Cyclic Calcination and Recarbonation of Calcined Dolomite Samuel Dobner, Lauris Sterns, Robert A. Graff,' and Arthur M. Squires Department of Chemical Engineering, The City College of The City University of New York, New York, New York 1003 1
Results on the cyclic calcination and recarbonation of calcined dolomite at atmospheric pressure and at 300 psig are presented. The effects of various reaction parameters (i.e., temperature, pressure, and gas composition) on the recarbonation rate and on the change in solid reactivity with cycling were explored. Addition of steam to the recarbonation atmosphere resulted in nearly a two order of magnitude increase in the recarbonation rate. Increasingthe temperature and the partial pressure of carbon dioxide for calcination gave lower recarbonation rates and accelerated the loss of solid reactivity with cycling. Reactivation by subsequent calcination in nitrogen was demonstrated.
Introduction Solid acceptors for carbon dioxide, used cyclically at high temperatures, have found applications in gas separation and purification and as a source for process heat. The solid acceptors are most often in the form of metal oxides prepared from naturally occurring minerals. Minerals which have received considerable attention are limestones and dolomites. Calcium oxide has been proposed as an acceptor for several reactions, including desulfurization, lignite gasification and carbonization, shift conversion, steam reforming of hydrocarbons, catalytic cracking of hydrocarbon oils, and the reduction of metallic ores. In some of these reactions the solid may serve to catalyze the reaction as well. In dolomitic acceptors only the calcium species are reactive. The magnesium species, other than their possible catalytic properties, are chemically inert and only serve to support the microstructure by finely dispersing the small crystallites of the active calcium species. Very often, the resulting superior durability or reactivity of dolomitic acceptors will on balance outweigh considerations of its lower theoretical capacity in comparison to limestones. The principal reaction in many of the applications of limestone and dolomitic acceptors is the cyclic calcination and recarbonation of the CaO via CaO
+ CO2 = CaCO,;
AH298 K
= -42.33 kcal/mol
The recarbonation reaction usually serves two functions: to provide a source of high-temperature process heat to endothermic reactions and to selectively remove diluent carbon dioxide generated by these reactions. This is best illustrated by the COz-acceptor process for gasifying lignite (Curran et al., 1967), where heat for endothermic gasification reactions with steam is provided by the in situ recarbonation of fully calcined dolomite. Removal of carbon dioxide generated during gasification also considerably upgrades the heating value of the make gas. Regeneration of the acceptor is carried out in a separate calciner where heat for calcination is provided by in situ combustion of a portion of ungasified char with air. Recarbonation of calcium oxide prepared from a range of materials has been investigated extensively over the past four decades (Bischoff, 1950a; Britton et al., 1952; Cremer and Nitsch, 1959; Curran et al., 1967, 1970; Davtyan et al., 1961; Dedman and Owen, 1962; Glasson, 1960; Gluud et al., 1930; Hyatt et al., 1958;Ketov et al., 1970; Nitsch, 1962;Ohno, 1957; Ohno and Fujiyama, 1957; Richer, 1954, Richer and Vallet, 1961; Schwob, l949,1950a,b; Shushinov and Fedyakova, 1958; Siske and Proks, 1958a,b; Zawadski and Bretsnajder, 1933a,b, 1935, 1938). There have also been studies reported for the recarbonation of fully calcined dolomite as well (Asboth, 1952;
Bischoff, 1950b; Britton et al., 1952; Curran et al., 1967,1970; Curran et al., 1964; Noll, 1950; Richer, 1954; Schwob, 1949, 1950a,b),although here the information is much more sparse. In general, it has been found that fully calcinated dolomite will recarbonate to a greater extent than any form of calcium oxide (Bischoff, 1950; Britton et al., 1952; Curran et al., 1967, 1970; Richer, 1954). This has been attributed to the microcrystallinity of the calcium species in dolomitic materials and to the microporous magnesium oxide supporting structure. The only sizeable body of data on the cyclic recarbonation of fully calcined dolomite is that provided by Consol Coal for their Con-acceptor process (Curran et al., 1967,1970; Curran et al., 1964) and by the Gesellschaft fur Kohlentechnik for their CO-shift process (Gluud et al., 1930). We therefore undertook exploratory experiments for cyclic calcination and recarbonation of dolomite. In particular, we examined the effects of calcination conditions, total pressure, and partial pressures of carbon dioxide and steam during recarbonation. Our work on recarbonation was preparatory to study of species derived from dolomite as acceptors for hydrogen sulfide, and served to shake down the experimental setup to be committed to the latter study. Experimental Section The calcination and recarbonation of fully calcined dolomite are most easily studied by following the weight changes taking place during reaction. Three different series of experiments are described below. Two samples of Greenfield dolomite were used in the experiments. The assays and particle sizes for the two samples are shown in Table I. These samples were used earlier by Pel1 (1971) and Ruth (1972) in their studies of the reaction of hydrogen sulfide with fully calcined and half-calcined dolomite, respectively. The first series of experiments were carried out in a duPont 950 TGA using sample A. The purpose of these experiments was to determine the effects of calcination conditions on the recarbonation rates in 1atm of carbon dioxide. In these tests calcination was carried out nonisothermally at 10 OC/min to a preselected temperature in the desired gas atmosphere and then held at the maximum temperature for a given period of time, again in some preselected gas atmosphere. The sample was then cooled in nitrogen to near room temperature, whereupon recarbonation was carried out nonisothermally a t 10 "C/min in 1 atm of carbon dioxide. At all conditions tested complete recarbonation was achieved below 900 O C. The effects of calcination atmosphere, hold atmosphere, hold temperature, and hold time were explored in this series of experiments. A second series of experiments was carried out using an Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4 , 1977
479
Table I. Particle Sizes and Assays of Greenfield Dolomite
100
I
I
I
I
I
"
Samoles
Particle size Atomic ratio, Ca/Mg Assay, weight % Magnesium Calcium Iron Silicon Aluminum Others
CYCLE
Sample A
Sample B
-80 +350 mesh 1.09
-250 +250 mesh 1.01
12.32 22.04 0.37 0.x 0.x 0.X or less
13.48 22.43 0.18
80
PER CENT CONVERSION TO
CP co3
60
40
0.x 0.x 0.OX or less
20 t/ /A/ n
Results The results of the first series of experiments are summarized in Table 11. The relative reaction rates for nonisothermal recarbonation may be expressed in terms of the temperature, T1/2,a t 50% recarbonation. The lower this temperature, the more reactive the material. Between 925 and 975 "C there appears to be only a small effect of increasing hold temperature on reactivity. Increasing the hold time from 1 to 4 h increased the value of TI12 by 50 "C a t 925 "C and by 20 "C a t 975 "C. The most reactive material is formed when calcination is carried out in nitrogen to 800 "C, with no hold period. This material is approximately 30 to 40 times more reactive than that produced by calcining in carbon dioxide. Addition of a hold period a t 975 "C in nitrogen increases TI12 by only 40 "C. In contrast, the addition of a hold period in carbon dioxide increases TI12 by more than
1 ATM CO,
L
"0 Ainsworth automatic recording balance (Pell, 1971). Calcination of the raw stone was carried out nonisothermally at 10 to 15 "C/min to 925 OC in 1 atm of carbon dioxide. The sample was then held a t 925 "C in nitrogen for an additional 15 to 30 min before being cooled to recarbonation temperature. Unlike the first series of experiments, recarbonation was carried out isothermally at some predetermined temperature. Conditions for subsequent calcinations varied from run to run. In this second series of experiments the cyclability of calcined dolomite (sample A) was explored a t two temperature levels and in various gas atmospheres containing carbon dioxide, nitrogen, and steam. The last series of experiments was carried out in a highpressure thermobalance incorporating a duPont 950 TGA as described by Dobner et al. (1976). Calcination was carried out nonisothermally a t 20 "C/min to the maximum temperature selected and held there for an additional 10 min in nitrogen. The sample was then cooled in nitrogen to the selected recarbonation temperature. In this last series of experiments the effects of calcination conditions and the partial pressures of carbon dioxide and steam during recarbonation were explored a t a pressure of 300 psig. Sample B was used for this series of experiments.
550°C
100 200 300 TIME, mins.
Figure 1. Cyclic recarbonation of calcined dolomite at 550 "C in C02 at atmospheric pressure.
150 "C, and is only 35 "C lower than that obtained by calcining in carbon dioxide. It therefore appears that the factors most strongly affecting the reactivity of the calcinate are hold atmospheres and calcination atmosphere (and hence the calcination temperature). In the second series of experiments the cyclic recarbonation of calcined dolomite was explored at atmospheric pressure. Figure 1 shows results for 9 cycles carried out a t 550 "C in carbon dioxide. The observed loss in reactivity with cycling is characteristic of the results obtained with this material. The catalytic influence of steam on recarbonation is readily seen from a comparison of the time scales in Figures 1and 2. The simple substitution of 50% of the carbon dioxide by steam can accelerate the recarbonation rate by a factor of 30 to 100 times. Cycling experiments were carried out in order to assess the extent of deactivation a t these more favorable reaction conditions. Two runs of 23 and 31 cycles were completed a t 550 and 700 "C, respectively. The results are shown in Figures 2 and 3. Calcinations in both these runs were conducted in the recarbonation gas atmosphere to 870 "C. The initial reaction rates do not appear to suffer with continued cycling. Deactivation a t these conditions occurs primarily by formation of inactive or "deadburnt" calcium oxide species in the solid. In the final series of experiments cyclic recarbonation a t pressure was explored. Figure 4 shows results for 7 cycles carried out a t 650 "C and 300 psig with 4 atm of carbon dioxide. A maximum temperature of 1030 "C is needed to effect calcination a t these conditions. Again, we observe the characteristic deactivation behavior. The effect of partial pressure of carbon dioxide on the recarbonation rate was explored. Results shown in Figure 5 for recarbonation a t 700 "C and a total pressure of 300 psig suggest that the reaction is approximately first order in partial pressure of carbon dioxide. Calcinations for these runs were
Table 11. Nonisothermal Recarbonation of Calcined Dolomite (Conditions: 1 atm of COz, 10 'C/min) Calcination Hold Hold Hold Temp @ 50% Run no. atm atm temp, "C time, h conversion, "C 82-6 82-1 82-2 82-4 82-8 82-7 82-3 82-5 82-9 480
coz
c02 CO2
co2 c02
coz N2 NZ NZ
cos coz cos co2 coz c02 N2 N2
cos
Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
400
925 950 940-970 975 925 975 800 max. 975 975
1 1 1 1 4
4 5 min 1
1
668 643 651 691 707 710 475 515 629
100
CYCLE NUMBER
4 ATM COz 300 PSlG
80
PER CENT CONVERSION TO cq co,
60
--
40
20
IO
'0
0
l
0
I
I
20
/
I
40
l
60
TIME, mins.
20
30
40
SO
TIM E, MINUTES Figure 4. Cyclic recarbonation of calcined dolomite at 650 "C in 4 atm of COZat 300 psig.
Figure 2. Cyclic recarbonation of calcined dolomite at 550 "C in 50/50 C02/H20 at atmospheric pressure.
'lT=-l
t
1
-1
80
PER CENT CONVERSION TO Ca CO, 40 2o
1
PER CENT CONVERSION TO Ca COS
700'C 0 . 5 ATM CO, 0.5 ATM HO ,
it
CI
1
40 20
L 30
OO
4
.o
IO
20
20
TIME, mins. Figure 3. Cyclic recarbonation of calcined dolomite at 700 "C in 50/50 C02/H20 at atmospheric pressure.
carried out a t the same total pressure to 930 "C in 1 atm of carbon dioxide. Additional data, not presented here, showed the first-order dependency on carbon dioxide pressure to be valid down to COSpartial pressures of 0.5 atm, but the order approaches zero a t 0.1 atm of CO2 (Dobner, 1976). The effect of total pressure is more difficult to assess. In general, we found that reaction rates at pressure are somewhat slower than at atmospheric pressure, for equivalent partial pressures cf carbon dioxide. This suggests that mass transfer may play a significant role in determining the reaction rates at these conditions. The catalytic effects of steam on the recarbonation reaction a t higher pressure was also explored. Results for addition of 4 atm of steam a t 700 "C, 300 psig in 4 atm of carbon dioxide are shown in Figure 6. A run at similar conditions but with no steam present is included in Figure 6 for comparison. We find that the addition of 4 atm of steam to 4 atm of carbon dioxide results in a 15 to 50 times improvement in recarbonation rate. In agreement with the earlier nonisothermal recarbonation experiments it was found that the calcination temperature strongly affects the reactivity of the calcined dolomite. This is illustrated in Figure 7 where the recarbonation rate of a stone calcined a t a pressure of 300 psig to 1030 "C in 4 atm of COZwas found to be significantly slower than for a stone that had been calcined at the same total pressure at 930 "C in 1atm of Con.
40
60
TIME, mins.
Figure 5. Effect of CO1 partial pressure on cyclic recarbonation of calcined dolomite at 700 "C and 300 psig. Discussion Our experimental results indicate that the kinetics for the recarbonation of calcined dolomite are complex. In large part the complexities can be attributed to the important role played by topochemistry in association and dissociation reactions, an area very often overlooked in discussions of gassolid kinetics. Where possible, the experimental results will be interpreted in terms of our knowledge of the topochemistries of the reaction. Effect of Partial Pressure of Carbon Dioxide on Recarbonation Reaction. The recarbonation reaction in the absence of steam appears to be first order in carbon dioxide, a t least for partial pressures above 0.5 atm. Reaction orders of zero (Dedman and Owen, 1962) and unity (Nitsch, 1962; Zawadski and Bretsnajder, 1933a,b, 1935, 1938) have been reported in the literature for the recarbonation of calcium oxide. Only a t very low partial pressures were reaction orders approaching zero obtained. Zero-order kinetics is consistent with a reaction model proposed by Hyatt et al. (1958)and later by Nitsch (1962), if the formation of active calcium oxide, CaO*, is assumed to be rate determining in their reaction sequence CaO * CaO* CaO* COZ e CaC03 A variable reaction order which increases from zero to unity with increasing partial pressure can also be explained by the not too unreasonable assumption that the rate constant for
+
Ind. Eng. Chern., Process Des. Dev., Vol. 16, No. 4, 1977
481
0 X
4ATM STEAM
---
NO
0 CALCINATION IN 1 ATM COz X CALCINATION IN SATM CO, FIRST CYCLE THIRD C Y C L E
STEAM
---
FIRST CYCLE THIRD CYCLE
1001
I
i
i
I
I
1
I
80 300 pssg, 7 0 0 ' C 4ATM C0,
PER CENT CONVERSION TO
Ca CO,
-
60
PER CENT CONVERS ION TO
cs co,
40 20
0
l
0
,
10
,
l
20
l
30
TIME, mins. Figure 6. Effect of steam on cyclic recarbonation of calcined dolomite at 700 O C in 4 atm of COz at 300 psig.
CaO* formation, k * , is itself dependent on the partial pressure of carbon dioxide (Cremer and Nitsch, 1962b; Nitsch, 1962; Zawadski and Bretsnajder, 1933a,b, 1935, 1938). We find our data to be well correlated by
k* = CY iPPco2 with @/CY = 3.1. In steam-containing atmospheres the initial recarbonation rates are extremely high (Figures 2, 3, and 6). Temperature and cycle number have little influence on the initial rates but strongly affect the final asymptotic level of conversion. The invariance of initial rate suggests that mass transfer effects may be controlling. Extent of Reaction and Reaction Rate. The rate of reaction, r, in gas-solid reactions is sometimes given by r =f(a)g(P,T)
where a is the fractional conversion of solid, and { ( a )is assumed to be invariant with temperature, time, and gas environment. While this simplification is often used in the analysis of nonisothermal scans (thermograms) (Sestak et al., 1973), it is in fact seldom valid except over a small range of experimental conditions. No simple correlation of the data presented here was possible using functions f( a ) having physical significance. In general, more than one equation is needed to fit the data for any single run. The kinetic constants are then functions of temperature, gas environment, and cycle number. The pattern of reaction behavior is also very different depending on whether steam is present or not. This becomes readily apparent when trying t o relate the behavior in Figure 1t o that of Figure 3. These difficulties are characteristic of true topochemical reactions where the progress of the reaction interface is governed not only by the local gas environment and temperature but also by the nature and time-temperature-environment history of the underlying solid reactant substrate and the developing solid product phase (Galwey, 1967; Garner, 1955; Pannetier and Souchay, 1967; Young, 1966). Temperature Coefficient of Reaction Rate. Analysis of results from the nonisothermal experiments using the integral method of MacCallum and Tanner (1970) and assuming the reaction to be first order in unreacted oxide gives an activation energy of about 20 kcal/mol. Arrhenius plots could be drawn 482
Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
TIME, mins. Figure 7. Effect of calcination conditions on cyclic recarbonation of calcined dolomite at 700 "C in 4 atm of CO? at 300 psig. for isothermal runs (not containing steam), to give a similar activation energy. The reactivities of the various solids described in Table I1 are therefore reflected only by the magnitude of their preexponential factors, which varied by a factor of 80. No activation energies were obtained for recarbonation in steam-containing atmospheres. Recarbonation of Calcined Dolomite Relative to Other Forms of CaO. There are two factors of importance in discussing acceptor utilization: reaction rate and capacity. The reaction rate is governed by the size and lattice strain of the crystallites of solid reactant as well as the macropore structure of the material. The capacity of the acceptor, on the other hand, is governed by the microporosity of the product layer formed behind the reaction interface. For CaO reacting to CaC03, the product layer will have a much larger specific molar volume than the substrate so that closing off of pores to the reaction interface will be a limiting factor to achieving high capacities (Glasson, 1960; Ohno and Fujiyama, 1957; Richer, 1954). Single crystals of calcite, for example, when calcined in nitrogen a t 700 OC will only recarbonate to 50% in 645 mmHg of carbon dioxide a t 700 " C and will require more than 10 h to achieve 40% conversion (Hyatt et al., 1958). The low surface area of the calcia accounts for the low reaction rates and the very compact growth of the recarbonated calcite phase prevents carbon dioxide from diffusing to the reaction interface. Precipitated CaC03 will produce finer crystallites of CaO, but here again the reaction falls short of completion and requires nearly 5 h to achieve 40% conversion (Hyatt et al., 1958). Lime reacts fairly rapidly under the same conditions requiring less than 1 h to achieve the same levels of conversion (Ohno, 1957; Shushinov and Fedyakova, 1958). Impedance by the product layer is still a problem and maximum conversion is limited to about 70 to 800?(Bischoff, 1950; Britton et al., 1952; Glasson, 1960; Nitsch, 1962; Ohno and Fujiyama, 1957; Richer, 1954; Siske and Proks, 1958a,b). Calcined dolomite (prepared in carbon dioxide), on the other hand, reacts completely at 700 "C, requiring only 30 min to achieve essentially complete conversion. Its high reaction rate relative to lime is attributed to the microcrystallinity of its CaO species. The size of the CaO crystals in calcined dolomite (first cycle) is only 400 A (Curran et al., 1970; Haul and Schoning, 1952) compared to 2000-3000 A for limes (Gasson, 1961). The MgO in calcined dolomite, while being inert toward COz a t these conditions, is capable by virtue of its ultrafine dispersion in the solid of forming a microporous structure which facilitates the ex-
100
I
/
I
1
I
I
PERCENT COWERSION TO cq
co,
60 40
-
2ol OO
100
200
4
C Q
-00°C
PER CENT CONVERSION TO
30
C
5
I5
IO
20
CYCLE NUMBER
ca CO, :O 5 ATM CO,1
700°C
10 5 ATM H 2 O j
300
TIME, mins.
Figure 8. Effect of steam addition during recarbonation of calcined dolomite a t 700 "C in 0.5 atm of CO? a t atmospheric pressure (tenth cycle). change of carbon dioxide to and from the particle, therefore resulting in high capacities for this material (Richer, 1954). Steam Catalysis of Recarbonation Reaction. Steam may be termed an "ideal catalyst" (Schwab et al., 1937) as a result of its ability to strongly catalyze both the decomposition of CaC03 and its formation from CaO (MacIntire and Stansel, 1953). Bischoff (1950), in particular, has demonstrated that steam catalyzes the calcination of half-calcined dolomite. The catalytic recarbonation of lime in steam is only briefly mentioned in the literature (Davtyan et al., 1961; Nitsch, 1962; Schwob, 1949, 1950a,b). Curran et al. (1970) suggested that steam also catalyzes the recarbonation of calcined dolomite but gave little data. Our results confirm the strong catalytic influence of steam on the recarbonation reaction. At 550 "C and atmospheric pressure the replacement of half the carbon dioxide by steam was found to increase the recarbonation rate by a factor of 30 to 100. We originally believed that steam plays a very important role in the formation of the CaC03 product layer. Other investigators (Garner, 1955; Young, 1966) have reported the ability of steam to promote fissurization of the product layer. The catalytic effect of steam was therefore thought to be the result of opening up of the reaction interface to the reactive gas atmosphere. Figure 8 shows the result of a test of this hypothesis. In this run, at 700 "C, steam was introduced after the solid had reached a steady level of conversion. As expected, the rate increased considerably with steam addition. After about 70 min the original gas atmosphere was returned. Fissurization of the product layer by steam should have reactivated the solid considerably, yet the solid remained just as inactive as prior to steam addition. Apparently, the catalytic effect of steam has no "memory". We may speculate that its catalytic nature derives rather from an ability to promote nucleation of the "active CaO", which is proposed to be rate limiting. Reactivity a n d Calcination Conditions. The effect of calcination temperature on the reactivity of calcined dolomite toward recarbonation is illustrated by the results shown in Table I1 and Figure 7 described earlier. In general, the higher the calcination temperature and hence partial pressure of carbon dioxide the less reactive the resulting oxide. Soaking of the calcined solid in carbon dioxide at temperatures above 925 "C will also result in a severe loss of reactivity. These observations confirm similar findings reported by other investigators (Bischoff et al., 1959; Cremer and Nitsch, 1962b; Curran et al., 1970; Ohno, 1957; Ohno and Fujiyama, 1957; Richer, 1954).
15
IO
20 25 30
CYCLE NUMBER
Figure 9. Effect of cycling on recarbonation of calcined dolomite in 0.5 atm of CO2 at atmospheric pressure and at 550 and 700 "C.
PER CENT CONVERSION
TO C a co,
'O1
4 56789 CYCLE NUMBER 2
3
Figure 10. Effect of cycling on recarbonation of calcined dolomite at 650 "C in 4 atm of COa at 300 psig.
Another effect of increased calcination temperatures is the deactivation of the solid with repeated cycling. The data in Figures 2,3, and 4 are replotted in Figures 9a, 9b, and 10 as a function of cycle number, N , at constant reaction times, t . The data are well correlated by at = aN-b
The slope, b, in the log-log plots may be looked upon as a form of deactivation rate constant. The larger this value the greater will be the falloff in conversion with repeated cycling. Figure 11 is a plot for data reported by Curran et al. (1967,1970) for calcined dolomite in a continuous reactor. Table I11 lists the values for b and their corresponding experimental conditions. The increasing value for b with increasing calcination temperature reflects a higher degree of deactivation of the solid at the higher temperatures. The effects of calcination conditions on reactivity may be understood in terms of the topochemistries taking place (Dedman and Owen, 1962; Glasson, 1958,1961; Hedin, 1961; Kovalenko, 1966; Nitsch, 1962; Yanev et al., 1970). When the CaCO3 species in half-calcined dolomite decomposes, there first occurs the epitaxial growth of CaO at points of high lattice strain, such as dislocations, holes, and grain boundaries. This metastable crystal structure results in the formation of a dense and amorphous-like product layer a t the reaction interface. Recrystallization of the CaO to its normal habit soon follows but usually lags behind C02 evolution. Depending on the temperature and gas environment the resulting CaO crystalInd. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
483
Table 111. Deactivation Rate Constant in Cyclic Recarbonation of Calcined Dolomite CY+
PT, atm
Reference Curran et al. (1967)a Curran et al. (1967) This work This work This work
= aN-b
Calcination conditions Pco2, Pn20, atm atm 3.6 2.6 4 0.5 0.5
20 11 21 1 1
T, "C
-
1060 1020
0.5 0.5
870 870
1030
Recarbonation conditions Pco2, Pn20, T, atm atm "C 0.75
3.9
0.87 4 0.5
2.8 0.5
0.5
0.5
-
815 815 700 700 550
b
0.592 0.305 0.350 0.296 0.255
Char combustion in calciner.
PER CENT CONVERSION
TO Ca COB
lo
L IO
20
40
GO 80
CYCLE NUMBER Figure 11. Effect of cycling on recarbonation capacity for calcined dolomite cycled in Consol Coal (CONOCO) continuous bench scale reactor (Curran et al., 1967). lites will begin to sinter. Sintering may be defined simply as a bulk migration of ions in the crystal lattice. The effect of sintering is twofold: an increase in crystallite size and the annealing of lattice imperfections, such as dislocations and holes. The former reduces the specific surface area of the solid formed while the latter reduces the concentration of potential nucleation sites for subsequent reaction. The rate of sintering of CaO in various gas atmospheres is reported to increase in the same order as its affinity for reaction with the gas atmosphere; SO2 > CO2 > H20 > 0 2 , air > N2 (Clark, 1949). On the basis of this criterion an oxide with superior reactivity will be obtained if the calcination is carried out in nitrogen. Carbon dioxide, in particular, was found to activate the sintering of CaC03 (Glasson, 1961; Proks and Jaskova, 1967; Proks and Siska, 1968). MgO (Bachmann and Cremer, 1961; Noll, 1950) and CaO (Bachmann and Cremer, 1961; Cremer and Nitsch, 1962a,b; Glasson, 1961; Hashimoto, 1961; Hedin, 1961, 1962; Ohno, 1957; Ohno and Fujiyama, 1957; Richer, 1954). Peterson and Cutler (1968) present data for activated sintering of CaO by steam a t 920 to 1123 "C. Surface migration in calcium carbonate can occur as low as 264 OC; bulk migration takes place a t 533 "C (Dedman and Owen, 1962). Above 800 "C annealing of calcium carbonate is quite rapid (Garner, 1955). As a result, soaking of calcium carbonate a t temperatures below decomposition will reduce its subsequent decomposition rate (Garner, 1955). Magnesium oxide crystallites formed from both magnesite and dolomite are reported to sinter and grow in carbon dioxide atmospheres above their decomposition temperatures (800 "C) (Bachmann and Cremer, 1961; Noll, 1950). Sintering of magnesium oxide in Con-free atmospheres becomes important a t temperatures above 975 "C (Pampuch, 1958). 484
Surface migration of ions in calcium oxide becomes significant above 675 "C (Glasson, 1961); bulk migration and crystallite growth take place rather suddenly a t 900 "C (Fischer, 1955; Glasson, 1958,1961; Pampuch, 1958;Tagawa and Sudo, 1959). This accounts for the drastic reduction in capacity reported by Curran et al. (1970) for isothermal cyclic recarbonation of calcined dolomite a t 954 "C in comparison to runs made below 887 "C. The equilibrium size of product crystallites in both calcination and recarbonation is governed by the interaction of the nucleation and growth rate, both of which may be functions of the gas atmosphere. In carbon dioxide atmospheres above the sintering temperature, the growth of product crystallites is favored over nulceation of new crystallites (Bachmann and Cremer, 1961). The effect of high rates of reaction, in general, results in much smaller crystallite sizes, and is consistent with the fact that annealing processes usually lag behind interfacial reaction rates (Dedman and Owen, 1962; Hedin, 1961,1962). As a result, smaller calcium oxide crystallites are obtained for calcination in vacuo and for calcination under rapid heating conditions. The above sintering phenomena help to explain the observed reduction in recarbonation activity with increasing calcination temperatures from 800 to 1030 "C, as well as the similar reduction in activity obtained with soaking of the calcinate in carbon dioxide a t 925 "C. In our high-pressure cycling experiments sintering was confirmed by caking of the powdered samples to a sponge-like mass for all runs having calcinations a t 1030 "C (4 atm of CO2). No caking was observed in any runs having calcinations a t lower temperature. Gluud and Klempt (1930) considered 1050 "C as an absolute upper limit for the calcining temperature in their CO-shift process. Significant deactivation in the CO2-acceptor bench scale continuous unit is likely the result of the high calcination temperatures employed (1060 "C) (Curren et al., 1967). In the recarbonation of calcium oxide the factors which govern the initial reaction rate are: the initial crystallite size, the concentration of potential sites, the nucleation rate, the rate of interfacial reaction with carbon dioxide, and the porosity of the material (Galwey, 1967; Garner, 1955; Pannetier and Souchay, 1967; Young, 1966). Subsequent reaction will be governed by the nature of the product calcium carbonate layer. Unlike the calcination reaction, recarbonation requires the transport of carbon dioxide through the solid to the reaction interface. Due to the much larger specific molar volume of calcium carbonate (36.9 cm3/g-mol) relative to calcium oxide (16.9 cm3/g-mol) the oxide will become covered by a layer of carbonate of much lower porosity. As a result, the carbon dioxide may not have free access to the reaction interface, but has to diffuse through pores and cracks in the product layer to reach the interface. As the reaction proceeds, its rate will become progressively slower due to the increasing thickness of the product layer. In cases where the effective diffusivity of the product layer is low, the reaction may cease altogether. This phenomenon is referred to in the literature as retention (Young, 1966) or
Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
impedance (Garner, 1955) and is responsible for the low capacities of limes in reactions with carbon dioxide and sulfur dioxide (Kruel and Juntgen, 1967). In the recarbonation of lime the reaction ceases after about 70 to 80% conversion. Inspection of the resulting solid revealed an almost complete loss of porosity and surface area as a result of product layer impedance (Glasson, 1960; Ohno and Fujiyama, 1957). The magnitude of this effect is governed by the crystallinity of the product layer formed. Conditions which promote fissurization or recrystallization will lessen the degree of impedance (Garner, 1955). The ability of calcined dolomite (first cycle) to achieve complete recarbonation is related to the ultrafine dispersion of magnesium oxide crystallites in the solid which helps to maintain good porosity during the entire course of reaction (Richer, 1954). Deactivation of acceptors with repeated cycling can take two forms, one of which affects the reaction rate and another which determines the extent of final conversion. In the cyclic recarbonation of calcined dolomite we observe both effects. Loss of reactivity with repeated cycling is illustrated by the results shown in Figures 1 and 4, while loss of capacity is illustrated in Figures 2 and 3. The former effect is governed by the change in size and lattice strain of the calcium oxide crystals with cycling. Since calcination is carried to completion, the nature of the product layer should not change with cycling except as a result of changes in the size of the calcium oxide crystallites. Loss of capacity is related to two distinct phenomena. If recarbonation is not carried to completion there will be some calcium oxide crystallites which will anneal during calcination to the extent that they will remain inert toward recarbonation. These crystallites will simply sinter and grow with each cycle, thus increasing the proportion of calcium oxide unavailable for reaction. This process is similar to the high-temperature process (1200 "C) of "deadburning". In addition, as the average crystallite size of the calcium oxide increases with cycling due to sintering in the calcination step, the level of conversion will decrease for a given penetration depth of carbon dioxide through the carbonate layer. This will result in increasing retention and hence a loss of capacity with cycling. Kriek et al. (1959) have shown that the admixing of CaO to MgO prevents the sintering and deadburning of the magnesium oxide by reducing the probability for particle-to-particle contact. This suggests that fully calcined dolomite should be able to resist sintering and hence chemical deactivation better than calcium oxide or magnesium oxide individually. A key factor for the durability of dolomite is, therefore, the fine scale of segregation of the calcium and magnesium species in the solid. Curran et al. (1970) describe the change in crystallite sizes taking place during the cyclic recarbonation of dolomite. After calcination, both the calcium and magnesium oxide crystallites are about 400 A in size. After only a few cycles a t process conditions the calcium and magnesium oxide crystals have grown to greater than 2000 A, the calcium carbonate crystallites remaining at 400 A. Samples taken from their continuous unit showed that after several cycles the calcium and magnesium species had segregated completely, the crystallites having reached the enormous dimensions of 10 to 20 p (calcination was carried out a t 1050 "C). The deactivation of calcined dolomite in cyclic recarbonation may therefore be attributed to sintering in the calcination step. The growth of the calcium oxide crystallites reduces both the reactivity toward recarbonation as well as its ability to withstand loss of capacity due t o product layer impedance. In the limit, complete segregation of the calcium and magnesium species in dolomite will yield a solid having chemical properties no different from that produced from limestone. In line with the above picture for deactivation of calcined
IO0
l
l
l
CYCLE 4 [AFTER CALCINATION
l
~
'
cyc~E
80
PER CENT CONVERSION TO
CP co3
60 40 5 5 0 "C
1 ATM CO,
loo 200 300 TIME, mins. Figure 12. Activating effect of calcination in nitrogen on cyclic recarbonation of calcined dolomite at 550 "C at atmospheric pressure.
dolomite, a reduction in the size of calcium oxide crystallites should be able to reactivate the solid. A test of this hypothesis is shown in Figure 12. After three cycles of recarbonation a t 550 "C in which calcination was carried out in 1 atm carbon dioxide to 925 "C, a calcination in nitrogen to 800 "C was made. The subsequent recarbonation proceeds extremely fast, even relative to the first recarbonation. By calcining in nitrogen the average crystallite size is reduced, thus producing a solid of much higher reactivity. However, complete conversion was not attained due to the presence of "deadburnt" calcium oxide formed during the earlier cycles which cannot be reactivated by simple calcination in nitrogen. Conclusions (a) The recarbonation reaction is approximately first order in partial pressure of carbon dioxide in the range of 0.5 to 4 atm of COz. (b) The activation energy for the recarbonation reaction is estimated to be about 20 kcal/mol. (c) Addition of steam to the recarbonation atmosphere results in a two order of magnitude increase in recarbonation rate. The catalytic effect of steam has no "memory" and is proposed to catalyze the interfacial reaction rate directly. (d) Increasing calcination temperatures and hence partial pressures of carbon dioxide result in a less reactive material for recarbonation. Active calcined dolomite, prepared a t lower temperatures (800 "C) in nitrogen, will quickly lose reactivity if allowed to soak a t higher temperatures (925 "C) in carbon dioxide. Visible sintering of a powdered sample is obtained a t a calcination temperature of 1030 "C (in 4 atm of carbon dioxide). (e) Deactivation of calcined dolomite during cyclic calcination and recarbonation results in both a loss of reactivity and capacity. This has been attributed to growth and eventual segregation of the calcium and magnesium oxide crystallites during repeated calcinations. Reactivation of a portion of the solid was achieved by calcining a t a lower temperature (800 "C) in nitrogen, thereby reducing the size of the calcium oxide crystallites. Acknowledgment This work was funded by Research Grant No. AP-00945 from the Office of Air Program of the Environmental Protection Agency. We wish to acknowledge the contribution made by Gary Weil, who performed the nonisothermal recarbonation studies used in the report. Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
485
~
Literature Cited Asboth, K., Austrian Patent 172 931 (Oct 25, 1952). Bachmann, L., Cremer, E., 2.Anorg. Allg. Chem., 309, 65 (1961). Bischoff, F., Monatsh., 81, 606 (1950a). Bischoff, F.. 2.Anorg. Chem., 262, 288 (1950b). Bischoff, F., Claus, D., Lehmann, H., Tonind.-Ztg. Keram. Rundschau, 83, 293
(1959). Britton, H. T. S.,Gregg, S. J., Winsor, G. W., Trans. Faraday Soc., 48, 70
(1952). Clark, L. M., Rec. Trav. Chim., 68, 969 (1949). Cremer, E., Nitsch, W., Tonind.-Ztg. Keram. Rundschau, 83, 579 (1959). Cremer, E., Nitsch. W., Sci. Ceram., I , 295 (1962a). Cremer, E., Nitsch. W., 2.Nektrochem., 66, 697 (1962b). Curran, G. P., Fink, C. E., Gorin, E., Adv. Chem. Ser., No. 69, 141 (1967). Curran, G. P., Fink, C. E., Gorin, E., "Bench-Scale Research on CSG Process, Phase II. Operation Of The Bench-Scale Continuous Gasification Unit, Dec 1, 1965 To July 1, 1968.R and D Report No. 16,Interim Report No. 3,Book 3,"Government Printing Office, 1970. Curran, G. P., Rice, C. H.,Gorin, E., Am. Chem. Soc., Div. Fuel Chem., Prepr., 8 (l),128 (1 964). Davtyan, 0. K., Ovchinnikova, E. N., Soboleva, N. M., Nauch. Ezhegodnik, Odessk. Gosudarst. Univ., Khim. Fak., No. 2, 128 (1961). Dedman. A. J.. Owen. A. J.. Trans. FaradavSoc.. 58. 2027 11962). Dobner. S , Ph D Dissertation The City Cdllege of The City University of New York (Ch E ), 1976 Dobner, S.,Kan, G., Graff, R. A., Squires, A.M., Thermochim. Acta, 16, 251
(1976). Fischer, H. C., J. Am. Ceram. SOC., 38, 264 (1955). Galwey, A. K., "Chemistry of Solids", Chapter 5, Chapman and Hall, London,
1967. Garner, W. E., "Chemistry of the Solid State", Chapter 8, Academic Press, New York. N.Y., 1955. Glasson, D. R., J. Appl. Chem., 8, 793 (1958). Glasson, D. R., J. Appl. Chem., IO, 42 (1960). Glasson. D. R., J. Appl. Chem., 11, 201 (1961). Gluud, W., Keiler, K., Klempt, W., Bestehorn, R., Ber. Ges. Kohlentechnik, 3,
211 (1930). Hashimoto, H.,Kogyo Kagaku Zasshi, 64, 250 (1961). Haul, R. A. W., Schoning, F. R. L.. 2. Anorg. Chem.. 269, 120 (1952). Hedin, R., Svenska Forskningsinst. Cement Betong Vid Kgl. Tek. Hogskol. Stockholm Sartryck, 16, 661 (1961). Hedin, R., Tek, Tidskr., 92, 101 (1962). Hyatt, E. P., Cutler, I. B., Wadsworth, M. E., J. Am. Ceram. SOC., 41, 70
( 1970).
Kovalenko. E. N., Mat. Sci. Res., 3, 485 (1966). Kriek, H. J. S., Ford, W. F., White, J., Trans. Br. Ceram. SOC., 58, 1 (1959). Kruel, M., Juntgen. H., Chem. Ing. Tech., 39, 607 (1967). MacCallum, J. R., Tanner, J., Eur. Polym. J., 6, 1033 (1970). Maclntire, W. H., Stansel, T. B., Ind. Eng. Chem., 45, 1548 (1953). Nitsch, W., Z. Nektrochem., 66, 703 (1962). NOH, W., Ang. Chem., 62, 567 (1950). Ohno, Y., Sekko To Sekkai, 1, 1366 (1957). Ohno, Y., Fujiyama, S.,Sekko To Sekkai, 1, 1469 (1957). Pampuch, R., Silic. Ind., 23, 119 (1958). Pannetier, R., Souchay. P., "Chemical Kinetics", p 393,Elsevier, Amsterdam,
1967. Pell, M., Ph.D. Dissertation, The City College of The City University of New York (Ch.E.), 1971. Peterson, R. 0.. Cutler, I. E.,J. Am. Ceram. Soc., 51 (l),21 (1968). Proks, I., Jaskova, V., Silikaty, 11 (3),201 (1967). Proks. I., Siska, V.. Silikaty, 12 (l),13 (1968). Richer, A., Compt. Rend., 238, 339 (1954). Richer, A., Vallet, P., Compt. Rend., 252, 1780 (1961). Ruth, L., P h D Dissertation. The City College of The City University of New York (Ch.E.), 1972. Schwab, G. M., Taylor, J. S.,Spence. R., "Catalysis from the Standpoint of Chemical Kinetics", p 16,D. Van Nostrand, New York, N.Y., 1937. Schwob, Y., Rev. Mater. Constr. Trav. Publics, Ed. C. 411, 409 (1949). Schwob, Y., Rev. Mater. Constr. Trav. Publics, Ed. C, 413, 33 (1950a). Schwob, Y., Rev. Mater. Constr. Trav. Publics, Ed. C, 413, 85 (1950b). Sestak, J., Satava, V., Wendlandt, W. W.. Thermochim. Acta, 7 (5), 447
(1973). Shushinov, V. A., Fedyakova, K. G., Uchenye Zapiski Gor'kovst. Gosudarst. Univ. im. N.I. Lobachevskogo, Ser. Khim., No. 32, 13(1958). Siske, V., Proks, I., Chem. Zvesti, 12, 201 (1958a). Siske, V., Proks, I., Chem. Zvesti, 12, 275 (1958b). Tagawa, H., Sudo, F., Kogyo Kagaku Zasshi, 61, 949 (1959). Yanev, I. P., Angelov, B. D., Radenkova, M. Z., Dokl. Bolg. Akad. Nauk, 23 (lo),
1219 (1970). Young, D. A., "Decomposition of Solids", Chapter 3,Pergamon Press, New York. N.Y., 1966. Zawadski, J., Bretsnajder, S.,Z. Phys. Chem., 822, 60 (1933a). Zawadski, J., Bretsnajder, S.,Z. Phys. Chem., 822, 79 (1933b). Zawadski, J., Bretsnajder, S.,Z. Phys. Chem., 840, 158 (1935). Zawadski, J., Bretsnajder, S.,Trans. faraday SOC.,34, 951 (1938).
Received for revieul July 22, 1976 Accepted May 6,1977
(1958). Ketov, A. N., Pechkovskii, V. V., Larikov, V. V., Obsch. Prikl. Khim., No. 3, 48
Absorption of Nitrogen Dioxide in Sodium Sulfite Solution from Air as a Diluent Hiroshi Takeuchi,' Katsuroku Takahashi, and Nobuo Kizawa Department of Chemical Engineering, Nagoya University, Nagoya, 464, Japan
A study was made of the absorption of nitrogen dioxide with air as a diluent into aqueous solution of sodium SUIfite at 25 O C and atmospheric pressure. An agitated vessel with a plane interface was used to make sure of the effect of additive as an antioxidant of sulfite. For the sulfite solution with no additive a significant difference in the rate of NO2absorption was found between the nitrogen and air diluents. Examining some organic compounds as the additive, it was found that hydroquinone and monoethanolamine have a considerable effect on the inhibition in the sulfite oxidation. Furthermore, in order to discuss the applicability of the kinetics of the reaction between NO2 and SOS2-, which was proposed in a previous paper, the absorption of NO2 diluted with air was carried out in both a bubble cap- and a sieve-tray column. The overall capacity coefficients calculated from the equation based on the dual film theory and gas absorption accompanied by a pseudo mth order reaction were in reasonable agreement with the observations for the bubble cap tray.
Introduction In the previous paper (Takeuchi et al., 1977), the authors discussed the mechanism of NOz absorption in aqueous sodium sulfite and/or bisulfite solutions according to gas absorption accompanied by a fast, pseudo mth-order reaction, and proposed the mechanism with competitive reactions involving the hydrolysis of NO2 and the reaction between NO2 486
Ind. Eng. Chem., Process Des. Dev., Vol. 16, No. 4, 1977
and sulfite or bisulfite ion. Nevertheless, using nitrogen as a diluent it was not necessary to consider an oxidation of sulfite. From a viewpoint of the processes for removing NO, from stack gases by sulfite solution, however, it is considered that the sulfite oxidation becomes significant from the presence of 0 2 remaining in the gases. The reaction between dissolved oxygen and sulfite ion is quite complex and not well under-
