Energy & Fuels 1994,8, 213-218
213
Pyridine Sorption Isotherms of Argonne Premium Coals: Dual-Mode Sorption and Coal Microporosity Thomas K. Green* and Trent D. Selby Department of Chemistry, Western Kentucky University, Bowling Green, Kentucky 42101 Received May 3, 1993. Revised Manuscript Received September 21, 1 9 9 P
Pyridine sorption isotherms were determined for three Argonne Premium coals at 50 “C. It is proposed that the isotherms are best interpreted in terms of a dual-mode sorption model originally developed to explain sorption of vapors into glassy polymers. This model considers two different populations of sorbed gas; one is dissolved into the polymer and is described by Henry’s law (i.e., sorption is linear with pressure) while the second population is considered to occupy unrelaxed free volume (micropores) within the polymer and is described by a Langmuir isotherm. We propose that the high-pressure linear portion of the pyridine isotherms for two coals represent dissolution of pyridine in the coal matrix and that the low-pressure regions represent primarily adsorption into micropores. The intercepts of the linear portion thus represent micropore volume accessible to pyridine. The pyridine-extracted coals all show a dramatic increase in intercept value, suggesting that pyridine extraction creates new micropores. 0-alkylation with bulky alkyl groups yield reduced intercepts, which correlate with the size of the added alkyl group. It is suggested that the added alkyl groups occupy micropores in the structure.
Introduction
Pyridine is one of the most effective and commonlyemployed solvents for investigating coal structure. We have recently determined sorption isotherms for several coal-pyridine systems.l The most striking characteristic of these isotherms are their high linearity over a very wide pressure range. The curves also have nonzero intercepts. We propose that these isotherms can be modeled by a dual-mode sorption mechanism which has been widely used to interpret the sorption isotherms of glassy polymers.24 In this model, the sorption mechanism is described in terms of one population of ordinarily dissolved sorbate which resides in the coal matrix and is described by Henry’s law (Le., sorption is linear with pressure) while the second population of sorbate is considered to occupy unrelaxed free volume (micropores) within the coal matrix and is described by a Langmuir isotherm. We propose that the linear portion of the pyridine isotherms presented in this paper represent dissolution of pyridine and that the intercepts are a measure of the coal’s microporosity available to pyridine. We present results on three Argonne premium coals, including the whole coals, pyridine extracts, extraction residues, and some 0-alkylated coals,to support our hypothesis. Porosity of Coals. The pores in porous solids are generally classified into three regime^:^ macropores with diameters greater than 500 A, mesopores with diameters e Abstract published in Aduance
ACS Abstracts, November 1, 1993. (1) Green, T. K. “Thermodynamics of the Solvent Swelling of Coal”, Final Technical Report, DOE/PC/88924, US. Department of Energy, Pittsburgh Energy Technology Center, 1991. (2) For a review of this model see: Vieth, W. R. Diffusion in and Through Polymers, Hanser Publishers: Munich, 1991; Chapters 2 and I. (3) Meares, P. J . Am. Chem. SOC.1954, 76,3415. (4) Barres, R. M.; Barrie, J. A.; Slater, J. J. Polym. Sci. 1958,27, 177. (5) Michaels, A. S.;Vieth, W. U.; Barrie, J. A. J . Appl. Phys. 1963,34, 1-12. (6) Vieth, W.R.; Howell, J. M.; Hsieh, J. H. J.Membr. Sci. 1976, I , 177. ~~
0887-0624/94/2508-0213$04.50/0
from 20 to 500 A, and micropores with diameters less than 20 A. This classification is not arbitrary and is based on the characteristics adsorption effects as manifested in the sorption isotherm. In the macropore range, the pores are so wide that it is virtually impossible to map out the isotherm in detail because the relative pressures of the adsorbate (adsorbed gas) are so close to unity. In mesopores, capillary condensation takes place, usually a t moderate relative pressures (greater than 0.3 relative pressure). In micropores, which are of molecular dimensions, an enhancement of the interaction potential takes place between the sorbent and adsorbate. The upper limit of size at which a pore begins to function as a micropore depends on the diameter of the adsorbate molecule u; for slitlike pores this limit is about 1.5 u but for cylindrical pores it lies at a pore diameter of about 2.5 u. As a consequence of this enhanced interaction potential, the micropore will be completely filled a t low relative pressures, frequently less than 0.01 relative pressure. This paper is concerned with the micropores of coal and this low pressure region where micropores are being filled. The pore structure of coals is well studied and reviews are a ~ a i l a b l e . The ~ , ~ most common methods for measuring coal porosities are mercury porosimetry and nitrogen adsorption. Van Krevelen and Zwietering,’ousing mercury porosimetry, established that coal contains both a macropore and micropore system. It is the micropore system which contributes most to the surface areas of most coals. For example, Lin et al.,” using small-angle X-ray scattering, found for the Illinois No. 6 coal that micropores contributed 140 m2/g surface area and that mesopores contributed only 10 m2/g. Microporosity therefore is of (7) Gregg, S.J.; Sing, K. S.W. Adsorption, Surface Area and Porosity; Academic Press: London, 1982. (8) Mahajan, 0. P. in Coal Structure; Meyers, R. A., Ed., Academic Press: New York; pp 51-87. (9) Marsh, H. Carbon 1987,25, 49-58. (10)Zweitering, P.; van Krevelen, D. W. Fuel 1954, 33, 331. (11) Lin, J. S.; Hendricks; R. W.; Harris, L. A.; Yust, C. S. J. Appl. Crystallogr. 1978, 11, 621.
0 1994 American Chemical Society
Green and Selby
214 Energy & Fuels, Vol. 8, No. 1, 1994
significant interest to coal scientists since any reagent or solvent must pass through the surface associated with micropores in order to reach the bulk of the coal. Gan, Nandi, and Walker1* studied the pore size distribution of a large number of American coals using both nitrogen adsorption and mercury and helium displacement. They found that the micropore volume of the coals (less than 12 A) range from 12 to 88% of the total pore volume of the coals, indicating that coals can act as molecular sieves. Of particular relevance to this paper is the recent suggestion of Larsen and Wernettl3 that many of the pores of the Illinois No. 6 coal are closed. These pores cannot be reached by diffusion through the pore network but must diffuse through the solid coal matrix. Carbon dioxide is an example of a soluble probe and surface areas determined by this probe are thought to be accurate. Pyridine is also capable of diffusing rapidly through coals but pyridine also swells coals to a large extent and this alters the pore structure. For example, small-angle neutron scattering experiments by Winans and Thiyagarajan14have demonstrated that a Pittsburgh bituminous coal contained a broad distribution of spherical pores but the coal swollen with pyridine contained elongated pores with a definite cross sectional radius. Kispert and coworkers'5 have developed a method to determine the pore size distribution in bituminous coals using nitroxide spin probes of different size and shape. The spin probes are dissolved in a swelling solvent and the solution is used to swell the coal. Removal of the solvent traps the spin probe, which is then detected by EPR. Their results on the Pittsburgh coal using pyridine as the swelling solvent generally agree with those of Winans and Thiyagarajan14 in that the number of spherical pores decreased but the number of elongated pores increased. Two groups of workers have used xenon gas adsorption and Xe NMR to investigate the microporosity of Botto and TsiaolGmeasured Xe adsorption isotherms on oxidized Illinois No. 6 coal and found that Xe adsorption decreased with degree of oxidation. They interpret this as a collapse of the micropores upon oxidation. Larsen et al. calculated the average micropore diameter of an Illinois No. 6 coal from the NMR chemical shift of adsorbed Xe as a function of Xe pressure.17 They calculated the average micropore diameter of the coal to be 5.4 A. Experimental Section
Sample Preparation. ArgonnePremium coals were obtained in ampoules of 5 g of -100 mesh. The coals were first dried under vacuum at 105O C to constant weight and then analyzed for carbon, hydrogen, and nitrogen. The results are shown in Table 1. Approximately 4.5 g of the sample was Soxhlet-extracted with dry pyridine under argon for several days until the siphon liquid was clear. The pyridine solution was then filtered through a 0.4-pm nylon membrane filter to ensure removal of particulates and colloidal material. The filter did not plug. Most of the pyridinewm removedby rotovaporizationunder reduced pressure at 70-80 OC. Approximately 200 mL of a methanol/water (80/20 vol) mixture and 2 mL of concentrated HC1 were added to the (12) Gan, H.; Nandi, S. P.; Walker, P. L. Fuel 1972,51, 272.
(13) Larsen, J. W.; Wernett, P. Energy Fueh 1988, 2, 719-720. (14) Winans, R. E.; Thiyagarajan, P.Energy Fuels 1988,2, 356-358. (15)Corray, L. S.; Kispert, L. D.; Wuu, S. K. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. 1988, 33, 32-37. (16) Tsiao, C.; Botto, R. E. Energy Fuels 1991,5, 87-92. (17) Larsen, J. W.; Wernett, P. C.; Yamada, 0.; Yue, H. J. Energy Fuels 1990, 4,412-413.
IO
".CYU.
Figure 1. Sorption apparatus. Table 1. Elemental Analyses of Aruonne Coals (Dry Basis) %C
%H
%N
H/C
65.7 77.9 60.5 64.6 64.3 69.1
4.7 5.6 4.4 4.8 5.7 6.6
1.2 1.5 1.3 1.2 0.8 0.7
0.85 0.87 0.87 0.89 1.06 1.15
whole extract
74.2 81.4 69.7
5.1 5.7 4.4
1.4 1.6 1.5
0.82 0.83 0.75
whole extract residue
67.5 72.9 68.2
4.9 5.3 4.8
0.8 0.8 1.2
0.86 0.87 0.84
coal
Illinois No. 6 whole extract
residue 0-methylated 0-butylated 0-octylated
Pittsburgh No. 8 residue Wyodak
flask and the mixture was stirred under nitrogen for 2 days. This treatment was used to remove residual pyridine as suggested by Buchanan.l* The solid extract was then filtered and dried under vacuum at 105 O C for 24 h. The extractability was 27.2 wt % for Illinois No. 6,29.8% for Pittsburgh No. 8, and 5.2 % for WyodakAnderson coal. The extracts were analyzed for carbon, hydrogen, and nitrogen. Most of the pyridine was removed from the extraction residues under vacuum. The residues were then treated with HCl/ methanol/water and dried in a similar manner as the extracts. Nitrogen analysis, shown in Table 1, reveal that most of the pyridine was removed from the pyridine extracts and residues. No evidence of residual pyridine was observed in the IR spectra of these materials. 0-alkylation Procedure. The procedure for 0-alkylation of the Illinois No. 6 coal was conducted according to Liotta's method.'$ Sorption Experiments. Sorption experiments on the coals were carried out using a quartz spring balance shown in Figure 1. The balance consists of a quartz spring, a 5-L flask, vacuum inlet system, and MKS pressure transducer (0-1000 Torr, 0.5% accuracy). The entire balance, including transducer, is housed in a Precision Scientificcirculating (forcedair) drying oven. The temperature is controlled by a PR temperature regulator, which activates a light bulb. The sample is suspended from the quartz spring and, as the sample sorbs solvent, the spring extends until equilibrium is reached. The extension of the spring is measured by an Eberbach cathetometer (travelling telescope). The spring is calibrated at the appropriate temperature using standardweights. The balance thus allows determination of the mass of solvent sorbed by the sample at agivenpartial pressure and temperature. The purpose of the 5-L flask is to minimize pressure changes caused by sorption of solvent by the sample. Quartz springs of the type used here have a linear extension versus suspended weight relationship (18)Buchanan, D. H.; Warfel, L. C.; Mai, W.; Lucas, D. Prepr. Pap-Am. Chem. Soc., Diu. Fuel Chem. 1987, 32 (l), 146. (19) Liotta, R.; Rose, K.; Hippo, E. J. J. Org. Chem. 1981, 46, 277.
Pyridine Sorption Isotherms of Coals
Energy &Fuels, Vol. 8, No. 1, 1994 215 500
300 260 -
8
400
,001
-
I
200-
I
. 0
I I
0
'
I I
I 0.80
I
#/po-
0.20;
0.40 I
I
I
0 0
20
40
60
80
0.81
0.60
100
120
Time, hours Figure 2. Incremental sorption of pyridine vapor by whole Illinois No. 6 coal at 50 "C. and exhibit no hysteresis within the range of weights for which the spring is designed. For the particular spring used in these experiments, the calibration factor was determined to be 0.48 mm/mg. The uncertainty of the cathetometer is hO.1 mm. Since two measurements must be made to obtain the weight of solvent, the uncertainty in the weight is 0.2 mm X 1mg/0.48 mm = f0.4 mg. In a typical experiment, 50 mg of sample was used, so the uncertainty per gram of sample is *0.4/0.050 g = h8 mg/g of sample. The experimental procedure was as follows. Approximately 50 mg of extract was placed in the quartz bucket and weighed on an electronic balance. The bucket and sample were then suspended on the spring. The hangdown tube was replaced and the system was evacuated to less than 1 0 3 Torr and brought to 50 0.02 "C. The system was allowed to evacuate overnight. Purified solvent was placed in the round bottom flask shown in Figure 1and frozen with dry ice/acetone. Stopcock B was closed, the stopcock C was opened to evacuate air from the flask. Then stopcock C was closed,and the pyridine was thawed and refrozen. Stopcock C was again reopened for evacuation. This procedure ensures removal of last traces of air. Stopcock A was closed and stopcocks B and C were then opened until the appropriate pressure of solvent was reached. After equilibrium was achieved, the pressure of solvent was again raised. This procedure was repeated until the entire pressure range was covered.
*
2i'
0;
200
I
I
L P
I
300
I I
I
I I
I I
,
100
I 0.90
0.94
I
;
'$0 Time,i loo Hours
40
140
3
Figure 3. Incremental sorption of pyridine vapor by pyridine extract of Illinois No. 6 coal at 50 "C.
I
n 0
20
40
, 60
I
, 80
100
120
140
160
Time, hours Figure 4. Incremental sorption of pyridine vapor by extraction residue of Illinois No. 6 coal at 50 O C . 500
, -*
WHOLE
* EXTRACT *
METHYLATED
Results The whole coals, extracts, and extraction residues were exposed to pyridine a t various vapor pressures a t 50 "C. Several incremental sorption experiments were conducted in that, once equilibrium was attained at a particular pressure, the pressure was raised and the system was again allowed to attain equilibrium. Plots of amount of pyridine sorbed versus time are shown in Figures 2-4 for the whole coal, extract, and extraction residue of the Illinois No. 6 coal. These plots are typical in that 1-2 days were generally required for each coal-pyridine system to reach equilibrium a t a given pressure. The 0-alkylated systems reached equilibrium much faster, typically in less than 1 h for the 0-butylated and 0-octylated coals a t each pressure. The increased sorption rate of the 0-alkylated coals are similar to the results of Green et al. using 0-alkylated Illinois No. 6 coal with benzene as sorbent.20 The equilibrium amounts of pyridine sorbed by the extracts are plotted against relative pressure of pyridine in Figures 5-7. For the Illinois No. 6 and WyodakAnderson whole coals, extraction residues, and extracts, (20) Green, T. K.; Ball, J. E.; Conkright, K. Energy Fuels 1991,5,610.
'0
0
0.2
0.6
0.4
0.8
1
PIP,
Figure 5. Pyridine sorption isotherms of Illinois No. 6 coals at 50 "C. the isotherms are linear or nearly linear over the relative pressure range of 0.2-0.8. For the whole Illinois No. 6 coal, the linear region was found to extend down to 0.05 relative pressure. Additionally, the slopes for the three materials for each coal are similar but the intercepts are quite different. The dramatic increase in intercept upon pyridine-extraction is particularly striking. The isotherms for the Pittsburgh No. 8 coal, shown in Figure 7, are not linear. However, a similar effect is observed upon pyridine extraction of the coal in that the extrapolated intercept is much higher for the extraction residue compared to the whole coal. The extrapolated intercept for the whole coal is near the origin for this coal, in contrast to the other two coals. We have also determined the sorption isotherms for the three 0-alkylated whole Illinois No. 6 coals. The isotherms, shown in Figure 5, are also linear over the range 0.2-0.6
Green and Selby
216 Energy &Fuels, Vol. 8, No. 1, 1994 600
600
+ WHOLE
500
11
!*
- 500
EXTRACT
-S RESIDUE
c = [kD + C ’ ~ b ] p
4oot
1
$e
represents sorption of normally diffusible species while the second term, CH,represents sorption in microvoids or *holesn. When bp < C 1, the isotherm reduces to a linear form (2)
I ’
3OOL
At sufficiently high pressures, the microvoids become saturated and will no longer sorb additional penetrant. When bp >> 1, sorption in the microvoids reaches a saturation limit, C’H and eq 1 reduces again to a linear form:
1
,
200‘
00
0 0.2
0.4
0.6
0.8
PIP.
Figure 6. Pyridine sorption isotherms of Wyodak-Anderson coals at 50 O C .
1
-
WHOLE
%
EXTRACT
,/
0.2
0
0.6
0.4
0.8
1
PIP.
Figure 7. Pyridine sorption isotherms of Pittsburgh No. 8 coals at 50 “C.
relative pressure. The intercept decreases with the size of added alkyl group.
Discussion Dual-Mode Sorption. We propose that the isotherms for the Illinois No. 6 and Wyodak-Anderson coals can be modeled in terms of a dual-mode sorption mechanism used to explain the sorptive behavior of glassy polymers. This mechanism, introduced by Meares in 1954: and further developed by Barrer et al.: Michaels et al.,5 and Vieth et a1.,6 describes the sorption mechanism in terms of one population of ordinarily dissolved sorbate which resides in the polymer matrix and is described by Henry’s law (Le., sorption is linear with pressure) while the second population of sorbate is considered t o occupy unrelaxed free volume within the polymer matrix and is described by a Langmuir isotherm. This unrelaxed free volume comes about due to restricted rotations of the polymer chains in the glassy state and represents the fixed microvoid or “holes” throughout the polymer. These microvoids act to immobilize a portion of the penetrant molecules by entrapment or by binding a t high-energy sites a t their molecular peripheries (similar to adsorption). The equilibrium isotherm of the dual sorption model can be expressed by the following equation:
C = C,
+ CH = k G + (C’Hbp)/(I+ bp)
C = kG
+ C’H
(3)
1
(I)
where C is solubility, kD is Henry’s law dissolution constant, b is the hole affinity constant, C’His the hole saturation constant, and p is the pressure. The first term, CD,
Thus the dual-mode sorption model predicts that a plot of C vs p will consist of a low-pressure linear region and high-pressure linear region connected by a nonlinear region. The sorption isotherms for methane in glassy polystyrene is an example of where this type of sorption behavior does indeed occur and that it is possible to quantitatively separate the two contributions to sorptioneZ1 Coal-Pyridine Isotherms. Equation 3 is an equation of a straight line with a slope Of kD,the dissolution constant, and an intercept of C’H, the hole saturation constant. For the pyridine sorption isotherms for the Illinois No. 6 and Wyodak-Anderson, shown in Figures 5 and 6, we propose that the slopes (kD) are a measure of the solubility of pyridine into the coal matrix and the intercepts (C’H) correspond to the micropore volume available to pyridine. Support for this interpretation includes the following. Isotherm Slopes. The amount of pyridine sorbed by the whole Illinois No. 6 and Wyodak-Anderson coals, pyridine extracts, and extraction residues increases linearly with pressure over a wide pressure range. This Henry’s law behavior suggest that dissolution is the dominant process over this pressure range. Moreover, the slopes of the isotherms are very similar for the Illinois No. 6 coal, extract, and residue. Since these materials are all of the same chemical constitution by elemental and NMR analysis,22similar slopes are expected if the linear portion of the isotherms represent dissolution of pyridine into the matrix. It is possible that mesopores are being filled in this highpressure region due to capillary condensation. However, mesoporous solids often exhibit an upturn in the isotherm in the midpressure region7which is absent in our isotherms. In addition, pyridine is known to be a good swelling solvent and we can expect the coal to swell (dissolve into the solid) in this pressure region as well. The process of swelling is expected to significantly alter the pore structure of coal in this high-pressure region.l4 The remarkable similarity of the slopes of the isotherms for all materials (whole coal, extract, and residue) is, in our opinion, best explained by a dominant process of dissolution of pyridine into the matrix. Isotherm Intercepts. The steep rise in the low-pressure region of the whole Illinois No. 6 coal suggests that it is micropores that are being filled, with a corresponding enhancement of the interaction potential and therefore of the enthalpy of adsorption. This enhanced interaction potential will occur a t about 1.5-2.5 times the diameter (21)Vieth, W. R.; Frangoulis,C. S.;Rionda,J. A., Jr. J . Colloidhterface Sci. 1966, 20, 454. (22) Davis, M. F.; Quinting, G. R.; Bronnimann, C. E.; Maciel, G. E. Fuel 1987,68,763-770.
Pyridine Sorption Isotherms of Coals of the pyridine molecule, depending on whether the pore is slitlike or cylindrical, respectively. The maximum diameter of the pyridine molecule is about 6.8 A.23 This suggests a maximum diameter of 17 A for the diameter of the pores being filled. The pressures covered in this study are too high to ascertain the shape of the isotherm in this low-pressure region. According to our model, the micropores of the whole Illinois No. 6 coal are being filled a t less than 0.05 relative pressure. This result is consistent with the behavior of the other microporous materials. For example, Cadenhead and Everett determined the adsorption isotherm of benzene on a microporous charcoal derived from anthracite and found that the micropores were essentially filled a t a relative pressure of 0.1.24 There are numerous examples of other microporous materials where the micropores are filled a t low relative pressures.' The exact pressure at which the micropores are filled will depend on the nature of the solid-sorbate interaction as well as the micropore geometry and the size of the sorbate molecule. The intercept of the isotherm of the Illinois No. 6 whole coal, when converted to a volume of pyridine per gram of dmmf coal, agrees well with the micropore volume (