Characterization of Pore Structure of Turkish Coals - Energy & Fuels


Characterization of Pore Structure of Turkish Coals - Energy & Fuels...

9 downloads 143 Views 129KB Size

Energy & Fuels 2001, 15, 331-338

331

Characterization of Pore Structure of Turkish Coals I˙ . Go¨khan S¸ enel, A. Gu¨niz Gu¨ru¨z,* and Hayrettin Yu¨cel Department of Chemical Engineering, Middle East Technical University, Ankara, Turkey 06531

Angelo W. Kandas and Adel F. Sarofim† Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received April 19, 2000. Revised Manuscript Received December 12, 2000

Pore structure of coal has a great influence on its behavior during mining, preparation, and utilization. Characterization of the pore structure of twelve Turkish coals from different geographic locations and with carbon contents varying between 61 and 84% (on dry ash-free basis) was carried out using different techniques. The volume and area of macropores were determined by mercury intrusion porosimetry. Mesopore volumes and areas were determined by N2 gas adsorption at 77 K using the Barrett-Joyner-Halenda (BJH) method. Brunauer-Emmet-Teller (BET) areas were calculated using the same data. Micropore volumes and areas were determined by the application of the Dubinin-Radushkevich (DR) equation to the CO2 adsorption data at 298 K. True and apparent densities of coals were measured by helium and mercury displacement. Pore size distributions were evaluated using data thus obtained. Small-angle X-ray scattering (SAXS) technique was also employed to determine the surface area of some samples. The highest BET surface area, 34 m2/g, was found for Tunc¸ bilek coal which has a significant mesoporous volume; while the corresponding values for the rest of the coals were less than 7 m2/g. DR surface areas which varied in the range 19-115 m2/g were larger than BET areas indicating molecular sieve character of coals. SAXS areas were larger than DR areas for some coals which can be explained by the presence of closed pores in these samples. For some coals having relatively small porosities, SAXS areas were found to be smaller than DR areas which is attributed to the inability of the method to distinguish ultramicropores of molecular dimensions which are probably accessible to CO2 molecules. SAXS surface area of Illinois No. 6 coal and a synthetic char (Spherocarb) were also measured and the values found agreed well with the ones given in the literature.

Introduction The knowledge of the pore structure of coals is of fundamental importance because of its marked influence on the behavior of coal in various existing and potential applications such as combustion, gasification, and liquefaction. The size distribution of pores in coals is very important since they govern the access and transport of reactants/reagents to the internal surface as well as the amount of surface available for reaction.1,2 It is generally accepted that the reactivity of coal in gasification reactions depends on the location of active carbon centers and catalytically active inorganic impurities; however, there appears to be no universal agreement whether they preferentially lie in micropores or larger mesopores.3-5 * Corresponding author. E-mail: [email protected] † Present Address: Chemical Engineering Department, University of Utah, Salt Lake City, UT 84112. (1) Walker, P. L., Jr. Philos. Trans. R. Soc. London 1981, 300, 6581. (2) Mahajan, O. P. Powder Technol. 1984, 40, 1-15. (3) De Koranyi, A. Carbon 1989, 27, 55-61. (4) Hurt, R. H.; Sarofim, A. F.; Longwell, J. P. Fuel 1991, 70, 10791082. (5) Hurt, R. H.; Sarofim, A. F.; Longwell, J. P. Energy Fuels 1991, 5, 1079-1082.

Coal is considered to be a suitable material for the production of activated carbon sorbents, thus pore structure of precursor coal and its evolution during carbonization and activation have been extensively studied for the preparation of carbonaceous adsorbents from coals.6-12 In recent years, attention has focused on preparations of microengineered/tailored sorbents from coal for important applications such as natural gas storage,13 flue gas cleanup,14,15 and organic pollutant removal from wastewater.16 These studies point that, (6) Finqueneisel, G.; Zimny, T.; Vogt, D.; Weber, J. V. Fuel Process. Technol. 1998, 57, 195-208. (7) Sausa, J. C.; Mahamud, J. B.; Parra, J. B,; Pajares, J. A. Fuel Process. Technol. 1993, 36, 333-339. (8) Sun, J.; Hippo, E. J.; Marsh, H.; O’Brien, W. S.; Crelling, J. C. Carbon 1997, 35, 341-352. (9) Pis, J. J.; Mahamud, M.; Parra, J. B.; Pajares, J. A.; Bansal, R. C. Fuel Process. Technol. 1997, 50, 249-260. (10) Munoz-Guillena, M. J.; Illan-Gomez, M. J.; Martin-Martinez, J. M.; Linares-Solano, A.; Salinas-Martinez de Lecea, C. Energy Fuels 1992, 6, 9-15. (11) Serrano-Talevera, B.; Munoz-Guillena, M. J.; Linares-Solano, A.; Salinas-Martinez de Lecea, C. Energy Fuels 1997, 11, 785-791. (12) Teng, H.; Ho, J.-A.; Hsu, Y.-A. Carbon 1997, 35, 275-283. (13) Sun, J.; Brady; T. A.; Rood, M. J.; Lehmann, C. M. Energy Fuels 1997, 11, 316-322. (14) Lizzio, A. A.; DeBarr, J. A.; Kruse, C. W. Energy Fuels 1997, 11, 250-25. (15) Rubio, B.; Izquierdo, M. T. Carbon 1997, 35, 1005-1011.

10.1021/ef000081k CCC: $20.00 © 2001 American Chemical Society Published on Web 02/01/2001

332

S¸ enel et al.

Energy & Fuels, Vol. 15, No. 2, 2001

in addition to some chemical factors,the nature of pores present in precursor coal has a great influence on properties of chars and activation products obtained because they carry fingerprints of pore structure of the precursor coal. In addition to the above applications, many properties of coal are determined or influenced by textural properties of coal, among which the ability of coal seams to hold methane, water uptake in coal-water slurries and spontaneous ignition of coal can be given.2 For the quantitative characterization of pore structure, estimation of pore surface area, pore volume, and pore size distributions together with the density measurements are needed. A convenient classification, recommended by IUPAC,17 divides pores sizes roughly into the following groups: micropores < 2.0 nm, mesopores in the range of 2.0-50 nm, and macropores > 50 nm. It may be also desirable to subdivide micropores into those smaller than about 0.7 nm (narrow micropores or ultramicropores) and those supermicropores in the range of 0.7 to 2.0 nm. In most cases, a combination of different techniques needs to be used for complete coverage of all sizes of pores. A large amount of information on the application of common physical methods on the characterization of texture of coals from deposits in various countries exist in the literature.18-29 These studies have shown that coal has a pore structure containing micro-, meso-, and macropores in various proportions depending on the origin of coal and its rank. Two most common methods used for the pore size characterization include mercury intrusion porosimetry and gas adsorption techniques. Total pore volume and porosity of coals can be determined by measuring apparent and true densities using mercury, and helium pycnometry techniques. The surface area of coals is customarily evaluated by the application of BrunauerEmmet-Teller (BET) and Dubinin-Radushkevich (DR) equations to N2 and CO2 adsorption data, respectively. The fundamentals and limitations of these characterization techniques for porous solids, in general, have been discussed extensively.30-32 In particular, since coal is a heteropolymeric substance having acidic and basic functional groups on parts of its surface, additional (16) Teng, H.; Hsieh, C.-T. J. Chem. Technol. Biotechnol. 1999, 74, 123-130. (17) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, T; Simieniewska, T. Pure Appl. Chem. 1985, 57, 603-619. (18) Toda, Y. Fuel 1972, 51, 108-112. (19) Toda, Y.; Toyoda, S. Fuel 1972, 51, 199-201. (20) Toda, Y.; Hatami, M.; Toyoda, S.; Yoshida, Y.; Honda, H. Fuel 1971, 50, 187-200. (21) Spitzer, Z. Powder Technol. 1981, 29, 177-186. (22) Gan, H.; Nandi, S. P.; Walker, P. L. Fuel 1972, 51, 272-277. (23) Walker, P. L.; Patel, R. L. Fuel 1970, 49, 91-94. (24) Debelak, A. D.; Schrodt, J. T. Fuel 1979, 58, 732-736. (25) Ng, S. H.; Fung, D. P.; Kim, S. D. Fuel 1984, 63, 1564-1569. (26) Walker, P. L.; Verma, S. K.; Utrilla , J.; Davis, A. Fuel 1988, 67, 1615-1623. (27) Unsworth, J. F.; Colin, S. F.; Jones, L. F. Fuel 1989, 68, 1826. (28) Gallegos, D. P.; Smith, D. M.; Stermer, D. L. In Characterization of Porous Solids I; Unger, K. K., et al., Eds; Elsevier Science Publishers BV: Amsterdam, 1998; pp 509-508. (29) Patel, R. L.; Nandi, S. P.; Walker, P. T., Jr. Fuel 1972, 51, 4751. (30) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity; Academic Press: New York, 1982. (31) Sh. Mikhail, R.; Robens, E. Microstructure and Thermal Analysis of Solid Surfaces; John Wiley & Sons: New York, 1983. (32) Lowell, S.; Shields, J. E. Powder Surface Area and Porosity, 3rd ed.; Chapman and Hall: London, 1991.

complications might arise due to specific interactions of fluid molecules with coal surface. It has been reported that CO2 surface areas may be influenced by the solubility of CO2 in coal26 which may be more pronounced for swelling coals. One important drawback of the use of gas-adsorption techniques is that they do not provide information on closed porosity. This drawback can largely be overcome by small-angle X-ray scattering (SAXS) technique. Since x-radiation “sees” the entire porosity, SAXS technique shows a promise for determining pore size distribution/area for both open and closed porosity.33-36 Turkey has considerable amount of coal reserves of which lignites constitute the major portion.37 Some previous studies about combustion, pyrolysis, liquefaction, and characterization of Turkish coals have already been reported in the literature.37-43 To our knowledge a systematic study on the pore structure of Turkish coals has not been performed, yet. The present study is part of a comprehensive research carried out on carbonization, activation, and reactivity of Turkish coals. The work presented here on the textural properties of Turkish coals, we hope, will form a basis for comparison of coals of different origins and will help to the understanding of different behaviors observed. Experimental Work Chemical Analyses of Coals. Coal samples representing a range of rank and geographic locations were obtained from two organizations: Mineral Research and Exploration Institute (MTA) and Turkish Coal Enterprises (TKI). No sample bank involving a formal coding system exists in Turkey. However these institutions are the major suppliers of coal samples to researchers. It is known that common sampling techniques for collecting bulk samples from fields are used by these institutions which themselves are also involved in coal research. Prior to analyses, coals were stored as received in sealed polyethylene bags. Samples before grinding had a size of 5-10 mm. Coal samples were ground to a particle size fraction of 0.250-0.500 mm. The carbon, hydrogen and nitrogen contents were determined by LECO-CHN-600 elemental analyzer. Total sulfur contents were measured by LECO sulfur analyzer. The proximate analysis of the original samples were carried out according to ASTM D-3173, D-3174, and D-3175 standards. Elemental and proximate analyses of coal samples are given in Table 1. Multiple analyses were carried out and with reproducibility of better than 5%, average values are reported. In the SAXS measurements, Illinois No. 6 coal and a synthetic char, spherocarb (Analabs Inc.), were also studied for comparison. (33) Kalliat, M.; Kwak, Y. C.; Schmidt, P. W. Small Angle X-ray Investigation of the Porosity in Coals. In New Approaches to Coal Chemistry; Blaustein, B. D., et al., Eds.; ACS Symposium Series 169, Amercan Chemical Society: Washington, DC, 1981. (34) Spitzer, Z.; Ulicky, L. Fuel 1976, 55, 21-24. (35) Lin, J. S.; Hendricks, R. W.; Harris, L. A.; Yust, C. S. J. Appl. Crystallogr. 1978, 11, 621-625. (36) Xu, Y.; Koga, Y.; Watkinson, A. P. Fuel 1994, 73, 1797-1801. (37) Coal(in Turkish); Kural, O., Ed.; Kurtis¸ Matbaası: I˙ stanbul, 1991. (38) Ekinci, E.; Yalkin, G.; Atakul, H.; Erdem-Senatalar, A. J. Inst. Energy 1988, 155, 189-191. (39) Yu¨ru¨m, Y.; Altuntas¸ , N. Fuel 1998, 77, 1809-1814. (40) Ceylan, K.; Karaca, H.; O ¨ nal, Y. Fuel 1999, 78, 1109-1116. (41) Yaman, S.; Cinpolat, E.; Karatepe, N.; Ku¨cu¨kbayrak, S. Thermochim. Acta 1999, 335, 63-68. (42) Gu¨ldogan, Y.; Durusoy, T.; Bozdemir, T. O ¨ . Thermochim. Acta 1999, 335, 75-81. (43) Gu¨ru¨z, G.; Yu¨ru¨m, Y.; Bac¸ , N.; Orbey, H.; Togˇrol, T.; S¸ enel, A. G. Turkish Scientific and Technical Research Council (TU ¨ BI˙ TAK) Project, TBAG-575/B, 1987.

Pore Structure of Turkish Coals

Energy & Fuels, Vol. 15, No. 2, 2001 333 Table 1. Chemical Analysis of Coals

coal

proximate analysis (db, wt %)

elemental analysis (dafb, wt %)

location, ID

geological age

rank

ash

volatile matter

fixed carbon

C

H

S

N

O by diff.

Zonguldak, Z Armutc¸ uk, A1 Amasra, A2 Tunc¸ bilek, T Mengen, M Seyito¨mer, Y Beypazarı, B1 Elbistan, E Soma, S Konya, K Orhaneli, O Go¨ynu¨k, G

Carbonifereous Subcarbonifereous Westphalian Miocene Miocene Miocene Miocene Pleisto-pliocene Miocene Pleisto-pliocene Miocene Miocene

bitum. bitum. bitum. lignite lignite lignite lignite lignite lignite lignite lignite lignite

18.4 7.6 15.0 20.1 18.7 18.6 49.5 25.8 37.5 31.1 16.2 45.3

25.9 32.3 31.5 38.1 44.7 44.0 32.4 46.6 44.9 43.6 53.3 35.0

55.7 60.1 53.5 41.8 36.6 37.4 18.1 27.6 17.6 25.3 30.5 19.7

83.1 82.5 74.5 71.4 69.3 66.6 63.4 63.2 62.8 62.4 62.0 61.3

4.9 4.8 5.3 4.8 6.9 4.7 5.3 4.4 5.1 5.6 5.6 4.6

0.6 1.0 0.9 2.7 10.8 1.5 5.5 4.4 1.4 5.5 4.0 1.8

0.7 1.0 1.5 2.3 1.6 1.3 0.9 0.8 1.2 1.4 1.4 1.3

10.7 10.7 17.8 18.8 11.4 25.9 24.9 27.2 29.5 25.1 27.0 31.0

Figure 1. Measurement Techniques Used for Pore Structure Characterization. Physical Characterization Techniques. Measurement techniques used for the characterization of pore structure of coals are summarized in Figure 1. Prior to characterization experiments, the samples were first oven dried at 383 K for 2 h and then outgassed under a vacuum of 0.13 Pa at 383 K for several hours. Gas Adsorption Measurements. N2 and CO2 adsorption data were obtained with an automated surface area analyzer (Model ASAP 2000 of Micromeritics Co., Inc.) at 77 and 298 K, respectively. Volumes and areas of mesopores were calculated by the available software which applies Barrett-JoynerHallenda (BJH) method44 to the nitrogen desorption data at 77 K. This method employs the Kelvin equation and a correction for adsorbed layer thinning during desorption from an assumed cylindrical pore model. In calculations, the cross sectional area of the nitrogen molecule26 was taken as 0.162 nm2 and IUPAC definition of mesopore size range of pores was taken into account. BET surface areas were calculated using N2 adsorption at 77 K in the relative pressure range of 0.05 to 0.25 with the help of software. The micropore area and volume of the samples were estimated by application of the DR equation to CO2 adsorption at 298 K. The cross sectional area23 and the density18 of CO2 molecules were assumed as 0.253 nm2 and 1.038 g/cm3, respectively. The saturation vapor pressure of CO2 at 298 K was taken as 6.4 MPa.20 Mercury Porosimetry. Pore volume and area of the pores in the macropore region as well as the apparent density of the samples were determined by mercury intrusion technique using a commercial mercury intrusion porosimeter (Model 9310, Micromeritics Inc. Co.). The minimum pressure at which the interparticle voids are completely filled was determined experimentally using nonporous glass powders having a size (44) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373-380.

range similar to that of the coal particles. This pressure was found as 0.151 MPa which corresponds to a pore diameter of about 8200 µm according to the Washburn equation.30 This pore diameter can be considered as the upper limit of macropore range which can be measured by the mercury porosimeter for the particle size range used. In the Washburn equation surface tension and contact angle are taken as 0.485 N/m and 130°, respectively. Apparent density values were calculated from the measurements of mercury displaced by the samples at 0.15 MPa since, in the determination of apparent density, the interparticle voids should be excluded while internal pores should be included. Helium Pycnometry. True densities of coal samples were measured by helium displacement using a commercial helium pycnometer (Model 1302, Micromeritics. Inc. Co.) based on the assumption of no closed porosity being present. Total specific pore volumes and porosities were calculated from apparent and true densities measured. SAXS Measurements. Small angle X-ray scattering (SAXS) generally refers to scattering phenomena observed at scattering angles of 0-6° 2θ. Scattering in this range of angles occurs when there are differences in the number of electrons per unit volume in different regions of sample. For coals there are two kinds of inhomogeneities which can act as scattering centers: pores and minerals dispersed within the carbonaceous matter. Scattering data from coals suggest that the pores and the mineral matter scatter independently of each other, so that when the contribution of the minerals to the scattering must be considered, the coals can be taken as two independent systems: coal-void and coal-mineral matter.33,35 Kalliat et al.33 measured the scattering of the mineral matter obtained from coals by low-temperature ashing, and concluded that its contribution to the total specific surface area can be ignored for most coals. In this study the contribution of mineral matter to the total surface area was also neglected and coal is assumed to be a two-phase system. SAXS data are often presented as a plot of the logarithm of the scattering intensity, I, as a function of the scattering vector, h, defined as

h ) (4π sin θ)/λ

(1)

where λ is the X-ray wavelength and 2θ is the scattering angle. The scattering in the region where scattering vector, h, sufficiently far from zero, corresponds to Porod’s law of scattering45 and provides information on the surface area. The Porod’s Law with sharp and nonfractal interfaces between the phases, is given as

I(h) )

2π(∆F)2S h4

(2)

where ∆F is the electron density difference, S is the total surface or interface surface of the scattering system. To avoid absolute intensity determinations, intensity values can be (45) Porod, S. Kolloid Z. 1951, 124, 83-114.

334

S¸ enel et al.

Energy & Fuels, Vol. 15, No. 2, 2001

normalized by using the total integrated intensity, known as “invariant”, Q. Since it is not experimentally possible to measure the intensity over the entire range of the scattering vector, the evaluation of Q can be carried out after splitting the whole range into three parts:

Q ) Q1 + Q2 + Q3 )

∫h ∫

∞ 2

0

I(h) dh )

hmax 2 h hmin



hmin 2

h I(h) dh +

0



I(h) dh +



hmax

h2 I(h) dh (3)

The middle range Q2 can be evaluated from the experimentally measured intensity data, corresponding to the minimum and maximum values of the scattering vectors. As the angle becomes very small where h f 0, the scattered intensity rises exponentially. In this region the extrapolation to the zero intensity which actually corresponds to the lower limit of the integration, can be performed with Guinier’s law:46

(

I(h) ) I(0) exp -

)

h2Rg2 3

(4)

where I(0) is the scattering intensity at zero scattering vector and Rg is the radius of gyration, a measure of the dimension of scatterers. Therefore, Q1 can be calculated from plot the of ln(I) vs h2, after obtaining the I(0) by extrapolating the experimental data to the origin. The upper range, Q3, can be approximated by Porod’s law which describes variation of the intensity proportional to the inverse fourth power of the scattering vector as h f ∞. The total surface area per unit mass (SSAXS) can then be estimated as

lim {h4 I(h)} π h f ∞ SSAXS ) FHe ∞ 2 h I(h) dh



(5)

0

where  and FHe are the porosity and the true density of the sample, respectively. The SAXS measurements were carried out at MIT with a rotating anode X-ray RIGAKU generator operating at 40 kV and 20 mA equipped with a copper target. The intensity of the scattered beam was recorded during 600 s with a twodimensional detector, under pinhole collimation conditions with a sample-to-detector distance of 28.5 cm. Powdered samples were sandwiched in a cell of 1 mm in thickness and diameter. The radially averaged experimental intensity values were then corrected for background scattering. Subtraction and analysis of the SAXS data were carried out by using a set of programs prepared by Glatter.47 Evaluation of surface area from SAXS data requires that the measurements of the scattered intensity be extended up to large angles, where the intensities are rather weak due to the relatively low porous nature of the coals and certain degree of absorbance of the X-rays by the coals. The background signal was high most probably due to scattering by air molecules, since the scattering experiments were conducted without sealing the system. Because of these effects, subtraction of the background signal was carefully conducted and verified in several regions of the scattering vector.

Results and Discussion Pore Volumes and Distributions. For all coals, N2 isotherms showed a type II behavior according to Branuer classification.30 All coal samples except Go¨ynu¨k(46) Guinier, A.; Fournet, G.; Walker, C. B.; Yudowitch, K. L. Small Angle Scattering of X-rays; Wiley: New York, 1955. (47) Glatter, O.; Kratky, O. Small-Angle X-ray Scattering; Academic Press Inc. Ltd.: London, 1982; p 472.

Figure 2. N2 Adsorption/Desorption Isotherms at 77 K for (a) Elbistan coal, (b) Tunc¸ bilek coal.

(G) and Tunc¸ bilek(T) showed hysteresis curves similar to type B according to deBoer’s classification32 as exemplified for Elbistan(E) coal in Figure 2a. This type of hysteresis is usually attributed to slit-shaped pores. Go¨ynu¨k and Tunc¸ bilek coals behave similarly and have type E hysteresis curves as shown in Figure 2b for Tunc¸ bilek coal. This type of hysteresis can be explained by the presence of ink-bottle pores.32 In pores of this shape, emptying of the wide portion will be delayed during desorption until the narrow neck can evaporate. Therefore the desorption curve shows a small slope at high relative pressures and a large slope where the wide part of the pore empties. Figure 3, parts a and b, show the mercury intrusion curves for twelve coals studied. No discontinuities in intrusion curves may be taken as an evidence of no significant micro crack formation at high pressures. In the region below 0.151 MPa, the rapid rise in the volume is caused by mercury filling the interparticle voids. Any further mercury penetration at higher pressure is thought to be due to the presence of smaller pores of coals and due to the compressibility of coals. Under these circumstances, the macropore volume of the samples was initially determined as

V′MACRO ) Vporew24.8 21MPa - Vporew0.151 MPa

(6)

The upper limit in this pressure range as “24.821 MPa”

Pore Structure of Turkish Coals

Energy & Fuels, Vol. 15, No. 2, 2001 335

Figure 3. Mercury intrusion curves of coals.

corresponds to the minimum pore diameter 50 nm according to the Washburn equation30 which is the lower limit in macropore range. A correction for the compressibility was introduced for the intrusion above 10.1 MPa pressure. Compressibility values, κ, of the coals were calculated using the gradients from the intrusion curves:

κ ) unit conversion factor × gradient × true density (7) The gradient was determined over the pressure range 10.1-24.821 MPa. Then macropore volumes were corrected for the compressibility effect using the equation

VMACRO ) V′MACRO - (gradient) × (∆P)

(8)

In Table 2, compressibility values, κ, and specific volumes of macropores corrected for compressibility of coal, VMACRO, are given. The compressibility values of the samples which vary between 2.3 and 5.9 × 10-10 Pa-1 except the value found for Elbistan coal, are in agreement with the values reported in the literature.19,25,27 In Table 2, mesopore volume, VMESO, micropore volume, VMICRO, and total pore volumes based on both pore volume distributions, VTOT,P, and density measurements, VTOT,D, are also presented. Mercury densities, FHg could be reproduced with (0.003 g/cm3 and helium densities, FHe could be reproduced with (0.002 g/cm3. Thus total porosity, VTOT,D which can be estimated from 1 - FHg/FHe has an uncertainty of approximately (5%. Total pore volumes based on density measurements are,

in general, close to the values based on pore volume distributions although each individual value is based on several methods involving different assumptions and experimental accuracy. Pore volume distribution data are also shown in Figure 4 as histograms. In general, the contribution of mesopores to the total porosity is much less compared to the others except for one coal sample, Tunc¸ bilek. No trend was observed with respect to rank and carbon content. When pore size distributions of two bituminous coals of similar rank, Zonguldak and Armutc¸ uk are compared, important differences were observed. Zonguldak coal is found to be macroporous with relatively high porosity, characteristic of coking coals. In fact this sample is from the main coking coal reserve of Turkey.37 The other sample, Armutc¸ uk, which is from a nearby locality, is of low porosity and has a higher percentage of microporosity. This coal has been reported to possess poor coking characteristics; this behavior can be attributed to its low porosity and absence of sufficient macroporosity of precursor coal in addition to some differences in their chemical structure.48 Density Measurements. Helium and mercury density values of the coals treated as the true and the apparent values are included in Table 2.Apparent and true densities of coals on dry ash free basis(dafb) are shown in Figure 5, parts a and b, respectively, together with the values reported by other workers; for the American,22 Canadian,25 Japanese;19 and those of Unsworth et al.27 for coals of different origins. For the conversion of densities to dafb, density of ash is assumed to be 2.7 g/cm3. Data are found to be in good agreement with literature for relatively high carbon content coals. For low carbon content coals there exist limited data for comparison but they seem to follow the general trend. In general, there is significant scatter of existing data which appears to be more pronounced for the low carbon range. Surface Areas. In Table 3, areas based on different techniques are given. Since SBET values for most coal samples were low (N2 adsorption is small) the uncertainty was, therefore, high, (20%, based on the reproducibility obtained for three repeated tests with one coal sample, namely Tunc¸ bilek(T). SDR values calculated from CO2 adsorption at 298 K were relatively high and thus could be reproduced within (10%. The uncertainty in SSAX values is estimated to be (15%. Macropore areas, SMACRO, are calculated from mercury intrusion curves assuming cylindrically-shaped pores.30 As expected, SMACRO values are low and in the range of about 0.2-0.5 m2/g. Comparison of SMESO and SBET values would be interesting. It should be recalled that both SMESO and SBET are obtained from nitrogen adsorption data at 77 K, but from different regions of relative pressure and using different models. SMESO are calculated from BJH model which is based on Kelvin equation of capillary condensation and takes into account adsorbed layer thickness, whereas SBET are calculated from monolayer capacity of the BET model which is based on multilayer adsorption. As can be observed from Table 3, SMESO and SBET values for the majority of coals are quite close. (48) Erbatur, G.; Erbatur, O. International Seminar on Coal Technology, I˙ stanbul, 1982.

336

S¸ enel et al.

Energy & Fuels, Vol. 15, No. 2, 2001 Table 2. Densities and Pore Size Distributions of Coals

coal ID

κ × 1010 (Pa1-)

VMACRO (cc/g)

VMESO (cc/g)

VMICRO (cc/g)

VTOT,P (cc/g)

FHg (g/cm3)

FHe (g/cm3)

VTOT,D (cc/g)

Z A1 A2 T M Y B1 E S K O G

2.95 2.33 3.07 4.10 5.91 5.20 2.81 10.10 5.55 4.10 3.07 3.97

0.136 0.008 0.007 0.040 0.004 0.021 0.048 0.102 0.026 0.028 0.006 0.014

0.001 0.001 0.005 0.042 0.006 0.003 0.005 0.003 0.010 0.003 0.001 0.006

0.012 0.017 0.028 0.023 0.009 0.019 0.012 0.017 0.032 0.005 0.010 0.012

0.149 0.026 0.040 0.145 0.019 0.062 0.065 0.122 0.068 0.036 0.017 0.032

1.298 1.304 1.232 1.228 1.331 1.302 1.533 1.262 1.468 1.343 1.288 1.589

1.543 1.348 1.297 1.448 1.348 1.398 1.649 1.498 1.597 1.399 1.348 1.698

0.122 0.025 0.041 0.124 0.009 0.053 0.046 0.125 0.055 0.030 0.035 0.040

Figure 4. Histograms for macro-, meso-, and micropore volumes of coals.

These results indicate that SBET of coal samples measures, principally mesoporous area and superporosity in samples is not present to a significant extent, otherwise, SBET >SMESO is expected. For Tunc¸ bilek(T) coal it is observed that SMESO > SBET which can be explained by the existence of ink-bottle pores in which a larger volume of gas is condensed in pores having a relatively small area.32 In fact, this samples was shown to exhibit hysteresis curves characteristic of ink-bottle pores (See Figure 2b) which supports the above explanation. SBET of coals were found to be less than 7.4 m2/g, except Tunc¸ bilek coal which has a relatively high SBET value (34 m2/g). This particular coal is the only sample which has a significant mesopore volume which contributes to the SBET. SBET of coals reported in the literature varies in the range 1-100 m2/g. Gan et al.22 reported values less than 10 m2/g for coals with carbon contents 76-84%. They found that SBET of coals were less than 1 m2/g on both sides of this range of carbon content. The anthracites having surface areas of 5-8 m2/g were the exception to this general behavior. SBET values reported by Ng et al.25 for the thirteen Canadian coals were less than 4 m2/g. Unsworth et al.27 found SBET values less than 20 m2/g for twenty-one coal samples (78-89% carbon-daf) from different countries. In the latter two studies, no relation between SBET with the coal rank was found. Dubinin-Radushkevich areas, SDR, of coals were found to be higher than the SBET. This behavior is usually attributed to the activated diffusion effects on the adsorption of N2 at 77 K. It was well documented in the literature49,50 that the mechanisms of adsorption of N2 and CO2 on carbonaceous materials are similar.

Figure 5. Comparison of apparent and true densities of coals on dafb. (a) Apparent density, (b) true density. Table 3. Surface Areas of Coals coal ID

SMACRO (m2/g)

SMESO (m2/g)

SBET (m2/g)

SDR (m2/g)

SSAXS (m2/g)

Z A1 A2 T M Y B1 E S K O G

0.23 0.19 0.23 0.41 0.27 0.55 0.53 1.38 0.33 0.21 0.22 0.31

1.0 1.0 4.3 38.9 5.0 2.2 2.6 1.5 7.8 2.4 0.2 4.6

1.0 1.1 4.1 33.6 7.4 2.3 2.6 2.0 6.8 2.2 1.1 4.5

44.6 61.3 101.6 81.3 30.7 67.8 43.5 59.7 115.4 18.7 34.0 42.6

114 31 52 146 15 63 102 162 69 48

However, if the micropores are very narrow the entry of the N2 molecules may be kinetically restricted at 77 K so that equilibrium cannot be established in a reasonable period of time allowed for adsorption. Critical dimensions of both molecules are very similar, 0.28 nm for CO2 and 0.30 nm for N2, but higher temperature (298

Pore Structure of Turkish Coals

Energy & Fuels, Vol. 15, No. 2, 2001 337

Figure 6. SAXS profiles of samples.

K) of adsorption for CO2 facilitates the entry into the narrow micropores. It appears that coal samples investigated contain an appreciable percentage of ultramicropores or pores with very narrow entrances which are inaccessible to N2 at 77 K; however, they are available for the adsorption of CO2 at 298 K. SDR of coals vary between 19 and 115 m2/g which is within the range of 18-250 m2/g found in the literature. Koranyi et al.3 reported SDR between 18 and 20 m2/g for three British bituminous coals. Sausa et al.7 reported SDR less than 70 m2/g for some Spanish coals, which varied independent of carbon content. Mahajan2 found a parabolic relationship beetween SDR and carbon content from 70% to 92%; however, no such a trend was observed in the present study. SAXS Areas. The scattering curves for some of the coals and spherocarb are shown in Figure 6. Spherocarb shows much higher scattering than the coals due to its well-known high microporosity. Guinier plots for the same samples are represented in Figure 7. To verify the validity of Porod’s law h4 I(h) versus h4 is plotted in Figure 8. The invariant plot does not show significant variation from Porod’s law at relatively high h values, exhibiting more or less a flat h4 I(h) profile; although with some scatter of data. To test the entire SAXS measurement and analysis system, the surface area of the Illinois No. 6 coal was measured and was found to be 186 m2/g. This value is in good agreement with those given by Spitzer and Ulicky,34 Lin et al.,35 and Glatter et al.47 as 179 m2/g, 140 ( 20 m2/g, and 195 ( 27 m2/g, respectively. SAXS area of spherocarb was also measured as it is a well-characterized, highly porous material. On the basis of the apparent and true density values of 1.5 and 2.1 g/cm3, the SAXS area of the spherocarb was found as 1380 m2/g which is in the expected range. For all coal samples, SSAXS were found to be larger than SBET which indicates that it measures micropore area in addition to mesopore/ macropore area. Kalliat (49) Garrido, J.; Linares-Solano, A.; Martin-Martinez, J. M.; MolinaSabio, M.; Rodriguez-Reinoso, F.; Torregrosa, R. Langmuir 1987, 3, 76-81. (50) Cazorlo-Amoros, D.; Alcaniz-Monge, J.; de la Casa-Lillo, M. A.; Linares-Solano, A. Langmuir 1998, 14, 4589-4596.

Figure 7. Guinier plots of samples.

Figure 8. Porod invariant plots for samples.

et al.33 argued that SAXS measures only mesopore and macropore area. This idea was not supported by the work of others.34-36 Our data also show that SAXS is capable of measuring micropore area; however, there may be a lower limit on the size of micropores which can be measured by this technique as will be explained later. When SSAXS and SDR values are compared, three cases were observed. For two samples (Go¨ynu¨k and Seyito¨mer) both areas were found to be comparable. The case of SSAXS > SDR was observed for samples (Zonguldak, Tunc¸ bilek, Beypazarı, and Elbistan) and can be explained by the presence of closed porosity inaccessible to CO2 molecules in adsorption measurement.The case of SSAXS < SDR was observed for samples (Armutc¸ uk, Amasra, and Soma) which is more difficult to explain. A plausible explanation is that these samples may contain very small pores of molecular dimensions so that scattering from these ultramicropores could not be distinguished from the background scattering and other nonhomogeneties in the samples. If these pores are accessible to CO2 at 298 K, they would contribute to SDR area and thus SSAXS < SDR results. These samples have comparatively small porosities with significant micropore contribution which may be taken as a support of the above hypothesis.

338

S¸ enel et al.

Energy & Fuels, Vol. 15, No. 2, 2001

Conclusions In this work a variety of characterization techniques were applied for the determination of pore structure of a series of Turkish coals. Some of these techniques were found to be complementary and useful information can be obtained by using a combination of them. For example, comparison of BET(N2,77 K), Dubinin-Radushkevich (CO2, 298 K), and SAXS areas give information about the nature of porosity; closed vs open and ultra- vs supermicroporosity present in coals. In this study, Dubinin-Radushkevich areas were found to be consistently larger than BET areas indicating molecular sieve character of Turkish coals. Comparison of Dubinin-Radushkevich and SAXS areas implied closed porosity in some samples. One particular coal, Tunc¸ bilek has a significant mesoporosity with an even pore size

distribution and may be expected to serve as a suitable material for adsorption of organic molecules from water. No noticeable trend was observed between pore size distribution and coal rank. In fact, textural properties of some coals with similar carbon contents were completely different indicating the complex nature of factors playing a role in the formation of pore network of coals. Acknowledgment. Support by TUBITAK Grant KTCAG-86 and METU Grant AFP 92-03-04-03 are gratefully acknowledged. The experimental assistance of Kerime Gu¨ney, Turgut Aksakal, and Selahattin Uysal is greatly appreciated. I.G.S. also thanks TUBITAK for support through a NATO fellowship which made it possible to carry out SAXS measurements at MIT. EF000081K