Energy & Fuels 1989,3, 24-28
24
the oxygen overpressure tests.
Conclusions Analytically, the advantages of this new method over previous methods are its greatly reduced test times and its excellent precision for replicate samples. Chemically, the advantages are its excellent correlations with lower temperature bottle testa, ita good discrimination of stability between fuels, and, most importantly, its ability to accu-
rately predict insolubles formation of various fuels during ambient storage. Systematic extension of the test to higher temperatures and shorter times will be the subject of a subsequent report.
Acknowledgment. We thank the US.Navy Energy Research and Development Office and David Taylor Research Center for financial support of this work. Registry No. Oxygen, 7782-44-7.
Single-Particle Surface Area Measurements in the Electrodynamic Balance D. R. Dudek, J. P. Longwell,* and A. F. Sarofim Chemical Engineering Department, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received March 1 , 1988. Revised Manuscript Received September 1 , 1988 ~~~
~
A technique is presented by which the microporous surface area of a single particle suspended in an electrodynamic balance can be measured in situ by using adsorption of carbon dioxide a t room temperature and the Dubinin-Polanyi equation. The agreement between surface area measurements performed on a single "Spherocarb" particle in the electrodynamic balance and on -150000 "Spherocarb" particles in a conventional volumetric adsorption system is very good when both are operated in the same range of carbon dioxide partial pressure. Eight separate single-particle surface area versus extent of reaction curves consistently show that the surface area of "Spherocarb" particles decrease from -950 to -650 m2/g on conversion from 0% to 80%. The surface area of a single Utah coal particle was also measured as a function of extent of reaction. The initial surface area of 150 m2/g increased steadily to -800 m2/g a t -75% conversion and then started to decrease.
-
Introduction The electrodynamic balance is a device capable of suspending a single 1-250-pm particle in an ambient gaseous environment and is now being developed as a tool to study single-particle, high-temperature, gas-solid reactions. Single-particle mass, diameter, density, and temperature measurements can all be performed as a function of the extent of particle reaction. Since the overall rate of reaction of the suspended particle is a function of ita surface area available to gaseous reactants, it is desirable to measure this surface area. Traditionally, methods for surface area measurement fall into three categories, (1)heat of wetting and calorimetric methods, (2) adsorption methods, and (3) smallangle X-ray scattering (SAXS). Heat of wetting methods lost their importance after it was discovered that the adsorption of methanol on coals involves specific interactions between the hydroxyl group of methanol and oxygen functionalities present on coal surfaces.' On the other hand, SAXS has not been used enough on coals and coal chars to allow for easy interpretation of the scattering data.2 Methods involving the adsorption of gases are used most commonly to measure surface areas of porous and nonporous solids. Nitrogen a t 77 K, carbon dioxide a t 195, (1) Marsh, H. The Determination of Surface Areas of Coals-Some Physichemical Considerations. Fuel 1966, 44, 263-268. (2) Kattiat, M.; Kwak, C. Y.; Schmidt, P. W. Presented at the Symposium on New Approaches in Coal Chemistry, Regional ACS Meeting, Pittsburgh, PA, 1980.
273, and 298 K, neon at 298 K, xenon a t 273 K, krypton a t 195 K, ethyl chloride a t 273 and 298 K, and other hydrocarbons at 298 K have all been used to measure the surface area of coal and coal char. Nitrogen and carbon dioxide are the most commonly used gases. Because of the low temperature (77 K) used with N2 adsorption, activated diffusion can be important in microporous networks, which then leads to experimentally unattainable equilibrium coverage of the N2 molecules on the microporous surface and, therefore, in some cases, unrealistically low surface areas.3 Anderson et al.4 have suggested that N2 adsorption at 77 K provides a measure of the surface area contained in pores having diameters greater than about 5 A. Also, Zwietering and van Kreveled found that the adsorption of N2 at 77 K in a particular coal had a half-coverage time of -1000 years. Marsh and Wynne-Jones6 have reported the opposite problem with N2 adsorption in activated carbons. When some pores in the carbon are several N2 diameters wide, N2is capable of filling these pores at extremely low relative pressures. This capillary condensation can cause the BET equation to predict an unrealistically high value of surface area. (3) Meyers, R. A,, Ed. Coal Structure; Academic Press: New York, 1982; pp 60-76. (4! Anderson, R. B.; Hall,W. K.; Leckye, J. A.; Stain, K. C. Sorption Studies on American Coals. J. Phys. Chem. 1956, 60, 1548-1668. (5) Zwietering, P.;van Krevelen, D. W. Chemical Structure and Properties of Coal IV-Pore Structure. Fuel 1954,33, 331-337. (6) Marsh, H.;Wynne-Jones, W. F. K. The Surface Properties of Carbon-I. The Effect of Activated Diffusion in the Determination of Surface Area. Carbon 1964, 1 , 269.
0887-062418912503-0024$01.50/0 0 1989 American Chemical Society
Single-Particle Surface Area Measurements
4
Energy & Fuels, Vol. 3, No. I, 1989 25
TO
MICROSCOPE
DETECTTOFI
DETECTORS
fF\
LASER co2
H e Ne
(N,, CO,, air) LASER Figure 1. Exploded view of the electrodynamic balance. GAS
Adsorption of COPat higher temperatures (195,273, 298 K) reduces the diffusion problems of N2 at 77 K for two reasons. First, C 0 2 is a smaller molecule than N2 with a minimum dimension of 3.3 A compared to 3.64 A? Second, diffusion a t 195, 273, and 298 K is much greater than diffusion a t 77 K. Thomas and Damberger' calculated the diffusion time of an N2 molecule a t 77 K through a pore 5 A in diameter and 10 pm in length to be lo5 times the diffusion time of a C 0 2 molecule at 195 K through the same pore. It should be noted, however, that capillary condensation may still be a problem when C 0 2 adsorption is used. When the BET equation for calculating surface area is used, it is essential to measure sorption isotherms up to a relative vapor pressure of about 0.2.8 For C 0 2 a t 273 and 298 K, the saturation vapor pressure is much higher than atmospheric pressure; therefore, to measure surface areas with the BET equation by using COz, a high-pressure adsorption system must be used. Marsh and SiemieniewskaQovercame this difficulty by using the DubininPolanyi (D-P) equation instead of the BET equation to calculate surface areas of coals by using C 0 2 adsorption a t 273 and 293 K. Walker and Patello measured surface areas of 10 different coals and chars with C 0 2 adsorption at 298 K by using both the D-P and BET equations in two different pressure ranges and found excellent agreement. Therefore, it can be concluded that a conventional volumetric adsorption apparatus and the D-P equation are as adequate for measuring C 0 2 surface areas of coals as a high-pressureadsorption apparatus and the BET equation. In this work, single-particle surface areas of a coal and a synthetic char versus extent of reaction are evaluated from the adsorption isotherm of C 0 2 at 298 K by using the Dubinin-Polanyi equation. Experimental Equipment The electrodynamic balance is sketched in Figure 1. Basically, it consists of two endcap electrodes and a ring electrode. A dc potential is applied across the endcap electrodes and an ac voltage is applied to the ring electrode. The electric field that is generated between the electrodes can suspend and hold a charged particle. Two optical microscopes and a camera are used for particle size measurements. One (70X) microscope is equipped with a graticule and can measure particle diameters to within *4 pm. Photographs (7) Thomas, T., Jr.; Damberger, H. H. Circ.-Ill. State Geol. Surv. 1976,No.493. (8)Brunauer, S.;Emmett, P. H.; Teller, E. Adsorption of Gases in 1938,60, 309-319. Multimolecular Layers. J. Am. Chem. SOC. (9)Marsh, H.;Siemieniewska, T. The Surface Areas of Coals as Evaluated from the Adsorption Isotherms of Carbon Dioxide Using the Dubinin-Polanyi Equation. Fuel 1966,44,355-367. (10)Walker, P.L., Jr.; Patel, R. L. Surface Areas of Coals from Carbon Dioxide Adsorption at 298 K. Fuel 1970,49,91-94.
taken from the 35-mm camera attached to the (loox)microscope can be used to measure diameters to within i 3 pm. The particle can be weighed by using a calibrated aerodynamic drag force technique.l1-l3 The error of this weighing procedure is approximately 10% for spherical particles. Correction of the drag coefficient for other shapes will introduce an additional source of error. An analysis of these sources of error is presented in ref 13. Also, a 20-W C02laser is available for particle heating, an electrooptical feedback control system automatically adjusts the dc voltage required for particle balancing, and a two-color (2 and 4 pm) infrared pyrometer provides the capability of measuring particle temperature. Details of this device are described elsewhere.14 Chromatographic-grade nitrogen, carbon dioxide, and various mixtures of nitrogen and carbon dioxide are supplied to the chamber via cylinders and their flow rate is monitored by a Brooks mass flowmeter (0-60 mL/min). Gas enters the chamber through a 5.56-mm hole in the center of the bottom dc electrode, passes around the suspended particle, and then exits to the atmosphere through a 5.56-mm hole in the center of the top electrode. Most of the surface area measurements presented in this paper were performed on a synthetic char, "Spherocarb". "Spherocarb" was developed as a chromatograph packing but is also finding applications as a model compound for the study of carbon oxidation'"l and gasification.18 It is produced by Foxboro/Analabs and the particles were found to be very nearly spherical. Procedure The use of the Dubinin-Polanyi equation is discussed by Lamond and Marshlg and is written as follows: BT2 log v = log v,- -log2 (P,/P) (1) P where V is the amount of COP adsorbed (mol/g) at the COz equilibrium pressure P, V, is the micropore capacity (mol/g), Po is the saturation vapor pressure of C02at temperature T (K), P is the affinity coefficient of COzrelative to Nz,and B is a constant. A plot of log V versus log2(P,/P) should yield a straight line with the intercept at log2(P,/P) = 0 equal to log V@ If the adsorption of COz is restricted to a monolayer, then Vomultiplied by the cross-sectional area of a C02molecule and Avagadro's number yields the microporous surface area. The cross-sectionalarea of the COzmolecule is a function of temperature and was taken from Walker and Kini.20 A vertical force balance on a particle perfectly centered in the chamber reveals that ma = qE (2) (11)Davis, E. J.; Zhang, S. H.; Fulton, J. H.; Periasamy, R. Measurement of the Aerodynamic Drag Force on Single Aerosol Particles. Aerosol Sci. Technol. 1987,6, 273-278. (12)D'Amore, M.; Dudek, D. R.; Sarofim, A. F.; Longwell, J. P. Apparent Particle Density of a Fine Particle. Powder Technol. 1988,56, 129-134. (13)Dudek, D. R. Single Particle, High Temperature, Gas-Solid Reactions in an Electrodynamic Balance. Ph.D. Thesis, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA, 1988. (14)Spjut, R. E.; Bar-Ziv, E.; Sarofim, A. F.; Longwell, J. P. Electrodynamic Thermoaravimetric Analyzer. Rev. Sci. Instrum. 1986, 57, 1604-1610. (15)Floess, J. K. The Effect of Calcium on the Gasification Reactions of Carbon. Ph.D. Thesis, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA, 1985. (16)Helble, J.; Neville, N.; Sarofim, A. F. Aggregate Formation from Vaporized Ash during Pulverized Coal Combustion. In Twenty-First Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1986. (17)Niksa, S.;Mitchell, R. E.; Hencken, K. R.; Tichenor, D. A. Optically Determined Temperatures, Sizes, and Velocities of Individual Carbon Particles under Typical Combustion Conditions. Combust. Flame 1984,60,183-193. (18)Mims, C. A.; Pabst, J. K. Role of Surface Salt Complexes in Alkali-Catalyzed Carbon Gasification. Fuel 1983,62,176. (19)Lamond, T. G.;Marsh, H. The Surface Properties of Carbon-11. The Effect of Capillary Condensation at Low Relative Pressures Upon the Determination of Surface Area. Carbon 1964,1,281-292. (20)Walker, P. L., Jr.; Kini, K. A. Measurement of the Ultrafine Surface Area of Coals. Fuel 1965,44, 453-459. ~
Dudek et al.
26 Energy & Fuels, Vol. 3, No. 1, 1989
i
0 “Spherocarb” A glass sphere
220 0.002
0
z
h
\
c
210
0
v
0.001
0
0
200
200
0
MASS
400
FLOW
I
0 000
600
0
2
.A
1
A
4
A
I A
8
I
0
~o~z~prJ/p)
METER READING
Figure 2. “Spherocarb”balancing voltage versus mass flowmeter
Figure 3. Plot of moles of CO, adsorbed/particle weight versus log2 (Po/P)for a “Spherocarb” particle and a glass sphere.
where m is the mass of the suspended particle, g is the gravitational acceleration, q is the number of excess charges on the particle, and E is the electric field strength in the vertical direction. For the chamber configuration utilized in this work E = cvdc/zO (3) where C is a chamber constant, v d c is the dc voltage between the endcap electrodes, and zo is the characteristic length of the chamber. Substituting eq 3 into eq 2 gives
Gas flow rates corresponding to mass flowmeter readings of 400,
reading for five different C 0 2 / N 2 gas mixtures.
mg = qcvdc/zO
(4)
Therefore, for a particle of a given charge ( m- m i ) / m i
(vdc
- vdci) / vdci
(5)
where mi and vdci are the initial mass and voltage, respectively, when no CO, has been adsorbed. In other words, the change in particle mass due to COz adsorption divided by initial mass is equal to the change in dc voltage required for balancing divided by the initial voltage. Experiments were initiated by capturing one particle in the center of the chamber. Degassing was then accomplished by passing ultrahigh-purity nitrogen through the chamber and over the particle at 42 mL/min for 30 min. While conventional BET procedures evacuate the sample to remove adsorbed gas, it was found that, for our conditions,the exchange of nitrogen with COP was sufficiently rapid. All the results in this paper will be presented as a Dubinin-Polanyi plot which is in the form of log ( n / W )vs log2 (Po/P),where n is the number of moles of CO, adsorbed and W is the COz-freeweight of the suspended particle. The n / W ratio can be measured directly from the balancing voltages (eq 5) and by dividing by the molecular weight of CO,. The temperature of the chamber and the gas mixture inside the chamber was taken to be in equilibrium with the surrounding atmosphere and was measured with a thermometer to hO.1 “C. This temperature was then used along with data from Perry1 to determine Po,the saturation vapor pressure of COB Five different gas mixtures were utilized in these adsorption studies, pure Nz, a 20%/80% C 0 2 / N 2 mixture, a 35%/65% C 0 2 / N 2 mixture, a 50%/50% C 0 2 / N 2 mixture, and pure CO,. Some raw data for a “Spherocarb” particle are plotted in Figure 2 in the form of balancing voltage versus the mass flowmeter reading. The mass flowmeter reading was used instead of volumetric flow rate because the mass flowmeter is only calibrated for nitrogen. A mass flowmeter reading of 600 corresponds to a nitrogen flow rate of 42 mL/min. Three balancing voltage measurements (at mass flowmeter readings of 400,500, and 600) were taken for each gas mixture. These measurements are shown as the open boxes in Figure 2. Since the chamber is not leak proof, in order to insure that the suspended particle is experiencing the same environment that is in the gas cylinders, a finite gas flow rate must be maintained. (21) Perry, R. H.; Chilton, C. H., Eds. Chemical Engineers’ Handbook, 5th ed.; McCraw-Hill: New York, 1973.
500, and 600 have been shown to be experimentally adequate.
Unfortunately, the introduction of a gas flowing upward past the suspended particle produces an aerodynamic drag force, Fd, on the particle, which affects the dc voltage required for stable balancing of the particle. This aerodynamic drag force can be described by using Stokes’ law Fd
= 31rrfiV
(6)
where r is the particle radius, fi is the viscosity of the gas mixture, and IJ is the velocity of the gas mixture over the particle. From Figure 2, we can see that the higher the gas flow rate, the less voltage required for balancing. This is because the aerodynamic drag force is pushing the particle upward, helping to suspend the particle. The solid lines in Figure 2 represent a least-squares linear regression that was performed on each group of three experimental points. By extrapolation of the experimental data to a zero mass flowmeter reading, the “true” balancing voltage without the aerodynamic drag force included can be obtained. These “true” balancing voltages can then be used directly to form a DubininPolanyi plot and thus a surface area calculation. One surface area measurement performed in this fashion takes approximately 1.5 h. Equation 5 predicts that the increase in balancing voltage with increasing COz partial pressure is due to particle mass increases caused by increased COz adsorption.
Results and Discussion In order to insure that the changes in balancing voltage shown in Figure 2 are really due to C 0 2 adsorption and not some hydrodynamic effect, the data from Figure 2 are plotted in (n/W) vs log2 (Po/P)form in Figure 3 along with similar data from a glass sphere. As expected, the glass sphere, which is known to have a very low surface area, absorbs practically no CO,; therefore, the adsorption of COz in the “Spherocarb” particle is assumed to be real. A comparison of the Dubinin-Polanyi plot between data collected in a conventional volumetric adsorption apparatus and data collected on a single particle in the electrodynamic balance is made in Figure 4.22 The conventional apparatus, which utilized 0.294 g of “Spherocarb” particles and operated at 273 K, could be operated over a much broader range of CO, pressures. Although not readily apparent, the points from the conventional apparatus do not form a straight line, but are curved concave upward; therefore, the intercept, and hence the surface area, will depend on which points are chosen. The in(22) Hurt, R. H. Personal communication, Conventional volumetric adsorption apparatus data, Department of Chemical Engineering, Massachusetts Institute of Technology, 1987.
Single-Particle Surface Area Measurements
Energy & Fuels, Vol. 3, No. 1, 1989 27 -2.0
+ *Ingle ‘Sphsroearb’
A
-2
Particle in CDB
duplicate ‘Spherofarb. paNcle
In EDB
0 284 g of ‘Spherocarb’ psrtlclsr
-2.5
h
;
M
E -3
\ C
\ C V
-
a 0
0
V
M
0 4
-4
x
+
-3.0
0 a
14% conversion
: 0
+ 23% conversion 0 52X conversion
-5 1 0
O I
10
20
1
I 30
l0g2(Po/P) Figure 4. Comparison of log (n/W) versus log2(Po/P)plots for runs performed in the EDTGA and a conventional adsorption apparatus.
tercept of the Dubinin-Polanyi plot, which is determihed from a least-squares linear regression, is used to determine surface area. Using only high values of COP pressure or low values of log2 (Po/P)will yield higher calculated values of surface area than using only low values of C 0 2pressure or high values of log2 (Po/P).It is not surprising, therefore, that the conventional apparatus (weighted more heavily toward the low values of C02pressure) yields a surface area for “Spherocarb” particles of 636 m2/g compared to values ranging from 821 to 1175 m2/g for single “Spherocarb” particles measured in the electrodynamic balance. The points from the electrodynamic balance shown in Figure 4 yield a surface area of 1160 m2/g. A surface area of 965 m2/g is calculated from the conventional apparatus data if only the two points with the highest values of C 0 2 pressure are used. It is estimated that the error produced by extrapolation of log2 (Po/P)combined with the error of extrapolation to zero gas flow rate is less than lo%, so when this error is combined with the 10% error in particle weight determination, an uncertainty of approximately 15% is introduced. I t is believed that refinement of this technique could reduce this error. It can be argued, therefore, that the absolute value of the measured surface area depends on the values of the C 0 2 pressure used in the Dubinin-Polanyi equation. Reasonable surface areas are obtained by use of any of the C 0 2 pressures reported in Figure 4; however, if valid comparisons are to be made from one particle to the next or from the same particle at different extents of conversion (fractional weight loss) or from the same particle using different apparatuses, care must be taken in choosing the same range of C 0 2 pressures. Figure 5 is a Dubinin-Polanyi plot for one “Spherocarb” particle at three separate extents of conversion, 14%, 23%, and 52 % . This particular reaction was performed in air a t 780 K. The reaction was stopped a t various extents of reaction, and surface area measurements were performed. The slope of these three lines is fairly constant, -0.162, -0.157 and -0.153, but the intercept decreases with conversion, indicating a decreasing surface area. Figure 6 is a plot of surface area per gram of “Spherocarb”remaining versus percent conversion for eight separate single-particle “Spherocarb” reactions. Five of the runs were performed in oxygen, two were performed in carbon dioxide, and one was performed in air. The reaction time required for 50% conversion ranged from 2 to 16 min for the oxygen runs and from 20 to 30 min for the carbon dioxide runs and was 110 min for the air run.
-3.5 2
0
4
6
10g2( P O P )
Figure 5. Plot of log ( n / W ) versus log2 (Po/P)for the same “Spherocarb” particle at three different conversions.
II 1000
?
+ + e . g
“E
v
0
o+ 0
0
0
w w d
500
rz;
0:
. 1 +x
?i
0
000
A p
single Spheroearb” in 0, s m i l e ’Spherocarb” tn CO, s m ~ l c Spheroearb” tn air
0
0
0
150 000 ‘Spherocarbs“ ~n conventional appamtus
100
50
PERCENT CONVERSION
Figure 6. Surface area versus percent conversion for “Spheracarb”particles. The temperatures corresponding to these reaction times are 750-830 K for oxygen, 1200-1250 K for carbon dioxide, and 770 K for air. The initial specific surface area varied from 821 to 1116 m2/g, with the average initial surface area being 960 m2/g. These surface areas can be compared to a value of 965 m2/g obtained from a 0.294 g sample of 150 000 “Spherocarb” particles by Hurt22in a conventional volumetric adsorption apparatus. The 0.294 g of “Spherocarb” particles were reacted in a conventional TGA apparatus. The surface area appears to increase slightly up to a conversion of 10-15% and then monotonically decrease to an average value of 660 m2/g a t about 81% conversion. A value of 647 m2/g was obtained at 65% conversion by Hurt22using a conventional volumetric adsorption apparatus. The data indicate that “Spherocarb” surface area evolution is not a function of temperature or reacting gas but only a function of conversion. Single-particle surface areas have also been determined for real coal particles. Figure 7 is a plot of surface area per gram of coal remaining versus extent of reaction for a Utah No. 1 subbituminous coal reacting in air. The initial surface area of 152 m2/g increases steadily with conversion until peaking at 800 m2/g somewhere between 70% and 80% conversion and then decreases.
-
Conclusions Single-particle C 0 2 surface area measurements performed in the electrodynamic balance are in agreement with the measurements performed on 150000 particles in a conventional volumetric adsorption apparatus when
-
Energy & Fuels 1989, 3, 28-37
28 loo0
&
h
0
M
0
\
“E
List of Symbols
0
v
0
0
500
-
0 0
w u 4 Lr, !x
0
2
0 0
50
Acknowledgment. Financial support for this work by Exxon Research and Engineering Company is gratefully acknowledged.
100
PERCENT CONVERSION
Figure 7. Surface area versus percent conversion for a single particle of Utah No. 1 subbituminous coal. both apparatuses are operated in the same range of COz partial pressure. Surface area versus extent of reaction curves are now possible for single particles reacting in the electrodynamic balance. As an example of the technique, the surface area of “Spherocarbn particles was shown to decrease from -960 to -660 m2/g when reacted from 0% to 80%.
B = constant C = chamber constant (dimensionless) E = electric field strength in the vertical direction (Vim) Fd = aerodynamic drag force on the suspended particle (N) g = gravitational acceleration (m/s2) m = mass of particle (kg) n = number of mols of C02 adsorbed (mol) P = COz equilibrium pressure (atm) Po = saturation vapor pressure of COz (atm) q = number of excess charges on the particle (C) r = particle radius (m) v = velocity of the gas mixture over the particle (m/s) V = amount of COz adsorbed (molfg) Vo = micropore capacity (molfg) VDc = dc voltage between the endcap electrodes (V) W = C02-free weight of suspended particle (N) zo = characteristic length of the chamber (m) ,f3 = affinity coefficient of C02 relative to N2 p = viscosity of gas mixture (kg/(m s)) Subscripts i = initial or C02-free Registry No. COz, 124-38-9; Spherocarb, 67417-49-6.
Catalysis of the Combustion of Synthetic Char Particles by Various Forms of Calcium Additives Y. A. Levendis,t S. W. Nam,$ M. Lowenberg,$ R. C. Flagan,t and G. R. Gavalas*p* Departments of Environmental Engineering Science and Chemical Engineering, California Institute of Technology, Pasadena, California 91125 Received June 20, 1988. Revised Manuscript Received September 30, 1988
The reactivity of calcium-laden carbonaceous particles to oxygen was investigated in the range 670-3000 K. The char particles were prepared from poly(furfural alcohol) (PFA) and were spherical and of uniform size. Three different methods were used to introduce the calcium additive: precipitation of calcium carbonate, impregnation with calcium acetate, and calcium ion exchange. Electron microscopy showed that the distribution of calcium was remarkably uniform in particles containing a bimodal distribution of micropores and transitional pores, whereas for particles with micropores only the Ca concentration was high at the surface and low a t the center. X-ray analysis indicated that the conversion of the carbonate to the oxide at low temperatures (below 900 K) takes place only after all carbon has been consumed. Combustion studies showed that the calcium catalyzed the oxidation reaction at all temperatures investigated by up to 2 orders of magnitude. The effectiveness of the catalyst introduced by the different methods was comparable, with the calcium ion exchanged chars being, in general, the most reactive.
1. Introduction An extensive literature has been published dealing with
the catalysis of carbon and coal gasification by various metals and metal compounds, especially by alkali and alkali-earth metals.’-l* Calcium, in particular, has been the subject of numerous investigations as a naturally occurring mineral and as a low-cost additive that can catalyze Department of Environmental Engineering Science. of Chemical Engineering.
t Department
0887-0624/89/2503-0028$01.50/0
the rate of and serve as a sulfur scavenger in gasification and combu~tion.’~A number of techniques (1) Jenkins, R. G.; Nandi, S. P.; Walker, P. L., Jr. Fuel 1973, 52, 288-303. (2) Hippo, E. J.; Walker, P. L., Jr. Fuel 1975,54, 245-248. (3) Tomita, A.; Mahajan, 0. P.; Walker, P. L., Jr Fuel 1977, 56, 137-144. (4) Linares-Solano, A.; Mahajan, 0. P.; Walker, P. L., Jr. Fuel 1979, 58, 321-332. ( 5 ) Takarada, T.; Tamai, Y.; Tomita, A. Fuel 1985, 64, 1438-1442. (6) Hengel, T. D.; Walker, P. L., Jr. Fuel 1984, 63, 1214-1220.
0 1989 American Chemical Society