that in the reduction process of phosphate pellets with silica and carbon, the ... neutralises the phosphoms to copper phosphide and water cleans the gas stream ...
OPTIMISATION O F P E L L E T REDUCTION IN A PHOSPHORUS F U R N A C E C. Dresen', I . H L . V o n c k e n ' , W . Schipper^ R. de R u i t e r ^ and M . A . Reuter' ' D e l f t U n i v e r s i t y o f Technology, Faculty o f C i v i l Engineering and Geosciences, Department o f A p p l i e d . Earth Science, R a w Materials Technology Group , Mijnbouwstraat 120, D e l f t 2628 R X The Netherlands ^Thermphos Intemational B V , PO B o x 406, V l i s s i n g e n 4380 A K The Netherlands
Abstract This paper discussed the reaction kinetics o f the reduction o f industrial phosphate pellets f o r P4 production. I n the reduction process o f various phosphate ores, the f o l l o w i n g parameters were investigated: •
D i f f u s i o n o f the phosphorus oxide and other gaseous products f r o m the reaction mixture after having been reduced on the reductant (coke) surface.
•
Influence o f silica on reaction kinetics.
•
Temperature influence.
•
Influence o f mineralogy and morphology ( d i f f e r e n t ores).
»
Particle size influence o f silica and coke.
Experiments were carried out i n a vertical tube fiimace w i t h 2 ores, 2 types o f silica and one type o f coke. Selected experimental conditions simulated fiimace conditions. The f o l l o w i n g conclusions could be made: •
Reaction rate increases vvith the addition o f SiOz into the pellet.
e
The reaction rate increases above 1 3 5 0 ° C ( m e l t i n g temperature o f t h e slag).
•
M i x i n g d i f f e r e n t ore types i n the pellets leads to differences i n reduction kinetics.
Particle size o f the coke particles influence reduction kinetics. The amount o f slag f o r m a t i o n is influenced b y silica addition, coke size and temperature.
EPD Congress 2002 Edited P.R. Taylor Fundamentals of Advanced Materials for Energy Conversion Edited by D. Chandra and R.G. Bautista TMS (The Minerals, Metals & Materials Society), 2002
435
Tntroduction The industrial production o f w h i t e phosphorus is facilitated i n submerged-arc ftimaces^ These fiimaces are f e d w i t h a mixture o f raw materials: apatite (phosphate rock), sihca
lÏltZr)
and coke (reducing agent). The phosphate ore is pe letised to - - e
tha
he f n m a c e bed remains stable and that permeability is maintained. A * o u g h
he
eaetion mechanism o f t h e P4 production is w e l l documented, some unclan y ex.s w i t h regard to the effect o f mineralogy on the reduction mechanism. T h i s paper Z u s s e s the influence o f particle size m i x i n g , temperature and ore source on the reaction kinetics. Phn.sphate Reduction Mechanism The general reduction reaction o f phosphate ore, can be summarised by the f o l l o w i n g reaction equation: Ca,,iPO,),F,+\SC
+ 9zSiO,-^3P^(g)
+ 9[CaOSiO,),]
+ CaF, + \5CO
(1)
A s illustrated by this equation, fluorapatite is reduced i n the P ^ - - - ° f f carbon to f o r m calcium fluorite, calcium silicates, phosphorus gas and CO-gas. The reaction given above can be s i m p l i f i e d to demonstrate the reaction l l ^ i S c i ^ m ^ o s ^ ^ .
mechanism
The first step is the liberafion o f the phojPhorus
fi^^^^
the ore as phosphoms oxide and at the same times the f o r m a t i o n o f ^ ^ I c m m silicates^ This step is considered to be the rate-determining step, since phosphorus ha
to d i f f i i e
f r o m the tricalciumphosphate ore i n a f o r m o f phosphorus oxide. M o s t o f t h e P.O,
is
assumed to be present as P2O5. Ca.iPO,), 2P,0^(g)
+3SiO,^3CaO-SiO,+P.O,Xg) + 2yC-^xP,{g)
+ 2yCOig)
(2) 0)
Finally the liberated phosphorus is transformed into its final f o r m b y cooling below 900°C. 2P,{g)^PAg)
^'^^
D u r i n g the different reactions various parameters seems to have a large influence o f the overall process, w h i c h include; ,
The d i f f i i s i o n ability, that describes the ability o f the phosphorus o'^ije to d i f & s e fiom the apatite to the carbon as w e l l as the removal o f t h e gaseous products the carbon surface after reduction o f P^Oy to P2.
.
The influence o f silica on the reaction kinetics.
a •
Temperature influence. The influence o f mineralogy o f the ores.
436
fiom
•
Partiele size influence o f silica, coke and phosphate ores.
A c c o r d i n g to K o n e v s k i i and Tyunni^"' one o f the most important aspects i n the reduction rate o f phosphate ore is the removal o f the gaseous reaction products. Laboratory tests demonstrated that f o r both sedimentary ore as w e l l as magmatic ore the reaction kinetics were higher when the gaseous products (phosphorus and C O ) were r e m o v e d w i t h a N2 gas current. M u et al.'^'''' concluded from their investigation, that i n the reduction process o f phosphate pellets w i t h silica and carbon, the reaction rate controlling step is the gaseous removal o f reaction products. Silica is an additive, w h i c h promotes the reaction, not o n l y by p r o v i d i n g a thermodynamic d r i v i n g force, but also by m o d i f y i n g the melting phenomena. Besides silica also alumina has a fluxing effect on the reaction kinetics, but less then silica. There are 2 types o f silica; 1) silica already naturally present i n the phosphate ore as feldspar and 2) silica added i n the f o r m o f sand, pebbles or quartz i n addition to the phosphate ore, to accelerate the reaction kinetics and lower the i n i t i a l reaction temperature. A c c o r d i n g to Dom^''^ the influence o f an increasing basicity Si02/CaO, ranging f r o m 0.75-1.1 is described as f o l l o w s : A n increase in basicity f r o m 0.75-0.90 resulted i n a P205-slag content decreasing f r o m l . l l % - 0 . 3 9 % . Further increase o f the basicity to 1.0 leaves the P2O5 content on a level o f 0 . 2 1 % , b u t w i t h this increase also a n e w reaction started to commence. Si02 is reduced by carbon to S i O ( g ) , w h i c h affects the refractories i n the furnace and dilutes the phosphorus product. T h e viscosity o f t h e slag is very important f o r the slag handling. I n the binary C a O - S i 0 2 diagrams an eutecticum can be f o u n d at-a Si02/CaO ratio o f 0.8-0.9. T h i s explains the slaghandling o p t i m u m . lacob and Reynolds'^^ demonstrated that the reduction reaction starts at 1150°C. I n a stable reaction environment the reduction reaction was completed i n 1 hour at a temperature o f 1325 °C and i n 10 minutes at 1500 ° C . R a w Material used i n the Experiments Based on the conclusions o f the previous investigation'*^ the experiments were cairied out w i t h a m i x t u r e o f 2 ore types: •
M a g m a t i c phosphate ore.
o
Sedimenatary phosphate ore
X R D analyses revealed that magmatic ores consist m a i n l y o f fluorapatite, calcite, and dolomite. Sedimentary ores consist o f fluorapatite and quartz w i t h sometimes calcite. V a n der Pas'''', i n a previous study conducted i n co-operation w i t h Thermphos Intemational B . V . , i t was established that a m i x t u r e o f magmatic/sedimentary ore o f 70/30 led to an o p t i m u m in reduction kinetics. For the present investigation, the ores were m i x e d i n a ratio o f magmatic/sedimentary o f 60/40, 70/30 and 80/20 to determine whether this o p t i m u m i n reduction kinetics f o r a 70/30 m i x t u r e was accurate. T o determine the influence o f silica addition into the phosphate pellets on the reduction kinetics, quartz was used.
437
A f t e r the experiments were earned out to determine the e f f e c t o f pure silica in the phosphate pellets, the effect of sand i n the pellets was also investigated T h e an^ vfa
;iSutir'
-
— p i f o s p h a t i t r s - z
For reducing the P.Oy f o r m e d in the pellets during the reduction process Chinese coke
'"^^^'.'g^tion were produced at Thermphos Intemational B V under for r J , S ^ c Ï " " ' " ^'--^"g*- Carbon crucibles w L used und Ï the n,>^ experunents to ensure that optimal reducing conditions were present under the nitrogen atmosphere m the furnace. The graphite material o f t h e crucible contams respectively 0.10 % and 0.60 % of ash and volatL matter
^ " Tfie experimental setup illustrating the t h e r r n d t a f a n c e for the measurement o f reduction kinetics o f phosphorous. Experimental .setup
Jlett'^'T
l""^'^"
"'"•^'^^
^
Carbohte ftimace. D u r i n g the
eantrbottf ' P^"'^"'^^^ ^ ^ ' ^ P - - d through two nintr ^ I " ' " ^ '""^P^"'^ (CUSO4) and water. Copper sulphZ neutralises the phosphoms to copper phosphide and water cleans the gas stream fJom the remaining phosphorus. D u r i n g the reduction process i t is essentfal that A e ea^ remova system o f the experimental set-up be under a slight neSive pressure accomplished by using a water j e t vacuum pump. pressure,
438
A n electronic balance continuously measures the w e i g h t loss o f the material in the crucible on a l i f t i n g tube. Since the entire system has to be airtight, the balance was placed i n a balance box, w h i c h is attached to the underside o f the furnace. The M e t t l e r Toledo balance is connected to a computer, permitting the loss o f w e i g h t to be registered i n an Excel spreadsheet. The balance box w i t h balance and l i f t i n g tube can be lowered using an elevator system. Before each experiment, the flirnace is heated to the required temperature and then the elevator system is lowered to place or replace the crucible w i t h sample. The entire system is elevated again and secured to the furnace w i t h bolts. The fiimace tube is made o f A l S i n t (dense sintered AI2O3), w h i l e the c m c i b l e is placed on a l i f t i n g tube o f silimanite. The length o f this is such that the crucible is placed i n the centre o f the hotspot o f the furnace. The top and b o t t o m flanges are cooled w i t h cooling water, w h i c h is circulated through the system. D u r i n g the reduction experiments, the f i i m a c e is kept under a N2 atmosphere, by flushing the system w i t h an N 2 f l o w o f 0 . 1 5 1/min. A n a l y t i c a l Techniques A f t e r the experiment the cmcibles were weighed to determine the loss o f carbon ( f r o m both c m c i b l e and coke) and the loss o f phosphate pellet. T h e samples from the reduction experiments were analysed w i t h different analytical techniques: •
X - r a y d i f f r a c t i o n ( X R D ) f o r mineralogy determination.
•
X - r a y fluorescence ( X R E ) f o r semi-quantitative elements content.
•
Some specific important samples were analysed w i t h electron microprobe analysis ( E M P ) f o r producing quantitative spot analyses, scans o f polished surfaces
from
sample cuts and photos o f mineralogical and morphological structures. •
Resin was poured i n the crucibles after w h i c h they were crosscut to demonstrate the reaction p r o f i l e . F r o m these crosscut a polished sample was prepared
for
microprobe analysis Experimental results The pellets produced f o r these experiments were also carefiilly examined w i t h respect to pellet strength and pellet porosity. A s the paper focuses on the reaction kinetics, these results are not mentioned i n this paper. Reduction Experiments T o compare the reduction kinetics o f the d i f f e r e n t pellets the assumption was made, that the loss o f weight o f the mixtures measured during the reaction w i t h the balance installation, represented the amount o f P2(gas) and CO(gas) leaving the fumace. The calculation o f the amount o f P2O5 reduction was made by assuming that the reduction reaction took place according the equations 2 to 4. U s i n g this equation the data obtained w i t h the online weight loss measurements could be transformed to graphs demonstrating the amount o f P2O5 reduced versus the time.
439
The amount o f phosphate reduction, calculated according to the reaction equations, d i d not equal the exact amount o f phosphate reduced, shown by X R F analyses o f the reaction products. The calculations made above haye to be con-ected w i t h a number o f con-ections, most notably f o r extra burn o f f o f ' c a r b o n , reduction o f other metals (impurities) i n the ore, such as FezOj, Z n O , M n O , SO3, K2O, etc. A l l data presented have been con-ected f o r these effects. Influence o f silica This paragraph discusses the influence o f silica i n the f o r m o f quartz (SiOa) powder on the reduction kinetics o f phosphate pellets. The experimental results w i l l be demonstrated w i t h P2O5 reduction versus time graphs and tables. The reduction experiments were carried out at 1350°C and 1400°C. F r o m Figure 2 i t can be observed that over a period o f 300 minutes the reduction kinetics are higher w i t h a 60/40 2 0 % S i O j m i x t u r e then w i t h a 60/40 3 0 % Si02 mixture. F r o m the microprobe photos (bottom) can be observed that the amount slag f o r m a t i o n is m u c h higher w i t h 2 0 % Si02 addition (bottom l e f t ) than w i t h 3 0 % Si02 addition (bottom right), i n w h i c h the pellet structure is still visible. I n case o f t h e 3 0 % Si02 addition the pellet tends to f o r m a slag layer at the outside o f the pellet, obstructing the reaction process. F r o m the shape o f the curves can be noticed that different reaction mechanisms take place. The course o f t h e 2 0 % Si02 addition curve can be explained b y the f o r m a t i o n o f p o l y c a l i u m silicates instead o f m o n o c a l c i u m silicates due t o the lack o f silica, w h i c h require a higher temperature and more energy. The reaction temperature should be increased to continue the reaction. To investigate the behaviour o f the pellet reduction process at a higher temperature, the same batch o f experiments was earned out at 1400°C. The experiments were earned out at 1400"C first f o r 90 minutes and later f o r 300 minutes. A t 1400°C, 300 minutes experiments, the addition o f quartz i n the pellet leads to an increasing amount o f P2O5 reduction. I n this case the 60/40 3 0 % Si02 m i x t u r e has been reduced almost completely, w h i l e i n the 60/40 2 0 % Si02 mixtures the amount o f P2O5 reduced f o r 75%. Backscattered electron images (not shown) o f the reaction products demonstrated that f o r (60/40 2 0 % S i O j ) the slag f o r m e d at the bottom o f t h e crucible is not u n i f o n n , b u t still contains phosphate grains and intermediate reaction products, whereas f o r the other (60/40 3 0 % Si02) the slag is completely u n i f o r m , w i t h no traces o f phosphate grains. A l t h o u g h i n i t i a l l y the reduction rate f o r the 20 % Si02 mixture is higher, is slows d o w n after a while. T h e earlier decreasing reduction rate o f the 60/40 2 0 % Si02 m i x t u r e can be explained b y the lack o f Si02 present to f o r m monocalcium silicates, so p o l y c a l c i u m silicates have to be f o m e d .
440
- ^ C O M O :0'V S1O2 {longl 60M0 ;0". S 0 2 (SO nm) S l O I (long] 60M0 30'.^ 5.03 (50m.n|
eOMO 30">
Fig. 2 - Resuits of reduction experiments at 135Ü"C. Shorter curves represent 90 minufe experiments, long curves represent 300 minutes experiments, ^ a g - showj at the right are photographs of crosscut crucibles, immersed in resm. White squares indicate areas which were selected for investigation with the electron microprobe, after of special polished sections from them. Photographs at the bottom are Backscattered Electron Images taken fi-om the polished sections.
reparS
Fig. 3 - Pellet reduction over a period of 300 minutes at 1400°C
441
Influence o f Coke Pnrtinlp Sj^rp
-
no extra coke, only the carbon crucible as reduction agent
^''^
'
ete'ritn"'
»
coke powder as used i n the previous experiments.
^'^^
P ^ ^ - '
60M0 30'.-. S . 0 2 c o l , , , ( , j j „ , „
rig. 4 - Reduction results o f
^Ste^^ïHdTsize
442
experiments
^o^" * e laboratory
D i f f e r e n t ore mixtures Based on preceding investigations (Van der Pas'*') experiments were carried witli tliree ^pes of oïe mixtufes; 60/40, 70/30 and 80/20 magmatic/sedimentaix The «^^^^^^^^^ to detei-mine an optimum in ore mixture composition. The amount of silica added was such th™ he ratio SiO2/CaO=0.88 (called "maximum SiO^"). It was observed that th h o s p h l pellets completely reacted to form a calcium silicate meR The -P-im^^^^^^^ reaction oroducts were weighed to determine the loss of carbon and phosphate ore, in hL aseTe remaining slag. This slag was analysed with XRF analysis to determme t reTaint I t e n t ^ f P.O.. With these measurements and analyses ca cdation could be can-ied out to determine the amount of reduced P ^ O ^ ' J h e ^ ^ ™ demonstrated in Table I . It appears that the reduction kinetics are the highest for the 60/40 mixture. Table I - ?2.0^ Reduction results ore mixtures tmaximum S1O2) ~ ~• Ore Mixture Magmatic/sedimentary
60/40 70/30 80/20
Percentage p^Q^ Rg^m-tion °T= 1350''C
of Percentage of P2O5 Reduction T=1400°C Time = 1 2 0 min
28:4%^
90%
23.8%
81 %
19.7o/„
J2%.
Plant Simulation Rxperiments In the previous paragraph the influence of quartz (silica) addition and the mixtures of differed 0 es iS the pellet on the reduction reaction have been investigated and SrTbed These experiments were carried out with quartz addition, which is an idea nme o m of silica but which for practical/economic reasons can not be used m plant OD eriSns In Ae real plant operations the SiO./CaO ratio ofthe feed to the fomace is S 1 ed by the additL of silica pebbles. To investigate the real the pellet instead of next to the pellet, a set experiments were carried « " ' ^ nrocess in the plant. Pellets were produced with addition of sand, in addition to sand S aced beside the pellets in the crucible. It appeared that the reduction Percentage of S'inc ses from 1 1 % with the 60/40 ref 3 mixture (all the silica is f d e d outside fhe pdle in form sand) to 9 0 % with the 60/40 3 0 % sand addition mixture^ Wife no sand ddition the pellet structore stays visible with a homogeneous pho phoms distribution. When sand is added fee pellet dissolves completely m a liquidus slag.
f ^ ^ Z
Discussion ofthe Reduction Mechanism To determine fee reaction mechanism ofthe pellet reduction first a f «nction^^ "^^^^^ between changing fee reaction parameters inside or outside fee pellet. Through observing and describing fee effects of fee pellet reduction separately a bette unders anding can be obfained of fee overall mechanism. The changing parameters tostde the peflets are the addition of quartz or sand and the Parameters outside the pellets are the size of coke particles and the emoval ot tne Z::T::^:^VroéJs. The temperatore parameter exerts influence bofe mside and
-^^l^lf^^^^^JS.
443
outside tlie pellet. A p o l y c a l c i u m silicates.
higher
temperatare
favours the
formation
o f mono-
or
The PxOy gas has to escape f r o m the pellet first before i t can be reduced by carbon at the coke surface. Before being removed, the P^Oy gas has to be liberated first f r o m the apatite. First a closer l o o k w i l l be taken to the process o f liberation and d i f f u s i o n f r o m the pellet o f the PxOy gas. A f t e r the P^Oy has been removed from the pellets the mechanism o f t h e reduction reaction at the carbon particles can be considered. Influence o f parameters changes inside the pellet
Possible In the previous % the parameiers c facts the conclu when calciums When the SIO; apatite and Ihe I also the silica the PvOv.
•
The percentage P2O5 reduction increases rapidly when a considerable amount o f calcium silicate slag is f o r m e d . W h e n the pellets start to react w i t h the slag, the reaction w i l l progress at a higher rate ^
9
A n increasing amount o f silica addition leads to an increasing amount o f calcium silicate slag formation and thus to an increasing amount o f reduced P2O5. (The reaction has progressed further.).
•
The amount o f slag f o r m a t i o n at 1 4 0 0 ° C is obviously considerable more than at 1350°C. (The reaction has progressed further). The range between 1350°C and 1400°C seems to be a transition area, i n w h i c h the liquidus temperature o f the calcium silicate slag f o r m a t i o n is exceeded.
•
T h e calcium silicate front is m o v i n g from the edges o f the pellet towards the centre: (silica present i n the pellet).
•
I n the reduction experiments w i t h d i f f e r e n t ore mixtures w i t h m a x i m u m Si02 addition, it can be observed that the 60/40 m i x t u r e has the highest values f o r the amount o f reduced P2O5. A n explanation f o r this is that the m i x t u r e 60/40 could benefit more from the qualities o f b o t h ores, as established i n a previous study [ 6 ] . ( A c c o r d i n g to this study, magmatic ore has a lower required reaction heat, but sedimentary ore has a higher reduction rate).
Influence o f coke outside the pellet T h e e f f e c t o f coke is clearly observed i n Figure 4. W h e n the coke size decreases more reaction surface w i l l be available, resulting i n a higher P^Oy reduction rate, w h i c h w i l l favour the liberation o f PxOy fi-om the apatite and finally the f o r m a t i o n o f calcium silicate slag due to the liberation o f CaO. Comparison o f sand and silica addition A comparison was made between the experiments w i t h 3 0 % quartz or sand addition into the pellet. F r o m the results i t is clear that the reduction kinetics f o r sand addition is higher than f o r quartz addition. W i t h sand addition, i n the first stage o f t h e reaction, the reduction rate is faster than w i t h the quartz powder. This can be explained w i t h the presence o f feldspar, w h i c h due to the presence o f alkali's w i l l melt at a lower temperature and react w i t h the calcium phosphate. This initial slag formation may fiinction as an accelerator. W i t h this early f o r m a t i o n o f a liquidus slag the PxOy can d i f f i i s e more easily.
444
In summary, the r pellets at a SiOi/ outside the pelle sufficient cokes 1 The reaction pellet towards the first with the Si02/CaO ratios (cle front the PxOy is li( liberated through which then reads wil gas a melt is formed slag structure via ga-, br reduced to P2 and CO(gai a l o w level, the pellci v. ill The rate-limiting step i n i o f t h e PxOy from the pelk Arguments in favour ot excess o f SiO. in the sla PxOy through reacting wii I n the binary diagram (I increasing amount ot Ca( also becomes less viscou liquidus temperature, ho" because o f t h e presenc o f fluorine (F) liberate F) are laiown to decrea
Possible reaction mechanism I n the previous sections d i f f e r e n t effects have been discussed conceming the effe.ct o f the parameters on the increase o f P2O5 reduction ftom the pellets. F r o m all the listed facts the conclusion can be made that the reduction reaction is increasing more rapidly w h e n calcium silicate slag is formed. W h e n the S i O . has reacted w i t h the calcium phosphate, the PxOy can d i f f u s e from the apatite and the remaining CaO can react to f o r m a calcium silicate slag. I n this process also the silica r i c h slag reacting w i t h the CaO c o u l d be the d r i v i n g force o f liberating the P x O y .
I n summary, the mechanism may be described as f o l l o w s . W h e n silica is added i n the pellets at a S i O j / C a O ratio o f 0.88, the reaction w i l l be much faster than w i t h silica outside the pellet. I n this reaction the temperature has to be at least HOO-'C and s u f f i c i e n t cokes have to be added (Stage 1).
the
The reaction front w i l l be m o v i n g i n the f o r m o f a m e l t f r o n t from the edges o f the pellet towards the centre (Stage 2). I n this melt f r o n t the S i O . i n the pellet w i l l react first w i t h the melt f o l l o w e d b y the CaO. This e f f e c t was observed w i t h measurmg SiO./CaO ratios (electronmicroprobe) i n the slag i n the m e l t i n g fi'ont. A t this reaction f r o n t the PxOy is liberated from the apatite. I t is not clear whether a l l the PxOy is liberated through the reaction o f CaO w i t h the slag or that some CaO is first f o r m e d , w h i c h then reacts w i t h the slag after the PxOy is liberated. D u e to the generated PxUy gas a melt is f o r m e d o f a f o a m i n g stmcture (Stage 3). The PxOy gas is f o r c e d out o f t h e slag stmcture v i a gas bubbles. The PxOy gas moves towards the coke where it is reduced to P . and C O (gas). W h e n the amount o f PxOy i n the pellet/slag is decreased to a l o w level, the pellet w i l l collapse and f o r m a l i q u i d slag (Stage 4 and 5). The r a t e - l i m i t i n g step i n this process can be either the liberation o f PxOy or the removal o f t h e PxOy from the pellet towards the coke. Arguments i n f a v o u r o f t h e parameter o f liberation o f t h e PxOy are the f o l l o w i n g . The excess o f S i O . i n the slag ( S i O . seems to dissolve first) favours the liberation o f t h e PxOy through reacting w i t h the CaO o f t h e apatite. I n the binary diagram (Figure 5) can be seen that starting out w i t h silica, w i t h an increasing amount o f CaO the liquidus temperature o f t h e slag is decreasing^ The slag also becomes less viscous. This m i g h t be the thermodynamic d r i v i n g force. The actual liquidus temperature, however, is about 50 degrees lower then i n the binary diagram, because o f t h e presence o f K and N a i n the fiimace feed, and because o f the presence o f fluorine (F) liberated from the apatite-ore (Ca5(P04)3F). A l l three elements (Na, K , F) are k n o w n to decrease the liquidus temperaUire o f silicate melts considerably.
can
445
Mole fraction S i O j
»•+
0.6
7-2CaO.SiO 0
20
CoO
F i g . 5 - The system CaO-SiO.. The effect o f P,Oy removal as rate controlling element can be explained assuming the viscosity o f the slag decreases w i t h more CaO addition. Conclusions F r o m the study the f o l l o w i n g conclusions can be drawn: «
The reduction rate w i l l increase w i t h the addition o f f i n e silica into the pellets, instead o f m the f o r m o f sand outside the pellet.
' «
n S O " ? " * °^
*^
considerable larger at 1400°C than at
M i x i n g o f magmatic and sedimentary ores i n the pellet causes differences i n reduction kmetics A t 1350°C and 1400°C the 60/40 (magmatic/sedimentary) w i t h 30/o S1O2 has the highest reduction rate. The particle size o f the coke particles influences the reduction kinetics W i t h a smaller coke particle size the reduction rate is increased. The differences i n particle sizes are large: coke powder versus particles screened at 10 m m . W i t h smaller coke sizes the reaction surface o f the reduction agent increases, w h i c h leads to more rxUy reduction.
\V^ltZ°7V^lV'V
P°f'•'•'"'^^ rate controlling parameters. Either the liberation o f the P^Oy f r o m the apatite or the d i f f e s i o n o f the P^Oy to the coke
446
•
particles. At this moment is not possible to determine, which parameter is overall rate controlling. Both have influence in this set o f experiments. The reduction kinetics of sand addition is higher than with quartz addition. References
1. M.R. Konevskii and M.P. Tyunni, "Influence of the conditions of removal of gaseous reaction products on the rate of phosphoms formation by reduction from melts", Russian Joumal of Applied Chemistry, translation from the Zhumal Prikladnoi Khimii (Plenum Publishing Corporation, 1985), pp. 355-356. 2. J. M u , F. Leder, W.C. Park, R. A. Hard. J. Megy and H . Reiss, "Reduction of Phosphate Ores by Carbon: Part I . Process Variables for Design of Rotary Kiln System", Metallurgical Transactions B, 17B (December 1986), pp. 861-868. 3. J. M u , F. Leder, W. C. Park, R. A. Hard. J. Megy and H . Reiss, "Reduction o f Phosphate Ores by Carbon: Part. I I . Rate Limiting Steps", Metallurgical Transactions B. 17B (December 1986V pp. 869-877. 4. F.W. Dom, "Gewinnung von Phosphor im ElektroflieBbetofen", Chemie-Ing.Tech. 45. lahrgang. 16 (1973), pp. 1013-1019. 5. K.D. Jacob and D.S. Reynolds, "Reduction of Tricalcium Phosphate by Carbon", Industrial and Engineering Chemistry. 20 (1928), pp. 1204-1212. 6. D. van der Pas "Kinetic study conceming the reduction of phosphate ores" (M.Sc. thesis Delft University of Technology 1999). Also: M . A . Reuter, D. van der Pas and R. de Ruiter; "Synergetic effects during phosphorous production in submerged-arc fiimaces". In: P.R. Taylor (eds.); Proc. o f t h e EPD Congress 2000. Sponsored bv the Extraction and Processing Division of the Minerals. Metals & Materials Society (TMS). 12-16"^ March 2000. (Nashville, Tennessee, USA), 2000, pp. 39-50.
447
