Académique Documents
Professionnel Documents
Culture Documents
477-483
A. H. RODRIGUEZ-HERNANDEZ,1)
MACIEL-CAMACHO. A. W. D. HILLS2) and R. D. MORALES3)
Graduate Center, Instituto Tecn016gico de Morelia, Av, Tecn016gico, Morelia Mich.. Mexico.
1) Instituto Polit6cnico Nacional-ESlalE. Department of Metallurgy. Apdo. Postal 75-874. M6xico D. F.. 07300, Mexico.
Instituto Polit6cnico Nacional-ESlalE, Department of Chemical Engineering. Apdo. Postal 75-874, M6xico D. F., 07300
Mexico. 2) Hills Metaliurgical Consultants. Sheffield, UK. 3) Instituto Polit6cnico Nacional-ESlalE, Department
of Metal]urgy, Apdo. Postal 75-874, M6xico D, F., 07300. Mexico,
The hydration rate of partially recarbonized lime pellets has been studied by employing a single pellet
into a stream of air/vapor water mixtures using a computer aided thermo-gravimettic technique. The hydration
reaction takes piace on a sharp, well defined interface between the product layer and the unreacted core.
It'has been found that the hydration mechanismconsists in two steps. The first observes very slow
hydration rates and the second higher ones. During the first step the water molecules diffuse through a
close porous carbonate structure and the second one begins when these molecules reach the calcium
oxide-calcium carbonate interface to continue through the lime core. A theoretical model was derived to
explain the results of this study. The experimental results were acceptably well predicted with calculations
performed using this model.
Onthe other hand, it wasfound also that the higher is the lime calcination temperature the thinner is the
recarbonized layer formed even under high C02 partial pressures. However, these thin layers are more
effective to increase the hydration resistance of lime pellets than thicker ones. This is an important experimental
finding since it indicates that low recarbonization extents are good enough to increase the hydration
resistance of steelmaking lime without impairing its quality.
KEYWORDS:
Iiquid steel; hydrogen; Iime; kinetics; recarbonization; hydration rate; diffusion coefficient;
layer thickness.
using SiC papers and cleaning them with alcohol. The part of the curve was around nine minutes in most of
sectioned samples were attacked with phenolphthalein the cases, at the higher temperature of I 123K it
was
which turns pink in the presence of alkali. The pink around 45 min.
calcium oxide core contrasted strongly with the white The effect of C02 partial pressure and calcination
calcium carbonate or calcium hydroxide layers (see Fig.
l).
1
o
UJ
a:
J 0.4 H~ 773 K
z
O
F ~ g23 K
o ~ 673 K
~
LL 0.2 ~ I 023 K
~ 112e K
o
o 1o 20 30 40 50 60
TIME (min)
experiments. However, it can be surely affirmed that the lime recarbonization, the extent of recarbonization will
recarbonization rate follows the following ascending be higher in smaller samples; Figure 5 shows the results
order: 22.3 kPa 101.3 kPa. Both
the reaction rate and the extent of reaction at
of experiments carried out at 773 and 101 kPa partial K
one pressure of C02' Again all cases showa similar behavior;
atmosphere (lOlkPa) partial pressure of C02 Were a fast first stage followed by a plateau.
substantially higher than those at low partial pressures. During the recarbonization process there was an
The lowest fractional recarbonlzation (0.07) was ob- increase in ternperature inside the sample, that fact is
tained at 22.3kPa of C02 partial pressure. Moreover, shown in Fig. 6 where it is observed an initial quick
all
curves show similar behavior at any partial pressure increase followed by a plateau and then decrease. Since
a
of C02reaching the plateau at around the sametime. recarbonization should be an activated process, due to
The initial calcination temperature of the pellet is an its diffusive nature, these high increases in temperature
important parameter during the recarbonization reac-
0.8
0.5
O• Iol,3 kPa
~~405 kPa T=673 K 0.7
> ~50 8 kPa Z
0.4 cs25.3 kPa O 0.6
z F
O
F: ~ 28 4 kPa
~L22 3 kPa
~
:~
N O 0.5
CO
z
O 0.3 a:
oo
:c O
UJ 0.4
C!:
T=773 K
OLU
~
Z 0.3 pco,=101 ,3 kPa
J 0.2 O
z F
O A5,8 mm
O o.a
F
O E [17,3
mm
~o. 1
0.1
09,1 mm
o
o o 5 1o 15 220
O 5 1O 15 220
TIME (min)
TIME (min)
Fig. 5. Effect ofsample size on rate and extent ofrecarboniza-
Fig. 3. Effect ofpco, on recarbonization rate. tion,
o.2 1 ~00
pc02=i02 3 kPa
*,'," ~ 1 1 oo
Z
O +''', ~
H 0.15
N ~? I ooo
Z
O
EO
9: a::)
O
UJ o. 1 H
900
OC
J 0_2uJ
Z
O
F 800
O
~ 0.05
LL
700
pco.=56,65 kPa
o 600o
o 5 Io i5 :20 5 Io i5 20
TIME (min) TIME (min)
Fig. 4. Effect ofthecalcination temperature on rateand extent Fig. 6. Temperature of the sample during recarbonization
of recarbonization .
reaction.
in the sample, a fact that should surely change the pellet - None
0.8 A0.1 mm
structure difficulting a fast recarbonization. Indeed, this >
o 0.4 mm
is also observed in Fig. 2, at higher temperatures, the zo D 0.5 mm
recarbonization rate is slower, although, the reaction F 0.6 O 0.6 mm
0.8
mm
extent is higher according to the diffusive nature of the o
>
phenomenainvolved. J
The recarbonated layer and the unreacted core showed zo
sharp interface whenviewed macroscopically, see Fig. F
04
a o
1.
The calcium carbonate layer thickness, a, wasobtained a:
3Wo(1-X) 1/3
(4)
Fig. 7. Effect of the layer thickness on hydration rate.
rl
47cps
0.2
Thepresence of a sharp interface reinforced the argument
for adopting this relationship. a[mm]
O1123 K O,35
3.2. The Hydration Rate ofPartially Recarbonized Lime A1273 K O.5
>
The effects of layer thickness, ~, and calcination tem- z
1373
cl
K0,2
O PH20=2.8 kPa
perature of the initial pellet on hydration rates are pre- H
sented in Figs. 7 and 8 respectively. Figure 7 shows the :c
o
results of six experiments carried out using samples >-
r 0.1
with different layer thicknesses. clear from this figure
It is J
that the hydration rate decreases as calcium carbonate zO
F:
layer thickness increases. There is a steady fractional o
hydration which is followed by a marked change in ~
slope, the hydration rate, however, being lower than
the unrecarbonated sample. The samples with greater
carbonate layer thickness showa very low hydration rate
for about 200 min which is then followed by a steady o
increase. o 500 1ooo i 500 2000
TIME (min)
Onthe other hand, the hydration rate tends to decrease
as the initial calcination temperature of the pellet is Fig. 8. Effect ofthe calcination temperature on hydration rate
for recarbonized samples.
increased a result that coincides with the hydration
behavior of unrecarbonated samples.3) This effect is
of unrecarbonized lime pellets3) assuming that the gas
shownin Fig. 8, where a comparison is madebetween
samples calcined at different temperatures. It is evident
film, shell diffusion and act together as
rate kinetics
resistances connected in series opposing to the diffusion
that the sample calcinated at the higher temperature has
of water vapor which is being driven by the difference
the greater hydration resistance.
between the C02partial pressure in the bulk phase and
Visual examination of these pellets revealed extensive
that in equilibrium on the pellet interface. In the present
cracking. The time to achieve the critical extent of
reaction before the calcium carbonate layer started to case a fourth resistance to the mass transfer of water
vapor is added following the same procedure3) under
crack depends on the layer thickness and calcination
steady state conditions. The final kinetic expression is
temperature.
given as:
3.3. Discussion
3.3.1. Overall Kinetic Model (PA,b-PA,.)
nA= -
As it is seen by the klnetic curves shownin Fig. there 7 1 ~ r r, 1
are very slow initial hydration rates followed by higher + 47~r + 471:r + 47zrfk
4ler2k
o g or * DA,.fl , ri DA,*ff
ones after an incubation period which is longer the thicker .(5)
is the recarbonized layer. To explain these results it is
necessary to develop a kinetic model for the hydration The terms of the denominator of this equation maybe
process of partially recarbonized lime pellets. identified respectively as the gas film, calcium carbonate
layer, calcium hydroxide shell and interface resistances.
Amodel based on the unreacted core shrinking model,
UCSM,was developed to study the hydration kinetics The relative importance of the gas film and interface
_ r~( I )2
(pA,b - PA,.)/ R Tr* p*
"(7)
A 13 1,5
dt
-, f +8r.(1 f) +, (f f 2)+ 1 ~ 0.1 * 3 2.4
O o 5 3.6
1'gkg roDA,.fl DA,*rr k Model
Integrating Eq. (7) and using a further rearrangement o -
gives the integrated rate equation for the hydration rate O 500 1OOO 1500
TiME (min)
of recarbonized llme pellets:
Fig. 9.
on hydration rate (comparison between
Effect ofpll,o
r~(~f- ~f2 +,f3) ~r.(3f'- ~f'2 +f3) experiments and theory).
3r~kg + 3,•.DA,*fl
+
/'.(~f'2 ~f'3) ,f' (pA,b -PA,.)
Table 1. Effective diffusivity for CaC03Iayer.
- + .
t (8)
6DA,*rr k RTr,p~ ..........
Exp. No Ca c n. T 5 (mm} Inoculation Porosity Effective diffusivity
considering that the gas-film is also negligible compared 4 1123 o 80 190 o 22 5.60
with the resistances to masstransfer offered by the layers 5 1123 o 35 90 0.22 2.30
of calcium carbonate, calcium hydroxide and hydration
s 1123 0.10 40 o 22 o 42
kinetics, this layer being calculated employing the ex-
perimental data already reported (Fig. 7of Ref. 3)), the 7 1123 0.47 90 0.22 4.10
Eq. (8) can be considerably simplified according to: 8 1273 o 47 375 0.00 0.98
A plot using the experimental data to calculate the left hydration of the partially
rate recarbonized samples
hand side (LHS) of Eq. (9) against the experimental occurs by the same mechanism of the unrecarbonized
results given by the term in brackets of this equation, ones.3) The only difference is that
an inoculation period
simply defined as e(t), should be a straight line. This is, exists during which water vapor is diffusing through
just the case as is observed in Fig. which shows a plot 9 highly coherent layers of calcium carbonate on the
of the LHSof Eq. (9) against O(t) for the experimental outside of lime.
results shownin Fig. 7. An important feature of this plot From the slopes of the straight lines in Fig. 9 the
is that the extrapolations of these straight lines to zero averaged diffusion coefficients or rather the global
value for O(t) do not coincide with the origin as is the diffusion coefficient. D*, for the steps of steady hydration
case for the hydration of non-recarbonized samples.3) In rates, after the incubation period, shownin Fig. 8, are
fact the interceptions with the LHSaxis correspond to easily derived and presented in Table I Thesecoefficients .
the so called incubation time periods, are four orders of magnitude lower than those deter-
Fromthe above discussion it results evident that the mined for the hydration of unrecarbonized samples3)
water vapor diffuses in a first step through a very low which is a reasonable result if the presence of the re-
porosity layer of calcium carbonate giving as a result a carbonized layer is taken into account.
time period of slow hydration rate called here as an Onthe other hand, giving the very slow hydration rate
incubation period. Once the molecules of water vapor during the incubation period it would be interesting to
are able to reach the calcium oxide~)alcium carbonate know, at least, an estimation of the magnitude of the
interface their diffussion through a more open lime diffusion coefficient of the water vapor in the calcium
structure is facilitated considerably observing, conse- carbonate layer. From the experimental data it is not,
quently, an increase in the hydration rate as is clearly unfortunately, possible to calculate this value. However,
is about 0,1
or O.2
mm, coefficients are of the order
the
o 500 1 OOO 1 500 :2000
of 10 ~ 12 m2s~ 1, whereasthey are an order of magnitude TIME (min)
greater when the layer thickness is about 0.5-0.6mm
0_1 5
regardless of the calcination temperature. Literature
values for the diffusion coefficient of water vapor in pure (b)
PH20=1.2 kPa
calcium carbonate could not be found, but experiments c:~ 1123 K
in zeolites7) have provided values of the order of A 1273 K
1373 K
l0~10m2s~1. Zeolites have an open structure that is o.1 ~ Model
relatively easy for water vapor to diffuse through, so that -
a diffusion coefficient value of l0~12m2s~1 could in- l D
dicate that water is diffusing through the calcium car- CJ
enormous difference between the fractional hydration dicted by the model match fairly well with the experi-
rates for a non-recarbonized sample and that recar- mental results until the calcium carbonate layer cracks
bonized forming a very thin carbonate layer of 0.1 due to the swelling of the calcium hydroxide shell.
thickness as can be seen in Fig. 7. The effective diffu-
mm The critical extent of hydration required to carbonate
sion coefficient for the first case is of the order of layer beginning to crack depends on its thickness, and
8 x 10~6m2 s~ I while in the second one it is of the order the time required to reach such extent of hydration also
of l0~12m2s~ 1. depends on its thickness and porosity. After the layer
The same type of mechanismcould explain the ac- has been broken the sample losses its hydration im-
celeration in the hydration rate, once the inoculation permeabilisation. Figure 7 shows that a O.1 layer
period is completed. This acceleration occurs once the thickness required an extent of reaction of O.2 for
mm
water vapor has penetrated the recarbonized layer and cracking. So, for a 18 at 1373K
is starting to react with lime. The compressive forces
mmpellet calcinated exposed
and a recarbonized layer thickness of 0.2 in
mm
created in the zone of this reaction would, again, place an I .2kPa partial pressure of water vapor atmosphere
the outer layers of the sample under tension, opening up will require more than 150 h to reach such level of hy-
4. Conclusions
k: reaction rate constant (m s~ 1)
kg : masstransfer coefficient in the gas film (m s~ l)
The hydration of partial recarbonized lime
kinetics nA : overall molar flow rate of water vapor to the
pellets as well as the carbonization rates of lime pellets pellet (kmol s~ l)
under streams of air/water vapor and airIC02 gaseous PA,b : Partial pressure of water in the bulk gas (Pa)
mixture respectively were studied using a computer-aided PA,. : Partial pressure of water in equilibrium with
thermogravimetric technique and the following conclu- lime (Pa)
sions have been established: r: radial coordinate (m)
(1) The recarbonization rate oflime is fast during an 1'o : Pellet radius (m)
initial stage, but it slows downas the reaction progresses. :
ri radius of the unreacted core (m)
The recarbonization rate also increases as temperature :
r* unrecarbonized core radius (m)
increases reaching a maximum rate at around 773 K. On R: universal gas constant (kJ mol~I K~l)
further increasing the temperature the carbonization rate t: time (s)
falls. Wo: initial weight of sample (kg)
(2) The higher the partial pressure of C02the higher W: weight of sarnple (kg)
the carbonization rate of lime. X: fraction hydration of lime (-)
(3) The higher temperature the lower
the calcination Greek symbols
the carbonization rate. With high calcination tempera- ~: difference between pellet radius and radius of
tures only a thin layer of calcium carbonate is formed the interface (m)
even at high partial pressures of C02' O: right hand side of Eq. (9)
(4) The masstransfer of water vapor takes place in p~ : molar density of solid (kmol m~3)
two stages; the first is a kind of incubation period where
the water molecules diffuse through the recarbonized REFERENCES
layer and the second one is that where these molecules l) R. H. Borgwardt, N. F. Roacheand K. R. Bruce: Ind. Chem.
are able to reach the calcium oxide~)alcium carbonate Eng. Fbinda'n., 25 (1986), 165.
interface. 2) D. W. Marsh and D. L. Ulrichson: Chem. Eng, Sci,, 40 (1985),
characterized for very small diffusion 423.
The first
stage is
3) A. Maciel Camacho,H. Rodriguez-Herhandez, A. W. D. Hills
coefficients giving, as a consequence, very small hydra-
and R. D. Morales: ISIJ Inl., 37 (1997), No. 5, 468.
tion rates while the second is characterized by higher 4) R. D. Pehlke, W. F. Porter and R. F. Urban: BOFStcelmaking,
diffusion coefficients in a moreopenporous calcium oxide 2, ISS, Warrendale, Pa., (1982), 819.
structure and promoting higher lime hydration rates. 5) O. Kubaschewski and C. B. Alcock: Metallurgical Thermo-
chemistry, PergamonPress, Oxford, England, (1979).
(5) Thin carbonization layers are more impermeable 6) J. Crank: The Mathematics of Diffusion, Clarendon Press,
to water vapor diffusion than thicker ones due, probably, Oxford, England, (1975), I 18.
to the state of stresses in the recarbonized layer which 7) R. M. B. Barrer and B. E. F. Fender: J. Phys. Chem. So!ids, 21
are higher in the later case promoting an opening up of (1961), 12.
the porous structure of calcium carbonate. This is of 8) The Making, Shaping and Treating of Steel USSteel, Pitt. Pa..
(1970), 1970.
enormous practical importance because only small re-