elemental forms. The stoichiometry for reducing phos-
phate values, expressed as P205, can be depicted as
follows:
(1) 2P205 + lOC----> P4 + lOCO
Suitable sources of carbon reductant are various cokes
derived from the pyrolysis of petroleum and coal.
About 9% to 10% phosphoric acid binder is usually
req~ired to give cured agglomerates of sufficient
strength and hardness. If it is assumed, for instance,
that a 9.6% level of 70% H3P04 is used and that the car-
bon, as coal derived coke such as coke breeze, and the
calcined phosphate fines contain 77% fixed carbon and
25% P20s, respectively, the charge material required to
satisfy equation (1) supra would consist of 14.7% of
coke and 85.3% of nodule fines. From an economic stand-
point, however, it is generally desirable to include an
excess of one or the other solid components. For
example, use of excess calcined phosphate fines would
ensure that the more expensive coke component undergoes
essentially complete reaction.
Silica is commonly added to the charge materials as
a fluxing agent which facilitates smelting of the phos-
phatic agglomerates. The flux forming reaction is
.. . . . .
~7~ 1 3 3 5 7 5 2
generally expressed in simplified form, as follows:
(2) 2Ca3(PO4)2 + lOC + 6sio2 ---> P4 + lOCO + 6CaSiO3
Representative of the phosphate fines used in
producing the furnace charge material of the invention
are the nodule fines recovered from a commercial phos-
phorus plant in which a charge of calcined phosphate ore
agglomerates, coke and silica are smelted in an electric
phosphorus furnace. Typically, such nodule fines assay
at about 32% CaO and about 24% sio2, corresponding to a
CaO/SiO2 mole ratio of about 1.4. The charge material
of the invention has yielded strong composite cured
agglomerates containing about 6.5 parts added silica per
100 parts of nodule fines. This gives a CaO/SiO2 mole
ratio of about 1.1, close to the stoichiometry required
by equation (2).
The particle size of nodule fines recovered from the
calciner unit of a commercial electric phosphorus plant
consists mainly of ~4.75mm (U.S.A. No. 4 Standard Sieve)
material. Another source of calcined phosphate fines is
burden dust recovered from the plant electrostatic
precipitators. Burden dust consists of <150Jum material
and ordinarily contains a few percent of coke dust.
Strong composite agglomerates, highly suitable as a
charge stream, containing calcined phosphate fines mixed
with up to 55% burden dust, have been produced by the
process of the invention.
Size distribution of representative nodule fines and
commercially available particulate carbons suitable as
carbon reductants are listed in Table I.
The cured agglomerates of the invention prepared
with nodule fines are introduced into a phosphorus fur-
nace as a supplemental furnace charge. The combination
- of supplemental and main furnace charges gives a yield
of phosphorus that is greater than that of the main
charge alone. Whereas the P205 assay of a conventional
, j, ", . : .. . . .
-8- 1 3 3 5 7 5 2
furnace charge is about 25% that of the invention is
about 27~. The use of such an enriched phosphatic
charge material translates into increased phosphorus
production on the order of several million pounds per
year for a typical commercial phosphorus plant.
Thus, the invention solves the problem of nodule
fines buildup while providing means of increasing phos-
phorus output without having to install added plant
capacity.
The invention is illustrated in further detail by
the following test procedures and examples in which
compositions are on a weight basis unless stated other-
wise.
General Preparations
Agqlomerate PreParation - Large cylindrical pellets
(ca 2.8 cm. diameter x 2.8 cm. height) were prepared for
the mech~nical strength evaluations which consisted of
Abradability and Tumble Tests (see below). Small (1.27
diameter x 1.1 cm. height) pellets were used in the
Furnace Reactivity Tests.
The large pellets were prepared in a Carver Press
from 35.0 gram portions of well mixed (Hobart Blender)
blends containing nodule fines, coke fines, 70% H3P04
and, optionally, burden dust and/or silica. In most
cases small quantities of water were also added to
facilitate the blending and pelletizing steps. The
small pellets were prepared from 2.0 gram portions of
the well mixed blends in a hand operated Pellet Press
(Parr Instrument Co.). The solids were used "as
received" for all of the m~r-h~n;cal strength tests and
for most of the reactivity evaluations. In some of the
latter, the solids were ground in a mortar and pestle to
obtain <25~um (U.S.A. No. 60 Sieve) material for small
pellet preparation. Typical size distributions for the
"as received" materials are given in Table I.
.. . . .
9 1 3 3 5 7 5 2
Curing Ske~ - All of the pellets in Table I were
routinely cured in a laboratory oven at 200C + 15C for
one hour. In one case, the cured pellets were
additionally "fired" by heating them at 1000C in a
laboratory tube furnace under nitrogen for 0.5 hour.
Test Procedures
Abradability Test - Four large pellets were weighed
and placed on a U.S.A Standard Series No. 6 Sieve
(3.35mm) equipped with a metal cover and receiving pan.
This assembly was shaken in a Portable Sieve Shaker
(Tyler Model RX24) for 20 minutes. The total quantity
(both >3.35mm and <3.35mm) of material abraded from the
pellets was determined by weighing, and calculated as
percent abraded based on the original sample weight.
Tumble Test - This test was performed on green
(uncured) and cured pellets. Four large green pellets
were weighed and rotated at 43 rpm for exactly one
minute in a 21.6 cm. diameter steel drum containing four
one inch high steel baffles. The 0.6 cm. material
re~in;ng after this treatment was obtained by sieving
the sample in a 0.6 cm. screen. This material was
weighed and the percent 0.6 cm. was calculated based on
the original sample weight. Four large cured pellets
were similarly treated for 10 minutes and the percent
plus 0.6 cm. material was calculated in the same manner.
Furnace ReactivitY Tests - Three or four small
pellets were weighed in a ceramic boat. The boat was
placed in a 5.1 cm. ID mullite tube equipped with end
plates containing 0.6 cm. stainless steel tubes for
entry and exiting of sweep gases. The tube was
positioned in a Lindberg 1500C horizontal tube furnace
in such a way that the ceramic boat and contents were in
the center (hot zone) of the furnace. The mullite tube
was blanketed with dry nitrogen. The temperature was
preset to the desired value and the power turned on.
~ -10- 1335752
The onset of the P4-forming reaction was noted by the appearance
of a small flame at the exit end of the apparatus. The reaction
was allowed to take place at the set temperature for three hours
(see Examples 5 and 6) after which the power was turned off. The
contents of the mullite tube were allowed to cool to room
temperature under nitrogen. The boat plus contents were then
reweighed and the percent weight loss accompanying the reaction
was calculated. The total percent P205 in the cured pellets and
that remaining ln the reacted agglomerates was determined by a
reliable titrimetric analytical procedure (see Anal. Chem., 20 p.
1052, 1948). The percent conversion in each case was calculated
from the loss in total P2O5 in the agglomerates by means of the
following formula:
(original spl,wt)(l- %wt.loss)(% P205 remaining)
% Conversion = 100 - 100
(original spl,wt)(original % P2O5)
100
The above formula represents:
% Conversion - g. P2O5 in original - g. P205 in final
g- P205 in original x 100
Example 1
Efect of Binder in Abradability Test
Large pellets were prepared at 27.58 KPa (4000 psi) from a
mixture containing 100.0 grams of nodule fines, 17.Z grams of
petroleum coke fines, 12.4 grams of 70% H3PO4 (9.6% binder level)
and 8.8 grams of water. The cured pellets abraded to the extent
of only 3.1% after 20 minutes. For comparison purposes,
similarly prepared and cured pellets containing no binder abraded
completely after only three minutes demonstrating that the 9.6%
binder level used in the example markedly decreased the tendency
of composite agglomerates to undergo abrasion under adverse
handling conditions.
~. ~; ~
~ ~ . ~ .
-11- 1 3 3 5 7 5 2
~xamPle 2
Effect of Binder with Silica Present
Pellets were prepared as per Example 1 except that
6.4 grams of silica per 100 grams of nodule fines was
incorporated into the blend. Cured pellets containing a
9.1% H3P04 (70%) binder level abraded to the extent of
only 3.5% after 20 minutes while those containing no
binder abraded essentially completely after three
minutes. These results demonstrate that phosphoric acid
is also an effective binder for agglomerates containing
approximately reactant proportions of silica and lime
(CaO/SiO2 mole ratio = 1.1) according to Equation 2.
Example 3
~ffect of "Firing" at 1000C
Large pellets were prepared at 41.37 kPa (6000 psi)
pressure in the Carver Press from a mixture containing
the following proportions of ingredients by weight:
100.0 parts of nodule fines, 14.6 parts of metallurgical
coke fines, 8.7 parts of green H3PO4 (70%) and 10.0
parts of water. The binder level, 70% H3P04, was 7.1%.
The pellets, after curing at 200C for one hour, abraded
to the extent of 2.5% in the Abradability Test. Several
cured pellets were additionally "fired" at 1000C for 30
minutes under a reducing atmosphere (ca 95% N2 + 5%
CH4). These pellets, after cooling, were evaluated in
the Abradability Test in which they underwent only 7.3%
abrasion after 20 minutes.
Prior Art Comparison
In another experiment, pellets were prepared using
molasses as a binder. The mixture contained 140 parts
of nodule fines, 60 parts of burden dust, 24.4 parts of
metallurgical coke breeze, 30.6 parts of molasses (12%
binder level) and 19.8 parts of water. After having
been cured and additionally fired at 1000C, these
pellets underwent 83.7~ abrasion in only ten minutes
.
-
1 335752
-12-
indicating that molasses, a typical carbohydrate, is an
ineffective binder for agglomerates consisting mostly of
calcined phosphate. This example demonstrates that
phosphoric acid is a uniquely effective binder for
composite agglomerates in that these agglomerates are
not significantly weakened at high temperatures unlike
those prepared with carbohydrate binders.
Exam~le 4
Tumble Test
The Tumble Test is somewhat more severe than the
Abradability Test described in the previous examples.
It was implemented to further determine whether
composite agglomerates prepared with H3PO4 binder are
likely to disintegrate in transit to the furnaces. A
mixture of nodule fines and burden dust containing 30%
by weight of the latter was combined with coke in pro-
portions such that the coke comprised 10.9% of the solid
blend. This mixture was combined with three levels of
phosphoric acid and additional water in separate experi-
ments. Large pellets were formed at 27.58 kPa (4000
psi) in the Carver Press. Tumble Test results on both
green and cured agglomerates are shown in Table II.
These results demonstrate that mechanical strength is
increased with increasing levels of the binder in both
cases.
ExamPle 5
Furnace ReactivitY of ComPosite Pellets
Small pellets containing approximately stoichio-
metric proportions (Equations 1 and 2) of total P2O5 and
coke carbon were prepared from the following blend:
Nodule Fines 72.2%
Coke Fines 12.4%
Green Acid (70%) 9.0%
Free Water 6.4%
-13- l 335752
The pellets were cured at 200C for one hour then
heated in a tube furnace for three hours at 1300C.
Coke fines were omitted from the pellets in one experi-
ment. Formation of elemental P4 was indicated by the
appearance o~ a flame at the exit end of the tube in the
experiments in which coke fines were present. When coke
was omitted the conversion value as defined previously
was negligible at 1.6% (see Table III) indicating that
little or no free P2O5 was volatilized at 1300C. When
"as received" coke and nodule fines were used, moderate
(31% to 58%) conversions were obtained. However, by
grinding these materials to <25~m (0.0098 inch) it was
possible to boost the conversions to 81% to 84% under
otherwise identical reaction conditions (Table III). It
was also observed that the pellets did not melt in any
of the 1300C experiments cited below.
ExamPle 6
Effect of Temperature
In this example, the phosphorus forming reaction was
2Q allowed to proceed at higher temperatures than Example
5. Large pellets were prepared in the Carver Press at
27.58 kPa (4000 pounds/in2) from a mixture containing
<18~um particles and green phosphric acid. The follow-
ing percentages of these ingredients were used to give
composite pellets containing a 7.9% molar excess of
carbon over that required to react with the nodule plus
binder P2Os content. The Sio2/CaO molar ratio was 0.97.
Nodule Fines 74.7%
Coke (98.7% fixed carbon) 10.2%
Silica (100%) 5.1%
Green Acid (66.7% H3PO4) 10.0%
The pellets were cured at 200C. One large (20.9
gm) pellet was heated in a molybdenum crucible, contain-
ed in a vertical high temperature furnace equipped with
a controller-programmer. The temperature-time relation-
1 335752
-14-
ships for the reactant pellet (obtained by optical pyro-
meter readings through a sight port in the reactor head)
were as follows:
Time Hours Reaction TemP. C
2.0 1070
3.0 1200
4.0 1340
5.0 1490
5.3 1530 (maximum)
6.0 1460
8.0 1190
10.0 850 (estimated)
The P4-forming reaction started at about 1080C and
a melt was observed at 1375C. The slag remaining after
the reaction contained only 0.09% P20s, corresponding to
99.7% conversion of the total P20s in the mixture to P4.
This example illustrates that very high conversions of
the P2Os in the composite feed are possible at maximum
temperatures, in the neighborhood of 1500C.
Ex~m~le 7
Effect of Particle Sizinq and Silica
This example demonstrates that neither fine (<18~m)
mesh particle sizing nor added silica is required to
achieve high conversion of P2O5 under high temperature
(1500-1550C) reaction conditions. In this case,
pellets were prepared as described in the previous
example, from <3.35mm solids and other ingredients, to
give an 8.0% molar excess of carbon and approximately a
0.70 SiO2/CaO molar ratio.
Nodule Fines 72.9%
Coke (83.3% fixed carbon)12.5%
Green Acid (66.8~ H3P04)10.7%
Water 3-9%
The pellets were cured at 200C and heated to a
maximum temperature of 1530C at a slightly faster rate
-15- 1 3 3 5 7 5 2
than in the previous example. The P4-forming reaction
started at about 1120C and a melt was observed at about
1250C. A weight loss of 37.3% occurred during the
reaction and the slag P205 content was only 0.4%,
corresponding to 98.9~ conversion of P205 to P4.
. .
-16- 1 3 3 5 7 5 2
TABLE I
PARTICLE SIZE DISTRIBUTIONS OF SOLIDS
Cumulative Weiqht % on Each Sieve
U.S.A. Standard Nodule Petroleum
Series Sieve Fines Coke Coke*
No. 8 16.3 57.0 0.7
No. 18 48.2 95.8 14.0
No. 30 59.9 99.6 27.6
No. 50 71.8 99.8 51.3
No. 100 82.8 99.8 72.0
* Low temperature coke; see U.S. Patents 3,140,241 and
3,140,242.
TABLE II
TUMBLE TEST RESULTS
Percent Plus 0.6 Cm. Remaininq
% Binder Green Pellets Cured Pellets
(70% H3PO4) (1 min.) ~20 min.)
8.0 34.9 35.9
10.0 51.3 81.6
12.0 76.3 89.7
-17- 1 3 3 5 7 5 2
.
TABLE III
FURNACE EXPERIMENTS AT 1300C (3 HOURS)
Solids
Experi- Coke Particle % Wt. % Con-
5 ment No. TyPe Size Loss version
1 None As
Added Received 4.7 1.6
2 *Coke As
Received 27.3 58.4
3 Coke As
(repeat) Received 26.8 54.7
4 Petro-
leum As
Coke Received 15.3 31.4
Coke -60 mesh 39.7 81.2
6 Petro-
leum
Coke -60 mesh 39.7 82.9
7 Metal-
lurgical
Coke -60 mesh 40.0 83.5
*See Table I
TABLE IV
EFFECT OF TEMPERATURE ON CONVERSION
OF P2Os TO p~a)
Reaction %
Temp. CConversion
1300 58.4
1300 54.7
1500 97.7
a) Reaction Time: 3 hours
.
