UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of coal composition upon gas sorption and transmissibility of bituminous coal Clarkson, Christopher Raymond 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-0378.pdf [ 5.31MB ]
Metadata
JSON: 831-1.0052727.json
JSON-LD: 831-1.0052727-ld.json
RDF/XML (Pretty): 831-1.0052727-rdf.xml
RDF/JSON: 831-1.0052727-rdf.json
Turtle: 831-1.0052727-turtle.txt
N-Triples: 831-1.0052727-rdf-ntriples.txt
Original Record: 831-1.0052727-source.json
Full Text
831-1.0052727-fulltext.txt
Citation
831-1.0052727.ris

Full Text

THE EFFECT OF COAL COMPOSITION UPON GAS SORPTION ANDTRANSMISSII3ILITY OF BITUMINOUS COALbyCHRISTOPHERRAYMOND CLARKSONB.A.Sc., The University ofBritish Columbia, 1992A THESIS SUBMITTED iN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPUED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Geological SciencesWe accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAAUGUST 1994© Christopher Raymond ClarksonIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. it is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of , cA( c —The University of British ColumbiaVancouver, CanadaDate /4i- 23DE-6 (2)88)11ABSTRACTThe effect of bituminous coal composition, particularly the organic fraction, upon gassorption and transmissibility is investigated. Micropore capacities of bituminous coals,determined from low pressure carbon dioxide adsorption, show a general increase with total andstructured vitrinite content. Conversely, micropore capacities generally decrease with an increasein inertinite and mineral matter content. High pressure methane monolayer capacities show asimilar trend. Micropore size distributions indicate an increase in the total number ofmicroporesand a slight decrease in mean pore diameter with vitrinite content.Mesopore volumes and surface areas, determined through nitrogen sorption, show ageneral decrease with vitrinite content and increase with inertinite content ofbituminous coal.Vitrinite therefore contains more microporosity and less mesoporosity than inertinite. Hysteresisloops of sorption isotherms indicate that mesopores of the coals studied are slit-shaped.Permeabilities of bituminous coals obtained through the use of a permeameter capable ofmeasuring permeabilities on a bed-by-bed scale show that brighter coal lithotypes are morepermeable than dull lithotypes. The order of decreasing permeability with lithotype is: bright>banded> fibrous > banded dull> dull. The increase in permeability with increased brightness ofcoals is due to the presence of abundant macrofracturing (cleating) in bright coal. For onesample, permeabilties were found to increase with vitrinite content.‘UTABLE OF CONTENTSABSTRACT.iiTABLE OF CONTENTS iiiLIST OF FIGURES viiiLIST OF TABLES xACKNOWLEDGEMENTS xiiCHAPTER 1 INTRODUCTORY STATEMENTS 11.1 INTRODUCTION 11.2 NATURAL GAS GENERATION FROM COAL 21.3 GASRETENTIONINCOAL 31.4 STRUCTURE OF THESIS 41.5 REFERENCES 5CHAPTER 2 GAS SORPTION THEORY 82.1 INTRODUCTION 82.1.1 Definitions 82.1.2 Concepts 122.2 THEORIES OF ADSORPTION 192.2.1 BET Theory 192.2.2 Type Ilsotherm - Dubinin Theory of Volume FillingforMicropores 252.2.3 Type IVlsotherm - BJH Theory 34iv2.3 CHOICE OF ADSORBATE .412.4 CONCLUSIONS 432.5 REFERENCES 44CHAPTER 3 VARIATION iN MICROPORE CAPACITY AND SIZE DISTRIBUTIONWITH COMPOSITION IN HIGH AND MEDIUM-VOLATILE BITUMINOUSCOAL OF THE WESTERN CANADIAN SEDIMENTARY BASIN:IMPLICATIONS FOR COALBED METHANE POTENTIAL 473.1 ABSTRACT 473.2 INTRODUCTION 483.2.1 Research Objectives 503.3 BACKGROUND 513.3.1 Dubinin Theory of Volume FillingforMicropores 513.3.2 Langmuir andBET Theory 543.4 METHODS 563.5 RESULTS 583.5.1 Gates suite 583.5.1.1 Proximate, rank, and petrographic data 583.5.1.2 Gas adsorption 633.5.2 Alberta suite 673.5.2.1 Proximate, rank, and petrographic data 673.5.2.2 Gas adsorption 753.6 DISCUSSION 823.6.1 Gates suite 823.6.1.1 Dubinin-Radushkevich Plots 863.6.1.2 Dubinin-Astakhov Differential Pore Volume Plots 87V3.6.1.3 Langmuir and BET Analysis .903.6.1.4 Equilibrium Moisture 913.6.2 Alberta suite 933.6.2.1 D-RPlots 953.6.2.2 D-A Differential Pore Volume Plots 953.6.2.3 Langmuir and BET Analysis 963.6.3 Comment on the Origin andNature ofMicroporosity in Coals 973.7 CONCLUSIONS 993.8 REFERENCES 101CHAPTER 4 VARIATION IN MESOPORE VOLUME AND SIZE DISTRIBUTION WITHCOMPOSITION IN A HIGH-VOLATILE BITUMINOUS COAL OF THEWESTERN CANADIAN SEDIMENTARY BASIN: IMPLICATIONS FORCOALBED METHANE TRANSMISSIBILITY 1064.1 ABSTRACT 1064.2 INTRODUCTION AND RESEARCH OBJECTIVES 1074.3 BACKGROUND 1094.3.] Barret, Joyner, andHalenda (BJH) Theory 1094.4 METHODS 1114.5 RESULTS 1124.5.1 Proximate, ranlç andpetrographic data 1124.5.2 Isotherms and hystersis loops 1174.5.3 BET andBJH surface areas 1194.5.4 Mesopore size distributions and volume 126vi4.6 DISCUSSION.1294.6.1 Relationship between mesopore volume andgas yieldsfrom desorpliontests 1334.6.2 Implicationsfor coalbedmethane transmissibility 1334.7 CONCLUSIONS 1364.8 REFERENCES 138CHAPTER 5 VARIATION IN PRESSURE-DECAY PROFILE PERMEABILITY-DERIVEDPERMEABILITIES WITH LITHOTYPE AND MACERAL COMPOSITION OFCOALS 1405.1 ABSTRACT 1405.2 INTRODUCTION AND RESEARCH OBJECTIVES 1415.3 EFFECT OF COAL STRUCTURE ON PERMEABILITY 1425.3.1 Cleat systems 1425.3.2 Microstructures 1435.4 THE PRESSURE-DECAY PROFILE PERMEAMETER 1455.5 METHODS 1475.6 RESULTS 1495.6.1 Lithotype, Megascopic Structure, andMeasurement SurfaceDescriptions 1495.6.2 Permeability Variation with Lithotype Composition 1615.6.3 Permeability Variation withMaceral Composition 1635.6.4 Effect ofRank upon Profile Permeability 1665.7 DISCUSSION 1675.8 CONCLUSIONS 169vii5.9 REFERENCES .171CHAPTER 6 SUMMARY AND CONCLUSIONS.1726.1 EFFECT OF COAL COMPOSITION UPON GAS SORPTION CAPACITYAND TRANSMISSIBILITY 1726.2 FUTURE WORK 1736.3 REFERENCES 174APPENDIX 175viiiLIST OF FIGURESFigure 2-1. Diagram illustrating the difference between internal and external surface area.Illustration is a hypothetical semifusinite maceral fragment 10Figure 2-2. Isotherms of the Brunauer, Demming, Demming, and Teller classification.Modified from Gregg and Sing (1982) 16Figure 2-3. Simulation ofmonolayer formation in a capillary 21Figure 2-4. Diagram depicting Polanyi’s equipotential lines 27Figure 2-5 Diagram illustrating variations in Dubinin-Radushkevich plots and correspondingpore size distributions 31Figure 2-6 Diagram showing the location of the adsorbed film and pore core in a cylindricalcapillary 36Figure 2-7 Revised de Boer hysteresis ioop classification showing the three most commonforms 40Figure 3-1 Gates coal petrography data on a mineral-free (a) and raw coal (b) basis 62Figure 3-2 Plots of carbon dioxide micropore capacities verses Gates coal composition on araw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d) 64Figure 3-3 Plots of carbon dioxide Langmuir monolayer volumes versus Gates coalcomposition on a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273K (c,d) 68Figure 3-4 Plots of carbon dioxide BET monolayer volumes versus Gates coal composition ona raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d) 69Figure 3-5 Plots ofmethane Langmuir monolayer volumes versus Gates coal composition ona raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d) 71Figure 3-6 Plots of Gates suite carbon dioxide D-R micropore capacities versus methanemonolayer capacities on a raw coal and mineral-free basis. D-R microporecapacities calculated from the 273 K isotherm (a) and 298 K isotherm (b) 72Figure 3-7 Alberta coal petrography data on a mineral-free (a) and raw coal (b) basis 74Figure 3-8 Plots of carbon dioxide D-R micropore capacities verses Alberta coal compositionon a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d).. .76Figure 3-9 Plots of carbon dioxide D-R micropore capacities verses Alberta coal compositionon a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d). . ..77ixFigure 3-10 Plots of carbon dioxide Langmuir monolayer volumes verses Alberta coalcomposition on a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273K(c,d) 78Figure 3-11 Plots of carbon dioxide BET monolayer volumes verses Alberta coal compositionon a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d). . ..79Figure 3-12 Plots of carbon dioxide Langmuir monolayer volumes verses Alberta coalcomposition on a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273K(c,d) 80Figure 3-13 Plots of carbon dioxide BET monolayer volumes verses Alberta coal compositionon a raw coal (b,d) and mineral-free (a,c) basis at 298 K (a,b) and 273 K (c,d). . ..81Figure 3-14 Dubinin-Radushkevich transformed isotherm plots for the Gates (a) and Alberta(b) suites. Calculations were made using the 273 K carbon dioxide isotherm 84Figure 3-15 Dubinin-Astakhov differential pore volume plots for the Gates (a) and Alberta (b)suites. Calculations were made using the 273 K carbon dioxide isotherm 85Figure 3-16 Plots ofAstakhov exponent (a) characteristic energy (b) and mean equivalent porediameter (c) versus Gates coal composition on a raw coal basis 88Figure 3-17 Plot of equilibrium moisture content versus Gates coal composition on a raw coalbasis 92Figure 3-18 Plots of carbon dioxide D-R micropore capacities versus Gates and Alberta coalcomposition, on a raw coal basis, calculated from the 298 K isotherm (a,b) and the273 K isotherm (c,d). Total vitrinite versus micropore capacity is plotted in b) andd); Structured vitrinite versus micropore capacity in a) and c) 94Figure 4-1 Diagram showing the location of the adsorbed film and pore core in a cylindricalcapillary 110Figure 4-2 Alberta coal petrography data on a mineral-free (a) and raw coal (b) basis 116Figure 4-3 Nitrogen isotherms obtained for samples *ACCC and ACCC-27 118Figure 4-4 Plots of 5-point BET surface areas versus total vitrinite (a,b) and structuredvitrinite (c,d) content. Mineral matter-free (a,c) and raw coal (b,d) values areplotted 122Figure 4-5 Plots of 5-point BET surface areas versus total inertinite (a,b) and semifusinite(c,d) content. Mineral matter-free (a,c) and raw coal (b,d) values are plotted ... 123Figure 4-6 Plot ofBJH cumulative surface area for pores between 2 and 50 nm diameterversus BET surface area 125Figure 4-7 Pore volume distribution curves for the a) adsorption branch and b) desorptionbranch of the isotherm. Samples are *ACCC and ACCC-27 127xFigure 4-8 Cumulative a) adsorption and b) desorption pore volume plots for samples *ACCCand ACCC-27 128Figure 4-9 Plots of cumulative adsorption mesopore volumes versus total vitrinite (a,b) andstructured vitrinite (c,d) content. Mineral matter-free (a,c) and raw coal (b,d)values are plotted 130Figure 4-10 Plots of cumulative adsorption mesopore volumes versus total inertinite (a,b) andsemifhsinite (c,d) content. Mineral matter-free (a,c) and raw coal (b,d) values areplotted 131Figure 4-11 Plot of gas yields from desorption canister testing versus mesopore volumes (rawcoalbasis) 134Figure 5-1 Diagram illustrating microstructures in coal 144Figure 5-2 Schematic diagram illustrating the Pressure-Decay Profile Permeameter (afterGeorgietal., 1993) 146Figure 5-3 Sample 1 showing top (a) and bottom (b) faces 151Figure 5-4 Permeability profiles for faces lA-i (a) and 1A-2 (b) 152Figure 5-5 Permeability profiles for faces lB-i (a) and 1B-2 (b) 153Figure 5-6 Sample 2 showing faces 2A (a) and 2B (b) 154Figure 5-7 Sample 3 showing faces 3A (a) and 3B (b) 156Figure 5-8 Permeability profiles of faces 4A (a) and 4B (b) 157Figure 5-9 Sample 5 showing faces 5A (a) and 5B (b) 158Figure 5-10 Sample 6 showing all faces on which points were measured 159Figure 5-11 Permeability profile of sample 7 160Figure 5-12 Plots of KI versus total vitrinite content of sample 1 on a; a) mineral matter-free,and ; b) raw coal basis 164xLIST OF TABLESTable 3-1 Results of proximate, sulphur, and equilibrium moisture analsyses 59Table 3-2 Lithotype classification, low-temperature ash, and x-ray diffraction results for theGates suite. Modified from Lamberson and Bustin (1993) 60Table 3-3 Gates and Alberta suite petrography data 61Table 3-4 Carbon dioxide Dubinin-Radushkevich (D-R) micropore capacities, Langmuir andBET monolayer volumes presented on a raw coal and mineral matter-free basis..65Table 3-5 Carbon dioxide Dubinin-Radushkevich (D-R) micropore capacities, Langmuir andBET equivalent surface areas presented on a raw coal and mineral matter-freebasis 66Table 3-6 High pressure methane monolayer volumes measured for the Gates suite 70Table 3-7 Astakhov exponents, characteristic energies, and mean equivalent pore diameterscalculated from the 273 K carbon dioxide isotherm 89Table 4-1 Results of proximate and sulphur analyses 113Table 4-2 Alberta suite petrography data 115Table 4-3 BET and Bill surface areas and mesopore volumes for the Alberta coals (raw coalbasis) 120Table 5-1 Lithotype classification used in the current study. Modified from Lamberson andBustin (1993) 150xiiAKNOWLEDGEMENTSThe current thesis was made possible through the help and contributions of severalindividuals. Firstly, I would like to thank my supervisor, Dr. R. Marc Bustin, for creating an idealenvironment in which to do research as well as for his many helpful and insightful discussions. Iwould also like to thank my supervisory committee, Dr. Roger Beckie, Dr. Rosemary Knight, andDr. John Ross for their support and helpful discussions. To my friends and colleagues, Dr.Michelle Lamberson and Dr. Maria Mastalerz, I extend a warm thank you for your help, support,and contributions thoughout the course ofmy study. I also deeply appreciate the support given tome by my family and friends. I would like to thank a very special person, Arlene Batuna, for herpatience and support.Technical support at UBC was provided by Ray Rodway, Bryon Cranston (thesis &hockey!), Marc Baker, Yvonne Douma, and Doug Polson; without you I would not havecompleted. Further technical support was provided by Core Laboratories of Calgary (JohnClow). I would also like to extend deep gratitude to Nancy Myra and the office staff ofGeological Sciences at UBC for keeping me “organize&’.Financial support was provided by the Natural Sciences and Engineering ResearchCouncil, Pan Canadian Ltd., the Geological Survey of Canada, and the Geological SciencesDepartment ofUBC.1CHAPTER 1INTRODUCTORY STATEMENTS1.1 iNTRODUCTIONCoalbed gas is currently being evaluated as a fuel source to supplement conventionalnatural gas reserves in several countries. The United States is the only country currentlyproducing coalbed gas commercially.Canada has an estimated 323 billion metric tons of coal resources (Kuuskraa and Boyer,1993) with estimates of coalbed gas resources varying from 1.42 x 1013 to 7.37 x i013 m3 (500-2600 Tel) (Schraufnagel, 1993). Much of the coal and hence coalbed gas resources is locatedwithin the Western Canadian Sedimentary Basin. The Alberta Geological Survey along withseveral Canadian gas and petroleum companies such as PetroCanada, Canadian Hunter, BP,Alberta Energy, Noreen, and Pan Canadian are currently in the process of evaluating coalbed gasas a natural gas supplement to current conventional gas reserves.Successful production of coalbed gas is dependent upon a complex interplay of geologicand economic factors. Among geologic factors affecting the ultimate recovery of coalbed gas are:coal seam thickness, continuity, geometry, and distribution; fracture permeability; rank; coal type;depth ofburial; gas saturation; and reservoir pressure and hydrologic conditions to name but afew. Controls of coal composition, particularly the organic fraction, upon the retention of gas hasonly briefly been investigated. The effects of macera! content of coal upon pore volumes and sizedistributions requires further investigation in order to completely understand the determinants ofgas content and producibility.This thesis investigates, through the use of the volumetric method ofmeasuring gassorption isotherms, the effects ofmacera! and mineral contents upon micropore and mesoporedistributions, capacities and associated surface areas. Further, the control of!ithotype andmaceral composition upon permeabilities, established through the use of a new permeameter2capable of obtaining permeability profiles on a bed (centimeter) scale, will be assessed. This laststudy investigates the effect ofmicro- and macrostructure upon coal permeability.1.2 NATURAL GAS GENERATION FROM COALA large quantity ofnatural gas is produced during the process of coalification, thebiochemical and thermal alteration ofplant material to peat, lignite, sub-bituminous, bituminous,semi-anthracite, anthracite, and meta-anthracite (Bustin et al., 1985). Natural gas produced fromcoalification, usually referred to as coalbed gas, is often rich in methane, but may also containsignificant amounts of other gases. Other, heavier hydrocarbons may also be produced duringcoalification. Composition of coalbed gas, in addition to associated products produced duringcoalification, is dependent upon the original organic matter type and the nature and degree ofbiochemical and thermogenic alteration. Further, migration of gases external to the system, suchas carbon dioxide derived from a magma source (Smith et al, 1985a; Kotarba, 1988, 1990) mayaffect the ultimate composition of coalbed gas.Two main types of coalbed gas exist: biogenic and thermogenic. Biogenic gas is primarilycomposed of methane and carbon dioxide and is formed through bacterial degradation of organicmatter (Kim and Douglas, 1972). Two main mechanisms exist for the formation of such gas:carbon dioxide reduction and methyl-type fermentation (Schoell, 1980; Woltemate et al., 1984;Jenden and Kaplan, 1986; Whiticar et al., 1986). Biogenic gas formation may occur at an earlyand late stage in the burial history of the coal (Rice, 1993).Thermogenic gas formation initiates at about the high-volatile bituminous stage (> 50°C)and continues throughout the geochemical stage of coalification (Hunt, 1979). Although methanegas is produced during the biochemical stage as a result of bacterial degradation of the originalvegetable matter, most of the gas produced is of thermogenic origin. The main gas componentsof the geochemical stage are methane, carbon dioxide, and water.The amount ofmethane produced during coalification is dependent upon coal compositionand assessments of gas produced will range depending on the estimation procedure. Estimatedvalues of total methane produced during. coalification range from 100 to 300 cm3/g (Juntgen and3Karweil, 1966; Juntgen and Klein, 1975; Hunt, 1979, Meissner, 1984; Welte et al., 1984; Levine,1987).1.3 GAS RETENTION IN COALCoal is unique in its ability to act as both a source rock and reservoir to natural gas. Thestorage capacity of coals varies with rank, pressure and temperature (Meissner, 1984). Much ofthe gas generated during coalification is lost to: a) surrounding sediments, possibly forming aconventional gas reservoir; b) the atmosphere; and c) groundwater flow through the coal seam.Some of the generated gas may be retained in the coal seam, depending upon the character of thecoal reservoir.Coalbed gas is retained in coal seams in the following ways: a) adsorption upon theinternal surfaces (i.e. in microporosity) or absorption within the molecular structure of the coal; b)as free gas, or gas in excess ofwhich can be adsorbed or absorbed, within cleats and fractures ofthe coal; and c) as a solute within groundwater present within the coal seam (Rightmire, 1984;Murray et al., 1991; Ertekin et al., 1991; Rice, 1993). By far the most important mechanism formethane retention is that of gas adsorption upon the internal surfaces of the coal, particularly withhigh rank coals. Hence the controls upon the micropore structure, and the pore structure ingeneral, of coals is hence of interest in determining the ultimate natural gas content of suchmaterial.The micropore system (pore diameters < 2 nm), which makes up the bulk of coalporosity at higher ranks, acts as a molecular sieve or as a clathrate cage (Van Krevelin, 1981).Gas retention within microporosity, and indeed its physical significance are a matter ofdebate. According to some workers (i.e. Dryden, 1963; Fuller, 1981; and Given, 1984)microporosity in coal may not be a fixed property of coal and is dependent upon the particularsorbate/coal system. Further, Levine (1993) states that sorption may be modeled as eitheradsorption (chemi- or physisorption) within the micropore network or as dissolution of sorbatewithin the molecular structure of the coal; sorption within coal is likely a combination of a variety4of different processes. The equations and sorption theories adhered to in this thesis are dependentupon physical adsorption taking place within coal porosity.In addition to microporosity, the pore structure of coal may be further broken down intothe following size classification as defined by the International Union ofPure and AppliedChemistry (JUPAC): mesoporosity, or pores with a diameter between 2 and 50 nm; andmacroporosity, or pores with a diameter greater than 50 nm. The physical mechanism ofgasadsorption appears to be dependent upon pore size. The dependence of these pore fillingmechanisms upon pore size will be discussed in Chapter 2.1.4 STRUCTURE OF THE THESISOne chapter is dedicated to the discussion of gas sorption theory and terminology andthree chapters are prepared as independent papers addressing the issues discussed above. Chapter3 investigates the effects of coal (maceral) composition upon the micropore capacity and sizedistributions and the implications for coalbed methane potential.Chapter 4 investigates the effects of coal (maceral) composition upon mesoporevolumes, size distributions and associated surface areas.Chapter 5 studies the variation of permeability with lithotype (megascopic) and maceral(microscopic) composition of coal.51.5 REFERENCESBustin, R.M., Cameron, A.R., Grieve, D.A., and Kalkreuth, W.D., 1985. Coal Petrology:Its Principles, Methods and Applications, Geological Association of Canada,Short Course Notes, Volume 3, Second Edition, 230 pp.Dryden, I.G.C., 1963. Chemical constitution and reaction of coal. In: H.H. Lowry (Editor)Chemistry of Coal Utilization Supplementary Volume. New York, John& Sons, pp. 232-295.Ertekin, T., Sung, W., and Bilgesu, H.I., 1991. Structural properties of coal that controlcoalbed methane production. In: D.C. Peters (Editor), Geology in Coal ResourceUtilization. pp 105-124.Fuller, E.L., Jr., 1981. Pf,’sical and chemical structure of coals: sorption studies. In: M.L.Gorbaty, and K. Ouchy (Editors), Coal Structure: Advances in Chemistry Series 192,Washington, D.C., American Chemical Society, p. 293-309.Given, P.H., P.H., 1984. An essay on the organic geochemistry of coal. In: M.L. Gorbaty et al.(Editors), Coal Science, v. 3. New York Academic Press, p. 63-252, 339-34 1.Gregg, S.J., and Sing, K.S.W., 1982. Adsorption, Surface Area and Porosity, Second Edition.Academic Press, New York. 303 pp.Hunt, J.M., 1979. Petroleum geochemistry and Geology: San Francisco, W.H. Freeman and Co.617p.Jenden, P.D., and Kaplan, JR., 1986. Comparison ofmicrobial gases of the Middle AmericaTrench and Scripps Submarine Canyon: implications for the origin of natural gas.Applied Geochemistry, 1: 631-646.Juntgen, H., and Karweil, J., 1966. Gasbildung and Gasspeicherung in Steinkohlenflozen, PartI and II. Erdol Kohie, Erdgas, Petrochem, 19: 25 1-258, 339-344.Juntgen, H., and Klein, J., 1975. Entstehung von Erdgas gus kohligen Sedementen. Erdol,Kohie, Erdgas, Petrochem, Erganzungsband, 1: 52-69.Kim, A.G., and Douglas, L.J., 1972. Hydrocarbon gases produced in a simulated environment.U.S. Bureau ofMines Report of Investigations 7690: iSp.Kotarba, M., 1990. Isotopic geochemistry and habitat of the natural gases from the UpperCarboniferous Zacler coal-bearing formation in Nowa Ruda coal district (LowerSilesia, Poland). In: B. Durand and F. Behar (Editors). Advances in OrganicGeochemistry, 1989. Oxford Pergamon Press, 1, pp.549-60.6Kotarba, M., 1988. Geochemical criteria for the origin of natural gases accumulated in theUpper Carboniferous coalseam-bearing fromations in Walbrzych Coal Basin (inPolish with English summary. Stanislaw Staszio Academy ofMining and MetallurgyScientific Bulletin 1199: ll9p.Kuuskra, V.A., and Boyer, C.M., II, 1993. Economic and parametric analysis of coalbedmethane. In: B.E. Law and D.D. Rice (Editors), Hydrocarbons from Coal,AAPG Studies in Geology # 38, pp. 373-394.Levine, J.R., 1993. Coalification: the evolution of coal as a source rock and reservoir rockfor oil and gas. In: B.E. Law and D.D. Rice (Editors), Hydrocarbons from Coal,AAPG Studies in Geology # 38, pp. 39-77.Levine, J.R., 1987. Influence of coal composition on the generation and retention of coalbednatural gas. Proceedings of the 1987 Coalbed Methane Symposium, pp. 15-17.Meissner, F.F., 1984. Cretaceous and lower Tertiary coals as sources for gas accumulationsin the Rocky Mountain area. Source rocks of the Rocky Mountain Region, 1984Guidebook, Rocky Mountain Association of Geologists, pp. 401-431.Murray, D.K., 1991. Coalbed methane: natural gas resources from coal seams. In D.C.Peters (Editor), Geology in Coal Resource Utilization, pp. 97-103.Rice, D.R., 1993. Composition and origins of coalbed gas. In: B.E. Law and D.D. Rice(Editors), Hydrocarbons from Coal, AAPG Studies in Geology # 38, pp. 159-184.Rightmire, C.T., 1984. Coalbed methane resource. In: C.T. Rightmire, G.E. Eddy, and J.N.Kirr (Editors), Coalbed methane resources of the United States. AmericanAssociation ofPetroleum Geologists, Studies in Geology, 17: 1-13.Schraufnagel, R.A., 1993. Coalbed methane production. In: B.E. Law and D.D. Rice (Editors),Hydrocarbons from Coal, AAPG Studies in Geology # 38, pp 341-359.Smith, J.W., Gould, K.W., Hart, G., Rigby, D., 1985a. Isotopic studies of Australian natural andcoal seam gas. Bulletin of Australasian Institute ofMining and Metallurgy, 290: 43-51.Van Krevelen, D.W., 1981. Coal (reprinted from 1961 edition). Amsterdam, Elsevier.514, pp.Welte, D.H., Schaefer, R.G., Stoessinger, W., and Radke, M., 1984. Gas generation andmigration in the Deep Basin ofwestern Canada. American Association ofPetroleum Geologists Memoir 38: 3 5-47.Whiticar, M.J., Faber, E., and Schoell, M., 1986. Biogenic methane formation in marine andfreshwater environments, CO2 reduction vs. acetate fermentation - isotopic evidence.Geochemica et Cosmochimica Acta, 50: 693-709.7Woltemate, I., Whiticar, M.J., Schoell, M., 1984. Carbon and hydrogen isotopic composition ofbacterial methane in a shallow freshwater lake. Limnology and Oceanography, 29:985-992.8CHAPTER 2GAS SORPTION THEORY2.1 INTRODUCTIONThis chapter outlines the basic principles of gas adsorption relevant to the determination ofthe internal surface area and pore size distributions of coal. Definitions of terms and descriptionsof concepts used in gas adsorption theory are given. In addition, the theories and equationsutilized in the current study to determine surface areas and pore size distributions based on gasadsorption are outlined; these include BET (Brunauer, Emmett, and Teller) Theory and associatedequations, Dubinin Theory ofVolume Filling for Micropores and associated equations, and BJH(Barrett, Joyner, and Halenda) Theory and associated equations. Finally, the choice of anadsorbate for coals is discussed.2.1.1 DefinitionsThe following list of terms and corresponding definitions is not meant to be exhaustive,but merely an introduction to the terminology applied in gas adsorption theory. These terms areones in general use in adsorption literature, and are not limited to a specific gas adsorption theory.Terms specific to each theory will be defined in a later section.1) Specific Surface Area:Specific surface area of a solid is defined as the surface area (internal and/or external) perunit mass of solid. The units used in the current study arem2/g.92) External vs. Internal Surface Area:External surface area of a solid containing surface irregularities is defined as that surfacearea including” all the prominences and all those cracks which are wider than they are deep”(Gregg and Sing, 1982, p. 23). The internal surface area of such a solid is thus defined as thesurface area which comprises “the walls of all cracks, pores and cavities which are deeper thanthey are wide” (Gregg and Sing, 1982, p. 24). The distinction between these two forms ofsurface area is arbitrary and forms the basis for the cut-off between inter- and intra-particleporosity. Figure 2-1 illustrates the difference between external and internal surface area using theexample of the coal maceral, semifusinite; the external surface area includes the outer surface ofthe maceral fragment, whereas the internal surface area comprises the inner walls of the pores.For many porous materials, including coal, the internal surface area far exceeds the externalsurface area of the material due to the area contribution of the pore walls and throats andmicrofractures in the sample.3) Porosity:Porosity of a solid refers to the ratio of the total pore volume of the solid to the solid’stotal volume. Porosity may be inter- or intra-particle porosity.4) Adsorption, Absorption, Sorption:These terms have often been used interchangeably in the literature and for the purposes ofthis thesis are defined in an unambiguous fashion as follows:10Internal S.A.Figure 2-1. Diagram illustrating the difference betweeninternal and external surface area. Illustrationis a hypothetical semifusinite macoral fragment.Cell cavityExternal S.A.11Adsorption: Gregg and Sing (1982, p. 2) define (physical) adsorption as “the enrichmentor depletion of one or more components in an interfacial layer’t. The distinctionbetween physical and chemical adsorption will be discussed in a later section.Adsorption is used in the current study as in Gregg and Sing (1982) to embrace thephysical uptake ofgas by either pore volume filling causing enhanced adsorption inmicroporosity or monolayer formation, both ofwhich involve surface adsorption, or bycapillary condensation.Absorption: This physical process refers to the actual incorporation or assimilation ofgas molecules into the solid’s molecular structure.Sorption: Sorption, as defined by Gregg and Sing (1982), is a general term which includes surfaceadsorption, absorption, and capillary condensation. Desorption is the opposite process.5) Adsorptive, Adsorbate, AdsorbentAdsorptive: This is a general term referring to any gas or vapour which is capable of beingadsorbed (Gregg and Sing, 1982).Adsorbate: This term is more specific and refers to the material that is physically or chemicallyadsorbed to the surface of the solid, such as a gas molecule occupying an adsorbedmonolayer. The adsorbate may have properties which differ from that of the adsorptive gas orbulk liquid.Adsorbent: This is the material upon whose surface adsorption takes place (e.g. coal).6) Adsorption Isotherm:12Adsorption isotherms are central to the discussion of adsorption theory, and as used in thisstudy, refer to plots of the amount (volume at stp, mass, number ofmoles) of vapour adsorbed(adsorbate) onto a solid (adsorbent) at a constant temperature, versus the relative pressure.Relative pressure is defined as the ratio of the equilibrium vapour pressure (P) to the saturationvapour pressure (P0) ofthe adsorbate gas and is used instead of equilibrium pressure if the gas isbelow its critical temperature. Brunauer, Deming, Deming and Teller (Brunauer et al., 1940)classified isotherms in terms of their fhnctional form. Isotherm types are dependent upon theparticular adsorbate-adsorbent system as well as the pore structure of the adsorbent.7) Sorption hysteresisSorption hysteresis refers to the non-coincidence of the adsorption and desorptionbranches of the isotherm curve.2.1.2 Concepts1) Forces ofAdsorptionA gas or vapour will be adsorbed to the surface of a solid through various mechanismsdepending upon the type of adsorption forces that govern the interaction between the adsorbateand the adsorbent. Dispersion and electrostatic forces are the most common forces governingadsorption (Gregg and Sing, 1982)Dispersion forces between atoms refer to those forces that arise from asymmetry of theelectron cloud of an atom over a short term (Fyfe, 1964; Gregg and Sing, 1982; Lowell andShields, 1984). An atom that is non-polar over a larger interval of time can be either polar ordipolar over a short period of time (Fyfe, 1964). If two atoms that exhibit dipolar behavior over ashort term are brought into proximity, the dipole moments may couple in phase and lead to asmall binding force. For example, helium has a spherically symmetric cloud consisting of two s13electrons in its electron shell. This spherical shape is the statistical average shape of the electroncloud described by a Schrodinger wave function (N’) (Fyfe, 1964). Over a very short period oftime, the average spherical symmetry of the electron cloud is not observed, but a transient dipolemoment is imparted to the helium atom (Lowell and Shields, 1984). The helium atom may inducea dipole moment in a neighbouring atom, leading to a net attraction.Dispersion forces are attractive in nature, but some repulsion is experienced due to theinter-penetration of the electron clouds of two atoms and the proximity of their nuclei. Theseforces are very small in magnitude relative to a typical covalent bond. The bonds created bydispersion forces are thus weak and easy to break.Electrostatic (coulombic) forces may also be important in determining adsorbateadsorbent interactions. Examples of such interactions are: polar solids with gas molecules thatpossess an induced dipole moment; polar solids with gas molecules which possess a permanentdipole moment; and polar solids with gas molecules possessing a quadrupolar moment (e.g. CO2)(Gregg and Sing, 1982; Lowell and Shields, 1984). Electrostatic forces are therefore highlydependent upon the nature of the adsorbate and adsorbent.2) Physical and Chemical AdsorptionFrom the discussion above, it is apparent that a variety of adsorbate-adsorbent interactionsare possible based on the nature of the forces involved. Two basic types of adsorption are defineddepending upon which of the two main groups of forces (dispersive or electrostatic) aredominant. Physical (or non-specific) adsorption occurs where dispersion and short term repulsiveforces predominate; chemical (or specific) adsorption occurs where electrostatic forcespredominate (Gregg and Sing, 1982; Lowell and Shields, 1984). Combinations of the two typesof adsorption occur, and Gregg and Sing (1982) give the range of possibilities based on the natureof the adsorbate and adsorbent. A continuum between chemical and physical adsorption probablyexists.14The two types of adsorption differ in several ways (Lowell and Shields, 1984):a) physical adsorption, due to the weak nature of dispersion forces, is reversible.b) physical adsorption is associated with a small heat of adsorption, whereas chemical adsorptioninvolves a larger heat of adsorption.c) chemical adsorption unlike physical adsorption involves true chemical bonding and has anassociated activation energy.d) the adsorbate is normally restricted to a single adsorbed layer in chemical adsorption, whereasin physical adsorption the adsorbate is less rigidly held to the surface and may form a number oflayers (multilayer adsorption).e) chemisorbed vapours are adsorbed to specific sites on the adsorbent surface, whereasphysisorbed adsorbates have a greater translational freedom.f) equilibrium is achieved more rapidly with physical adsorption than with chemical adsorption,except perhaps in the case ofmicropores where activated diflhsion processes may occur.Physical adsorption is thus desirable for surface area measurement due to the nonlocalized nature of adsorbate and hence greater surface coverage, as well as the lower equilibriumtimes and reversibility of the process.153) Heat ofAdsorptionThe potential energy of an adsorbate interacting with an adsorbent reaches a minimum atsome point close to the adsorbent surface (Gregg and Sing, 1982). This potential “well”represents the equilibrium or adsorbed position of the adsorbate.The process of adsorption is necessarily an exothermic one due to the loss of translationalfreedom of the adsorbate. The kinetic energy lost is converted to heat and the enthalpy change (AH) is necessarily negative. Heat of adsorption is related to this process with the exactthermodynamic derivation given in Gregg and Sing (1982). Heats of adsorption can bedetermined experimentally and are important in separating physical and chemical adsorption,distinguishing pore structures in which adsorption is enhanced (such as in micropores), and inmonitoring completion ofmonolayer formation, etc.4) Classification ofAdsorption IsothermsFive basic types of adsorption isotherms were described by Brunauer, Deming, Deming,and Teller (Brunauer et al., 1940) and are shown in Figure 2-2. Most adsorbate-adsorbentsystems yield isotherms that fall into this basic classification. Type I, II, III, IV, and V isothermsare described below.16Figure 2-2. Isotherms of the Brunauer, Demming,Demming and Teller classification.Modified from Gregg and Sing (1982).Am0UntadS0rbedRelative pressure17Type I isotherms, also referred to as Langmuir isotherms (Langmuir, 1916), are produced byadsorption onto microporous solids or by adsorbate-adsorbent systems in which adsorption isrestricted to a few monolayers. In physisorbed systems, the adsorbent must contain a very finepore structure with a small external surface to approximate the Type I isotherm shape (Figure 2-2). In such systems, enhanced uptake occurs at low relative pressures due to the overlapping ofadsorption potentials between pore walls of pores with diameters only slightly wider than theadsorbate gas molecule (Lowell and Shields, 1984). This effect is illustrated by the initial steepslope of the Type I isotherm. Adsorption fall off once the micropore system has been filled andlittle additional adsorption occurs until the system reaches its saturation point (P/P0=l).Hysteresis is normally absent from this type of isotherm (Orr, 1977). Type I isotherms are alsoproduced by chemisorbed systems where adsorption is necessarily restricted to a single monolayer(Lowell and Shields, 1984). The “plateau” section of the isotherm is then interpreted to representthe completion of a single monolayer (Orr, 1977).Type II isotherms, also referred to as sigmoid or S-shaped isotherms (Brunauer et al., 1940), areproduced by adsorption onto non-porous or macroporous solids. Adsorption is believed to occurthrough the formation of layers of adsorbed gas which are only one adsorbate molecule thick (amonolayer). The first point of inflection of the Type II isotherm is believed to be approximatelycoincident with the BET monolayer capacity (volume of adsorbate gas occupying a layer ofmolecular thickness) (Orr, 1977; Gregg and Sing, 1982; Lowell and Shields, 1984). At higherrelative pressures, multilayers are formed on the nonporous surface until saturation is achieved(Lowell and Shields, 1984). The Type II isotherm is described by classical BET Theory.Type ifi isotherms display an increase in adsorption with the total amount adsorbed due to agreater interaction of the adsorbate with the adsorbed layer than with the adsorbent. In such asystem the heat of adsorption is greater than the heat of liquefaction of the adsorbate (Lowell andShields, 1984).18Type IV isotherms are produced by adsorption onto mesoporous solids, i.e. solids with pores inthe 1.5 - 100 nm range. The initial portion of the isotherm is similar to the Type II isotherm, butenhanced adsorption occurs at higher relative pressures due to the onset of capillary condensation(discussed later). Type IV isotherms are also distinguished by the presence of a distinct hysteresisioop at higher relative pressures, which indicates non-coincidence of the adsorption anddesorption branches of the isotherm. Hysteresis is thought to occur subsequent to the completionof the first adsorbed monolayer (P/P0 0.3). As will be seen, the shape of the hysteresis loop ischaracteristic of the pore shape of the adsorbent.Type V isotherms result from weak adsorbate-adsorbent interactions and are rare.5) Pore Size ClassificationA pore size classification was defined at The International Union ofPure and AppliedChemistry (IUPAC) meeting in Washington, D.C. on July 23, 1971 (Orr, 1977). The meetingestablished the definition ofmicropores, mesopores (or transitional pores), and macropores asfollows:Micropores: pores with diameters of less than 2 nm.Mesopores: pores with diameters between 2 and 50 nmMacropores: pores with diameters greater than 50 nm.The pore size classification is arbitrary, but has a convenient application for manymaterials in the chemical industry. This classification is the one adhered to in the current study.Dubinin (1982) proposed a size classification for pores that is based on the linear sizes ofcarbonaceous adsorbents. This classification is as follows:19Micropores: pore (radii) less than 0.6 - 0.7 nm in size.Supermicropores: pores between 0.6 - 0.7 nm and 1.5 - 1.6 nm.Mesopores: pores between 1.5 - 1.6 nm and 100 - 200 nm.Macropores: pores greater than 100 - 200 nm in size.The Dubinin classification is such that the pore sizes correspond to the interpretedmechanism of pore filling for a carbonaceous adsorbent. For example, the microporeclassification is utilized for pores in which Dubinin’s Theory ofVolume Filling for microporesapplies (see later), and the mesopore range is coincident with multilayer formation and capillarycondensation (Dubinin, 1982). Supermicropores are ones in which “cooperative” effects occur(Gregg and Sing, 1982), and macropores are pores in which the capillary condensation mechanismcannot feasibly apply. Marsh (1987) warns, however, that “Close distinctions between the classesof porosity cannot be rigorous since they are based on adsorption behaviour, adsorbate withadsorbent, rather than a physical measurement”.2.2 THEORIES OF ADSORPTIONThe following section includes a description of the basic concepts underlying the theoriesof adsorption used in the current thesis. The main formulas used in each theory are given as wellas the range of applicability (in terms of relative pressure) and limitations of the theories.2.2.1 BET TheoryThe Brunauer, Emmett, and Teller (BET) Theory has enjoyed widespread use in the fieldof surface area measurement since its introduction in 1938. The theory is a modification ofLangrnuir’s kinetic model of adsorption (Gregg and Sing, 1982). The BET equation wasdeveloped to describe a Type II isotherm.20The BET equationThe BET Theory assumes that the surface of an adsorbent is simply an “array ofadsorption sites” (Gregg and Sing, 1982, p 42), where the most energetic sites are occupied firstas the pressure increases (Lowell and Shields, 1984). Physical adsorption is achieved through theformation of incomplete monolayers (Figure 2-3) stacked outward from the surface; the greaterthe number ofmonolayers formed, the greater the area of adsorbent surface covered. A dynamicequilibrium is thought to occur whereby the rate of evaporation from the first formed monolayeris equal to the rate of condensation upon the adsorbent surface (Brunauer et al., 1938). Theequation representing the state of equilibrium with the adsorbent surface for the first adsorbedlayer is (Brunauer et a!., 1938; Gregg and Sing, 1982; Lowell and Shields, 1984):Nm9lvl[elT] = A1icPO0where Nm is the number of adsorbate molecules occupying a completed monolayer, e1 is thefraction of surface sites occupied by the adsorbate, v1 represents the frequency of oscillation ofthe adsorbate molecule perpendicular to the adsorbent surface, E1 is an average adsorptionenergy for the first layer, A1 is the condensation coefficient, K is a constant derived from thekinetic theory of gases (Gregg and Sing, 1982; Lowell and Shields, 1984), P is the equilibriumadsorptive gas pressure, R is the Universal Gas Constant, and T is temperature.In the second and successive layers, the adsorption energy is assumed to be equal to theheat of liquefaction of the adsorbate, EL, and the constants v and A remain constant (Lowell andShields, 1984). The rate of condensation onto the first layer is assumed to be equal to the rate ofevaporation from the second layer (Brunauer et al., 1938) and the rate of condensation on the nthlayer is assumed to be equal to the rate of evaporation from the n+ 1 layer. Also, the number ofadsorbate layers at saturation is assumed to be infinite. After algebraic manipulation (Gregg andSing, 1982; Lowell and Shields, 1984), the following relation, the BET equation, is arrived at:B.•B.L•jMonolayer(adsorbate)AdsorptiveVapourMoleculesAdsorbentFigure2-3.Simulationofmonolayerformationinacapillary.Arrowsindicatedirectionofinteractionofadsorptivemolecules.Interactionisassumedtooccurintheverticaldirectiononly;arrowswithcrossthroughindicatenointeraction.22(1) P = 1 + C-i [PIP0][V(P0 - P)] VmC VmCwhere P0 is the saturation pressure of the adsorbate, V is the volume adsorbed at equilibrium, andVm is the volume of adsorbate occupying a monolayer (monolayer volume, or capacity). Aversion of equation (1) in which a finite number of adsorbed layers is assumed, has beendeveloped (Brunauer et al., 1938; Gregg and Sing, 1982; Lowell and Shields, 1984). Anapproximation of C is taken to be:(2) C=exp{[El-EL]/RT}where the terms are defined as above. The difference between E1 and EL is equal to the net heatof adsorption (Gregg and Sing, 1982).A plot of the left side of equation (1) versus relative pressure should yield a straight line.The values of C and Vm can be obtained from the slope (slope = [C11/VmC) and from theintercept (intercept = l/VmC). The monolayer capacity may then be converted to surface area ifthe adsorbate cross-sectional area is known (Lowell and Shields, 1984).In summary, the major assumptions made in the derivation of the BET equation are: 1) theenergy of adsorption is equal to the heat of liquefaction of the adsorbate for every layer but thefirst; 2) the conditions of dynamic equilibrium are the same for the second and higher layers; and3) at saturation, the number ofmultilayers is infinite, i.e., the adsorbate condenses to a bulk liquid(Gregg and Sing, 1982).23Significance of the BET C valueThe BET ‘C value is a parameter that may be used to predict the shape of the isothermand thus the nature of the adsorbate-adsorbent system. For example, for values of C greater thantwo, the isotherm described by the BET equation (plot ofVIVm vs. relative pressure), conformsto the shape of a Type II isotherm (Gregg and Sing, 1982). For large values of C, the knee of theisotherm becomes sharper (Figure 2.1, Gregg and Sing, 1982, p.46). If the value of C is less thanabout 20, it is thought that estimation of the monolayer capacity from either the BET equation orthe Point B method may be in error (Gregg and Sing, 1982). This is understandable, since thepoint of inflection of the Type II (and IV) isotherm is thought to be approximately coincident withthe completion of the first monolayer; if the point of inflection of the isotherm is not welldeveloped, the monolayer may not be complete at that point.An estimation of the relative affinity of an adsorbate for adsorption onto an adsorbent mayalso be obtained from the C value. The BET C value, as discussed above, is estimated by therelation: C = exp {[E1 - EL] / RT}. The C value will increase as the net heat of adsorptionincreases, or in other words, as the affinity of the adsorbate for adsorption upon the adsorbentsurface increases. For example, for a Type I and II composite isotherm obtained from amicroporous material described in Gregg and Sing (Figure 4.11, 1982), the initial part of theisotherm was steep due to enhanced adsorption. This is reflected in large values of C calculatedfrom the BET equation for the initial portion of the isotherm.Range ofApplicability of the BET EquationThe BET equation is useful for a variety of different isotherms, but application of theequation is generally limited to the relative pressures at which monolayer formation is believed tooccur. The range of relative pressures that correspond to nearly complete monolayers for Cvalues between 3 and 1000 is 0.05 < P/P0 <0.35 (Lowell and Shields, 1984). This range isapplicable to most experimental isotherms, and generally good agreement between experimentally24derived isotherms and the calculated BET isotherms has been achieved. Various examples ofdeparture of the BET linear plot (left side of equation (1) vs. relative pressure) from linearitybelow relative pressures between 0.2 and 0.3 do exist, however (Gregg and Sing, 1982). TheBET equation also fails to reproduce experimental isotherm data in the multilayer region (relativepressures> 0.3).Criticisms ofBET TheoryThe main assumptions of the BET equation were given earlier. Although theseassumptions simplify the BET treatment, they are the main source of criticism of the theory.Lateral interactions between adsorbate molecules are ignored in favour of the adsorbateadsorbent interactions in BET Theory. Although adsorbate-adsorbent interactions may benegligible far from the adsorbent surface, this is not so within the adsorbate monolayer.BET Theory assumes that surface adsorption sites are energetically identical. Due to theheterogeneous nature of many solid surfaces (e.g. coal), however, this is likely an erroneousassumption.The assumption that the heat (or energy) of adsorption in all layers but the first is equal tothe heat of liquefaction of the adsorbate may also be in error. As pointed out by Lowell andShields (1984), polarizing forces are likely to enhance adsorption potentials within at least the firstfew monolayers and not just the first layer.Finally, the BET Theory seems to be applicable mainly to a range of relative pressures andacisorbate-adsorbent systems in which monolayer formation occurs. For example, at low relativepressures in an adsorbate-adsorbent system in which the adsorbent is microporous, it is likely thatmonolayers do not form due to the enhanced potential between pore walls of pores ofmoleculardimensions.252.2.2 Type I Isotherm - Dubinin Theory of Volume FillingforMicroporesThe initially steep portion of the Type I isotherm, at low relative pressures, is due toenhanced adsorption within a pore with pore walls that are only a few adsorbate moleculediameters apart. Dubinin (1966) envisioned this process as being due to the overlapping ofadsorption potentials between the pore walls. Dubinin (1966) also concluded that this processcannot be adequately described by monolayer formation as described by the Langmuir and BETtheories. These conclusions were reached from the adsorption of vapours upon carbonaceousadsorbents, in particular activated carbon, at various stages ofburn-out. The differential heats ofadsorption were found to be considerably higher for the porous activated carbon than for the nonporous carbon black (Figure 3, Dubinin, 1966). The mechanism ofvolume filling was thusinvoked for pores of diameters less than about 2 nm.Gregg and Sing (Sing, 1982) referred to the process of volume filling as the primaryprocess of adsorption for slit-shaped micropores that are approximately .3 - .7 nm in width asdetermined from nitrogen at 77 K. They also stated that the degree of enhancement of theinteraction potential and thus enthalpy of adsorption is “dependent upon the nature of theadsorbate-adsorbent interaction and the polarizeability of the adsorbate” (Gregg and Sing, p.242).Gregg and Sing went onto define a secondary process for slightly wider slit-shaped pores (.7 - 1.8nm, obtained as before) in which cooperative effects enhance adsorption to a lesser degree thanthe primary process. At still larger pore diameters, the process of capillary condensation isbelieved to occur. The degree of enhancement of adsorption at low relative pressures is thus notstrictly a function of pore diameter, but of the ratio of pore diameter to the adsorbate moleculediameter.Dubinin (1975) states that the process of adsorption in micropores is thermodynamicallyanalogous to the process of solution. The adsorbate-adsorbent system may be treated as auniphase system in which no interface exists between the adsorbate molecule and the adsorbentsurface. The concept ofmicropore surface area is thus thought to be meaningless. The maincontrol upon gas adsorption in such pores is therefore pore volume, not micropore surface area.26Dubinin (1975) also stated that the main difference between the ‘theory of volume filling’ formicropores and multilayer adsorption in mesopores, in reference to binary mixtures of vapours,specifically, is that for microporous solids, the selectivity of one adsorbate over the other occursthroughout the entire micropore space whereas in mesoporous materials, the selectivity withrespect to the adsorbate is mainly restricted to the first adsorbed monolayer. The adsorptionforcefleld in micropores may be viewed as continuous throughout the system.Dubinin and various co-workers went on to formulate equations that described adsorptionin the low to medium pressure region of the isotherm starting from Polanyi’s potential theory ofadsorption, described in the next section.Polanyi’s Potential Theory of AdsorptionThe Polanyi Theory ofAdsorption is described in Dubinin (1975) and Lowell and Shields(1984) and is only briefly touched on here.Polanyi envisioned the surface of an adsorbent (Figure 2-4) as having an adsorptionpotential gradient that extended from the surface to a distance at which the equipotential line forthe adsorbate in question is equal to zero (Lowell and Shields, 1984). The adsorbate molecule isthus assumed to occupy a space, referred to as the adsorption volume, between the surface andthe zero equipotential line.A critical parameter is A, defined initially as the adsorption potential (Lowell and Shields,1984), but later referred to as the differential molar work of adsorption by Dubinin (1966). A isgiven by the expression:A = RT1n(P0/P)An important postulate is that the volume adsorbed at equilibrium relative pressure is dependentupon A. Plots of the adsorption volume vs. A are called “characteristic curves”..-.-.-.-.-.-.-..En=O..Figure2-4.DiagramdepictingPolanyi’sequipotentialplanes.PointsAandBrepresentasurfaceimpurityandirregularity,respectively.EsandEnareequipotentiallines.ModifiedfromLowellandShields(1984).—a28Dubinin-Radushkevich EquationSeveral postulates are important to the development of the Dubinin-Radushkevichequation (Dubinin, 1965,1966; Gregg and Sing, 1982).The first postulate is that the micropore volume is filled through volume filling of thepores, not by conventional multilayer adsorption as described by BET Theory. The parameter erepresents the degree of filling of the micropores and is equal to the ratio of the volume filled byadsorbate at equilibrium pressures to the limiting micropore volume (WIW0).Secondly, characteristic curves, or plots of e vs A, are invariant with temperature({dAIdT} = 0). Thus, it is assumed that the forces governing adsorption are van der Waalsforces, because these are temperature invariant (Marsh, 1987). This postulate is supported by theplotting of characteristic curves for various adsorbate-adsorbent systems at differenttemperatures. The curves for each adsorbent-adsorbate pair at different temperatures coincide(Figures 5 - 7, Dubinin, 1966).An important parameter, defined by Dubinin, is 13, the relative differential molar work ofadsorption or affinity coefficient of the characteristic curve. This parameter is an outcome of athird postulate given by Dubinin, where {A/A0}= 13. According to Dubinin (1965,1966), “atequal filled volumes of the adsorption space, W, the ratio of the differential molar work ofadsorption A of a given vapour to the differential molar work of adsorptionA0 of the vapourchosen as the standard is a constant value” (Dubinin, 1966, p.60). The standard adsorbate is oftenchosen to be benzene, with 13 1.A fourth proposition used in the derivation of the Dubinin-Radushkevich equation is thatthe distribution of pore sizes (or more correctly, the distribution of differential molar works ofadsorption (A)) is Gaussian (Dubinin, 1965, 1966; Gregg and Sing, 1982; Lowed and Shields,1984; Marsh, 1987). The equation thus assumes that the microporous carbonaceous adsorbent ishomogeneous (Dublin, 1982) and that the pore size distribution is narrow and does not include29supermicroporosity (Marsh, 1987). Equations to describe wider pore size distributions have beendeveloped (Dubinin, 1982; Rozwadowski and Wojsz, 1984).Using all of the above propositions, the following equation was formulated:WIW0=9 = exp[ - k(A/,13)2]Substituting the equation for A, this becomes (Dubinin, 1966; Gregg and Sing, 1982):(3) W/W0= 8 = exp[ -k/132(RT1nP0IP)]orW/W = 9 = exp[-B(T/f3)21og(P0/P)jwhere: B = 2.303R/k; k is a structural parameter related to pore size (energy) distribution of theadsorbent.Equation (3) may also be written in the following form for plotting purposes:logW = log W0 -B(T/j3)2[log(PIP)]According to Gregg and Sing (1982), W is equal toJp*,where p is the adsorbate density. Ifthe temperature ofmeasurement is well below the critical temperature of the adsorbate, p*maybe taken as the density of the liquid adsorptive. The limiting volume of the adsorption space maybe obtained from equation (3). A monolayer capacity (volume) and monolayer equivalent surfacearea may also be calculated. The value ofW is also referred to as the micropore volume which isobtained from the micropore capacity, the amount of vapour adsorbed into the micropores. Thecalculated micropore volume may be in error if the effect of the proximity ofmicropore wallsupon the degree of packing of the adsorbate is not taken into account (Sing, 1989). The validityof a micropore surface area has also been questioned by some authors (Marsh, 1987) due to its30dependence upon method ofmeasurement. Micropore surface area must then be referred to asthe equivalent surface area.A plot of logW versus log2(P°IP) should yield a straight line if the theory ofvolume fillingofmicropores is obeyed. The intercept will giveW0, the limiting volume of the adsorption space,and the slope will yield the ratio B/132. The gradient of the Dubinin plot is thought to be related tothe average pore size and width of the Gaussian distribution (Marsh, 1987); the width of thedistribution is given by the parameter k (Gregg and Sing, 1982). Dubinin (1966) found that alinear fit was applicable for a range of adsorbate-adsorbent systems, and that for a particularadsorbate system, the value ofW0 should remain constant for a variety of adsorbates. As Dubininexplains (1966), this fact is not an outcome of the Gurvich rule, since the original rule wasformulated for non-microporous adsorbents whose pores filled through capillary condensation.Marsh (1987) illustrates several examples of non-linearity of the transformed Dubinin plot,and discusses the effect upon obtained pore size distributions. Figure 2-5 shows these deviationsand gives explanations for them.Dubinin-Astakhov EquationIn an attempt to rectify the problem of non-linearity of the transformed Dubinin plot foradsorbents with a broad pore size distribution, Dubinin and Astakhov (Dubinin and Astakhov,1971; Dubinin, 1975) introduced the Dubinin Astakhov equation, a generalized version of theDubinin-Radushkevich equation:(4) W/W0= 0 exp [-(RT/E)’ ln(P0fP)]31A)LogW Frequency- micropores filled continuoslyB)- larger micropores absent orentire micropore system filledlow relative pressures.C)enhanced adsorption causedby filling of a) supermicropores,b) mesopores, or by multilayerformation.D)- activated diffusion ormolecular sieve effects.Figure 2-5. Diagram illustrating variations in DubininRadushkevich plots and corresponding poresize distributions. Explanations are obtainedfrom Marsh (1987). Modified from Marsh (1987).E)JfrThN\Log2(p°/p)- system not at equilibriumPore diameter32where E is the characteristic free energy of adsorption and is equal to the differential molar workof adsorption for e .368 (Gregg and Sing, 1982). This value is thought to be an inversefunction of the average micropore size (Stoeckli et a!., 1989) and has been related to the half-width of slit-shaped micropores using the radius of gyration obtained from small-angle scatteringx-ray techniques (Dubinin, 1982).The pore size distribution, or more correctly, the energy distribution, in the case of theDubinin-Astakhov equation is assumed to be Weibull (Dubinin, 1975; Greg and Sing, 1982), notGaussian. The value of n in the Dubinin-Astakhov equation is optimized between the values of 1and 4 to obtain a best fit to the linear regression obtained for the transformed Dubinin plot (plotof logW vs logrl(Po/P)). For the case of n = 2, the Dubinin-Astakhov equation assumes the formof the Dubinin-Radushkevich equation. The value of n can thus give an indication of the nature ofthe pore size distribution. Stoeckli et al. (1989) state that as n varies from 3 to 1.5, theheterogeneity of the micropore size distribution increases; n = 3 for truly homogeneous molecularsieve activated carbons, which is contrary to the assumption made by Dubinin (1966) that n = 2for a homogeneous micropore system.Range ofApplication of the Dubinin EquationsThe Dubinin-Radushkevich equation is generally thought to be valid over the relativepressure range of about i0 <P/P0 <0.1 (Rozwadowski and Wojsz, 1984). This range isconvenient if one uses carbon dioxide as an adsorptive, since the carbon dioxide has a saturationpressure at 298 K of—S 48,200 mmHg. The high saturation pressure allows measurements to betaken below one atmosphere (760 mmHg). Most other methods of isotherm interpretation arenot valid at relative pressures below 0.02.The Dubinin-Astakhov equation has been shown (Dubinin, 1975) to have a lowerboundary of application at pore fillings (9) of about 0.15 - 0.2. The second Dubinin postulate({dA/dt} = 0) is not obeyed for lower values of filling.33Criticisms ofDubinin TheorySing (1989) states that” at present, there is no reliable procedure available for thecomputation of the micropore size distribution from a single isotherm”. This statement stemsfrom the fact that there is no strict mathematical description of the adsorption process inmicropores that takes into account the variability of all adsorbate-adsorbent systems.The Dubinin equations attempt to determine a micropore size distribution based on thedistribution of adsorption potentials which is assumed to obey a standard distribution type(Gaussian, Rayleigh, or Weibull). As Marsh (1987) states “the fact that so many adsorptionisotherms can be linearized in Dubinin-Radushkevich coordinates (whereas random curvesresembling isotherms cannot be linearized) is telling us that some property ofmicroporosity isbeing exhibited”. This is contrary to some criticisms (Sutherland, 1967) that accuse the Dubininequations of linearizing random curves (Freeman et al., 1970). The Dubinin-Radushkevichtransformed isotherm plot is not linear for many adsorbate-adsorbent systems, which is in directcontradiction to the accusations made by Sutherland (1967). The fact that the Dubinin plots arenot linear for many systems leads to the following criticism, however.Deviation of the Dubinin-Radushkevich plot from linearity in some systems may resultfrom several reasons. Some degree of heterogeneity in the micropore system may cause thedeviation, in which case the Dubinin-Astakhov equation may provide a better model for thesystem. Other modifications of the Dubinin-Radushkevich equation have also been developed toaccount for micropore heterogeneity (Dubinin, 1966; Dubinin and Stoeckli, 1980). Deviationfrom linearity in the Dubinin-Radushkevich plot can also be caused by chemical adsorption whichmay occur in addition to physical adsorption. Since chemisorption is temperature dependent, theamount of adsorbate uptake is also dependent upon temperature (Marsh, 1987). The gradient ofDubinin-Radushkevich (D-R) plots appears to be dependent upon temperature in that the gradientdecreases with increasing temperature for polar adsorbates. Thus, if chemisorption is occurring, itcan be predicted through the use of the D-R plots, and therefore the results may be examined34critically. One additional cause ofD-R plot deviation at low relative pressures is activateddiffusion or molecular sieve effects (Figure 2-5).Although the semi-empirical Dubinin equations are not able to model all adsorbateadsorbent systems accurately, their value comes in their ability to predict the nature of themicropore size distributions and adsorbate-adsorbent interactions.2.2.3 Type IVlsotherm - BJH TheoryThe Type IV isotherm is unique in that at relative pressures of—0.42 and above (fornitrogen as an adsorbate at 77 K), a hysteresis loop is encountered (Gregg and Sing, 1982).Mesoporous solids typically yield a Type IV isotherm. Significant enhancement of adsorptionmay occur at relative pressures above and below the point of closure of the hysteresis ioop for amesoporous solid as compared to the equivalent non-porous solid. This is thought to be due tothe occurrence of capillary condensation within the mesopores. The concept of capillarycondensation and the Kelvin equation are key to the BJ}I (Barrett, Joyner, Halenda) Theory aswell as many other theories describing the Type IV isotherm. Before BJH Theory is discussed,however, the Kelvin and Halsey equations are examined.The Kelvin EquationThe basic form of the Kelvin equation used in examination of the Type IV isotherm is(Gregg and Sing, 1982):lnPfP0 (-2yV )/(RTrm)where P/P0 is the relative pressure, ‘y is the surface tension of the liquid adsorptive, VL is themolar volume of the liquid adsorptive, and rm is the mean radius of curvature of the meniscusbetween the liquid adsorptive and its vapour at equilibrium.35The Kelvin equation, when applied to cylindrical pores, takes the form:lnP/P0 = (-2yVL) cos9(rk RT)where r (Figure 2-6) is the Kelvin pore radius (or core radius), and 8 is the angle of contactbetween the capillary condensate and an adsorbed film on the pore wall, which is often assumedto be equal to zero (Sato, 1981; Gregg and Sing, 1982; Lowell and Shields, 1984). Theassumption that the contact angle is equal to zero has been questioned (Gregg and Sing, 1982)but is widely used.The Kelvin equation was derived on thermodynamic grounds, and accounts for themechanical and physicochemical equilibrium established between a liquid and its vapour across ameniscus at a particular relative pressure. The equation states that at equilibrium pressures lessthan the saturation pressure, the vapour may be in equilibrium with the condensed liquidadsorbate, depending on the radius of curvature of the (concave) meniscus. At saturation vapourpressure, the radius is infinite, and the equation describes the equilibrium between the vapour andthe bulk liquid across a planar surface. The Kelvin equation therefore gives the radius of the coreof a pore in which capillary condensation occurs at a given P/P0.Pore size distributions are obtained by relating the curvature of the liquid/vapour interfaceto the radius of the pore. Assuming that the pore shape is cylindrical, and that the angle ofcontact between the capillary condensate and the adsorbed film is zero, the Kelvin radius is takenas being equal to the radius of the pore core (Figure 2-6). The Kelvin radius is thus not equal tothe radius of the pore itself but of the pore core since an adsorbed film already exists on the porewalls at the given relative pressure. The pore radius is then given by:= k + t36A)B)Figure 2-6. Diagram A) shows the location of the adsorbed film andpore core in a cylindrical capillary; B) illustrates thedifference between the Kelvin (rk) and pore (rp) radii.37where rp is the radius of the pore, r is the Kelvin radius, and t is the thickness of the adsorbedfilm. The thickness of the adsorbed film may be calculated by various methods, but only theHalsey equation will be discussed here.The Halsey EquationThe Halsey equation is based on the assumption that the thickness of an adsorbed layer ona planar surface is the same as that on the the internal surface of a pore (Sato, 1981; Lowell andShields, 1984). The thickness of the absorbed film is given by the following expression:t = W/W X twhere t is the thickness of the adsorbed film, Wa is the amount adsorbed at the given relativepressure, Wm is the amount adsorbed in a layer of adsorbate molecular thickness (BET monolayercapacity), and t is the thickness of the monolayer. The thickness of the monolayer may beobtained from:= VjjSwhere VL is the adsorbate molar liquid volume, and S is the surface area occupied by spreading amole of liquid adsorbate over a planar surface to a thickness of one adsorbate molecule. Fornitrogen, the monolayer thickness is .354 nm. A plot ofWa/Wm versus the relative pressureyields a Type II isotherm which may be described by the Halsey equation in the form of (Lowelland Shields, 1984):t .354 x [5/ln(P0JP)]11338BJH TheoryThe BiB Theory (Barrett et al., 1951) is based upon the Wheeler equation which may beexpressed as:V0 -V=irj(r-t)2L(r)drwhere the integration is carried from1Pn’ the radius of largest pore filled at a given pressure, toinfinity; V0 is the volume of adsorbate adsorbed at saturation vapour pressure; V is the volumeadsorbed at the equilibrium pressure; L(r) is the length of pores with radii lying between r and r +dr; t is the multilayer thickness at equilibrium pressure.BJH Theory does not make the assumption that the pore size distribution has a definiteshape (Gaussian or Maxwell) or that the adsorbed layer does not change thickness, as assumed inearlier theories. The theory does, however, make two fundamental assumptions: the pores arecylindrical in shape, and the two mechanisms of capillary condensation and multilayer formationlead to pore filling.The step by step description of how BJH calculates pore size distributions, volumes andsurface areas are discussed by Barrett et al. (1951) and Gregg and Sing (1982).Range ofApplicability ofBJH TheoryThe range of applicability of the BJH method for predicting pore size distributions isintimately related to the range of applicability of the Kelvin equation. Gregg and Sing (1982)discuss the various different controls upon the range of the Kelvin equation including thecurvature and the tensile strength effects. Gregg and Sing also state that although a theoreticallimit for the upper range of the Kelvin equation does not exist (if 9 <90°), a practical limit doesoccur. In using nitrogen as an adsorbate at T = 77 K, uncertainty in temperature measurements atrelative pressures close to unity may lead to large errors in the calculation of rm. Barrett et al.39(1951) impose a practical upper limit of about 30 nm for pore radius measurements. The lowerlimit imposed for the Kelvin equation stems from the uncertainties involving adsorbate molarvolumes and surface tensions in very fine pores and is usually set at about 1 - 1.5 nm (Lowell andShields, 1984). Kadlec (1989) has proposed a more generalized version of the Kelvin equation.Type IV Isotherm HysteresisAs mentioned earlier, the desorption and adsorption branches at relative pressures> 0.3are not coincident for the Type IV isotherm, and therefore the process of adsorption-desorption isnot reversible. Pore shape is interpreted as being the cause of hysteresis in mesopores asevaporation and condensation occur in different portions of the pore at the same relative pressure(Gregg and Sing, 1982). The fact that two relative pressures occur for the same degree of uptakeis cause for concern in trying to obtain pore size distributions through the use of the Kelvinequation. The desorption branch of the isotherm is often chosen for pore size distribution analysison thermodynamic grounds (Lowell and Shields, 1984). There are exceptions to this rule,however (see Gregg and Sing, 1982).Five types of hysteresis loops are given by de Boer (1958) based on various different poreshapes. Three of the common hysteresis loops are given in Figure 2-7. Type A hysterisis isgenerally associated with agglomerates with narrow pore size distributions (Sing et al., 1985);Type B hysteresis is caused by slit shaped pores; and Type E hysteresis is caused by “bottleneck”pores. Gregg and Sing (1982) analyze the various different pore shape models and resultinghystersis by utilizing the Kelvin equation.40Figure 2-7. Revised de Boer hysteresis loop classification showingthe three most common forms. Also shown are thecorresponding pore shapes. Possible low pressurehysteresis is indicated with dotted lines. Modified fromGregg and Sing (1982).Am0UntadS0rbedRelative pressure412.3 CHOICE OF ADSOREATEAs can be inferred from the discussion of the forces governing adsorption, the choice ofadsorbate in a particular adsorbate-adsorbent system is critical in determining the type ofinteraction during adsorption. Indeed, if the interaction between the adsorbent-adsorbate pair isnot governed strictly by dispersion forces (i.e. van der Waals forces), such as would be the case ifthe adsorbate possessed a permanent dipole moment and the adsorbent were polar, then the aboveequations (BET, Dubinin etc.), which are based on physical adsorption, are not valid. Thefollowing discussion will address this issue as well as others in choosing the proper adsorptive forthe meso-microporosity of coals.Nitrogen gas at analysis temperatures of 77 K is a popular choice for determining BETsurface areas and pore size distributions ofmesoporous solids. The properties of nitrogen gasthat make it an effective adsorbate are: 1) small enough BET C value to preclude localizedadsorption, but not too small to be excessively mobile at the adsorbent surface (Lowell andShields, 1984); 2) the saturation pressure of the gas is sufficiently large that a wide range ofrelative pressures may be obtained accurately (Gregg and Sing, 1982); 3) nitrogen gas is inert; 4)the cross-sectional area of the molecule is well established from liquid density calculations (0.162rn2) and is relatively small; and 4), the analysis bath temperature of 77 K is easily achieved withliquid nitrogen.Some problems are encountered with the use of nitrogen as an adsorbate for microporoussolids such as coals. Nitrogen has been shown to exhibit a positive temperature dependence withrespect to uptake (Mahajan, 1991). This is thought to be due to activated diffusion effects inwhich a significant activation energy for diffusion must be overcome by the nitrogen moleculebefore entry into fine pores is allowed. The activation energy for diffusion of nitrogen has beenshown to be significantly larger than that for carbon dioxide, despite the small difference in theiraverage diameters (0.365 nm for nitrogen, 0.33 nm for carbon dioxide; Mahajan, 1991). Rao(1991) gives the example that the energy barrier for carbon dioxide for entry into a pore 0.542 nm42in diameter is zero, whereas for nitrogen the energy barrier is 24.3 kJ/mol. Thus, microporousmaterials such as coal display molecular sieving characteristics towards the nitrogen molecule.Due to the problem of activated entry of nitrogen into micropores, the carbon dioxidemolecule was suggested as an adsorbate to be used for microporous materials (Walker and Kini,1965; Mahajan, 1991). At the temperatures commonly used in the measurement of surface areawith carbon dioxide as an adsorbate gas (273 - 298 K), the carbon dioxide molecule does notappear to display activated diffusion effects, and therefore is adsorbed more readily than thenitrogen molecule at 77 K (Marsh, 1987). Mahajan (1991, p. 736) states: “adsorption ofCO2 at25°C should invariably measure essentially the total surface area of coals” and that at the sametemperature “essentially the entire pore volume of all the coals studied would be filled with C02”.Use of carbon dioxide as an adsorbate for microporous adsorbents has been criticized forseveral reasons, however. Firstly, the carbon dioxide molecule is known to possess a quadrupolarmoment, and may interact with the hydroxyl- and oxygen- containing functional groups at the coalsurface to form chemical bonds. Not all authors are in agreement with this view, however(Mahajan, 1991). Secondly, the carbon dioxide molecule is thought to induce swelling behaviourin coal due to imbibition, and therefore lead to artificially high surface areas. Mahajan (1991, p.740) states, however, that: “CO2 gives higher surface areas compared to other adsorbatesbecause, through imbibition, it is available to both open and closed porosity, and at CO2 pressure<760 mmHg and short ‘equilibrium’ times the contribution of swelling to total surface area issmall”. The reporting of carbon dioxide and surface areas are suspect, however, because thecross-sectional area of the molecule is not well established in micropores.In summary, for the current study, nitrogen is used as an adsorbate gas for thedetermination ofmesopore size distribution and surface area analyses, due to its universalacceptance for this purpose, and carbon dioxide is used for micropore size distributions andmonolayer capacities due to lack of activated diffusion characteristics. In the results of Chapter 3,the potential problems associated with using carbon dioxide as an adsorbate are considered.432.4 CONCLUSIONSIt is obvious from the above discussion that the process of adsorption of vapour onto theinternal porosity of adsorbents, particularly microporous adsorbents, is a complex process. Thereis no universal mathematical treatment that adequately describes all aspects of the adsorptionprocess for all adsorbate-adsorbent systems under all conditions, nor for microporous systems, isit possible to observe the process. The current theories of adsorption for meso- andmicroporosity make many assumptions, often tacit, about the process of adsorption, and theircritical application is recommended. The TUPAC has published a guide to reporting adsorptiondata that attempts to minimize the ambiguity that is often found in the literature (Sing et al.,1985). Indeed, pore size distributions and surface area results obtained are a function of thetheories of adsorption and experimental procedure used in their calculation.442.5 REFERENCESBarrett, E.P., Joyner, L.G., and Halenda, P.P., 1951. The determination of pore volume and areadistributions in porous substances. I. Computations from nitrogen isotherms. TheJournal of the American Chemical Society, 73: 373-380.Brunauer, S., Deming, L.S., Deming, W.S., and Teller, E., 1940. On a theory of van der Waalsadsorption of gases. The Journal of the American Chemical Society, 62: 1723-1732.Brunauer, S., and Emmett, P.H., 1937. The use of low temperature van der Waals adsorptionisotherms in determining surface areas of various adsorbents. The Journal of theAmerican Chemical Society, 59: 2682-2689.Brunauer, S., Emmett, P.H., and Teller, E., 1938. Adsorption of gases in multimolecular layers.The Journal of the American Chemical Society, 60: 309-319.de Boer, J.H., 1958. In: D.H. Everett and F.S. Stone (Editors), The Structure and Properties ofPorous Materials. Butterworth, England, p. 68.Dubinin, M.M., 1965. Theory of the bulk saturation ofmicroporous activated charcoals duringadsorption of gases and vapours. Russion Journal ofPhysical Chemistry, 39: 697-704.Dubinin, M.M., 1966. Porous structure and adsorption properties of active carbons. In: P.L.Walker, Jr. (Editor), Chemistry and Physics of Carbon, Volume 2. Edward Arnold, Ltd.,New York. pp. 51-120.Dubinin, M.M., 1975. Physical adsorption of gases and vapors in micropores. In:D.A.Cadenhead, J.F. Danielli, and M.D. Rosenberg (Editors), Progress in surface andmembrane science, Volume 9. Academic press, New York. pp. 1-70.Dubinin, M.M., 1982. Microporous structures of carbonaceous adsorbents. Carbon, 20: 195-200.Dubinin, M.M., and Astakhov, V.A., 1971. Description of adsorption equilibria of vapors onzeolites over wide ranges of temperature and pressure. Advances in Chemistry Series,No. 102: p102.Dubinin, M.M., and Stoeckli, H.F., 1980. Homogeneous and heterogeneous micropore structuresin carbonaceous adsorbents. Journal of Colloid and Interface Science, 75: pp. 34-42.Emmett, P.H. and Brunauer, 5., 1937. The use of low temperature van derWaals adsorptionisotherms in determining the surface area of iron synthetic ammonia catalysts. TheJournal of the American Chemical Society, 59: 1553-1564.Freeman, E.M., Siemieniewska, T., Marsh, H., and Rand, B., 1970. A critique and experimentalobservations of the applicability to microporosity of the Dubinin equation ofadsorption. Carbon, 8: 7-17.45Fyfe, W.S., 1964. Geochemistry of solids. Mcgraw-Hill Book Company, New York, 199 pp.Gregg, S.J., and Sing, K.S.W., 1982. Adsorption, Surface Area and Porosity, Second Edition.Academic Press, New York. 303 pp.Jaroniec, M., Madey, R., Choma, J., McEnaney, B., and Mays, T.J., 1989. Comparison ofadsorption methods for characterizing the microporosity of activated carbons. Carbon,27: 77-83.Kadlec, 0., 1989. On the theory of capillary condensation and mercury intrusion in determiningcarbon porosity. Carbon, 27: 14 1-155.Langmuir, I., 1916. The constitution and fundamental properties of solids and liquids. TheJournal of the American Chemical Society, 38: 2221-2295.Lowell, S., and Shields, J.E., 1984. Powder Surface Area and Porosity, Second Edition.Chapman and Hall, London, 234 pp.Mahajan, O.P., 1991. CO2 surface area of coals: the 25 year paradox. Carbon, 29: 735-742.Mahajan, O.P., and Walker, P.L., Jr. 1978. Porosity of coal and coal products. In: C. Karr, Jr(Editor), Analytical Methods for Coal and Coal Products, Volume I. Academic Press,New York. pp. 125-162.Marsh, H., 1987. Adsorption methods to study microporosity in coals and carbon - a critique.Carbon, 25: 49-58.McEnaney, B., 1987. Estimation of the dimensions ofmicropores in active carbons using theDubinin-Radushkevich equation. Carbon, 25: 69-75.Orr, C., 1977. Pore size and volume measurement, In: I.M. Kolthofl P.J. Elving, and F.H.Stross (Editors), Treatise on Analytical Chemistry Part III, Volume 4. John Wiley andSons, New York. pp. 3 59-402.Orr, C., 1977. Surface area measurement. In: I.M. Kolthofl P. J. Elving, and F.H. Stross(Editors), Treatise on Analytical Chemistry Part III, Volume 4, John Wiley and Sons,New York. pp. 321-358.Rao, M.B., 1991. Diffusion through carbon micropores - 4 years later. Carbon, 29: 813-815.Rodriguez-Reinoso, F., Garrido, J., Martin-Martinez, Molina-Sabio, M., and Torregrosa, R.,1989. The combined use of different approaches in the characterization of microporouscarbons. Carbon, 27: 23-32.Rozwadowski, M., and Wojsz, R., 1984. An attempt at determination of the structuralheterogeneity ofmicroporous adsorbents. Carbon, 22: 363-367.46Sato, T., 1981. Methane recovery from coal beds: surface and physical properties ofwesternUnited States coals; M. Sci thesis, The University ofNew Mexico, 78 pp., unpublished.Sing, K.S.W., D.H. Everett, Haul, R.A.W., Moscou, L., Pierotti, R.A., Rouquerol, J., andSiemienewska, T., 1985. Reporting physisorption data for gas/solid systems with specialreference to the determination of surface area and porosity. Pure and Applied Chemistry,57: 603-619.Sing, K.S.W., 1989. The use of physisorption for the characterization ofmicroporous carbons.Carbon, 27: 5-11.Sutherland, J.W., 1967. In: R.L. Bond (Editor), Porous Carbon Solids. Academic Press, NewYork. p. 1.Stoeckli, H.F., Kraehenbuehl, Ballerini, L., and De Bernardini, S., 1989. Recent developments inthe Dubinin equation. Carbon, 27: 125-128.Wojsz, R., and Rozwadowski, M., 1989. The micropore structure analysis of active carbons.Carbon, 27: 135-139.47CHAPTER 3VARIATION IN MICROPORE CAPACITY AND SIZE DISTRIBUTION WITHCOMPOSITION IN HIGH AND MEDIUM-VOLATILE BITUMINOUS COAL OF THEWESTERN CANADIAN SEDIMENTARY BASIN: IMPLICATIONS FOR COALBEDMETHANE POTENTIAL.3.1 ABSTRACTThe effect of lithotype, maceral composition and mineral content upon the microporecapacity and size distribution is investigated for a medium-volatile bituminous coal from the midCretaceous Gates Formation of northeast British Columbia and a high-volatile bituminous coalfrom the Cretaceous ofAlberta. Vitrinite content ranges from 18 to 95 % (volume %, mineralmatter-free) for the Gates coal and 36 to 85 % for the Alberta coal. Ash yields vary from 4.4 to33.7 (weight %) for the Gates coal and 1.2 to 10.6 % for the Alberta coal. Dubinin-Radushkevichcarbon dioxide micropore capacities, measured at 298 K, range from 21.7 to 39.8 cm3/g (mineralmatter-free) for the Gates coals and 34.1 to 49.7 cm3/g for the Alberta coal. Low -pressureDubinin micropore capacities, Langmuir and BET monolayer volumes, measured at 298 and 273K, generally increase with total and structured vitrinite content and, conversely, decrease withinertinite and mineral matter content. A similar trend is found for high-pressure Langmuirmethane monolayer capacities determined for the Gates coals; the methane monolayer capacitiesare smaller but correlatable with the carbon dioxide micropore capacities. The increase inmicropore capacity with vitrinite content is due to an increase in the number of micropores, asdemonstrated by Dubinin-Astakhov micropore size distributions. For the Gates suite, a samplewith both a high total vitrinite content and semifusinite content yielded the largest microporecapacity which may be due to the creation of micropore capacity through burning (charring)during semifusinite formation. Micropore heterogeneity and mean pore size increase with anincrease in inertinite and mineral matter content. Coal composition thus appears to be an48important control upon the micropore capacity, size distribution, and hence, the gas content ofbituminous coals.3.2 INTRODUCTIONLarge quantities ofnatural gas, often mainly methane, are produced during thebiochemical and geochemical or thermogenic stage of coalification. Much of this gas is lost tosurrounding sediments, groundwater or to the atmosphere but significant quanitities may beretained, depending upon the character of the coal reservoir.Gas may be retained in the coalbed reservoir in several forms: 1) as free gas (gas in excessof that which can be sorbed, in the cleats, fractures and porosity of the coal) 2) as a solute ingroundwater occupying the coal reservoir; and 3) as sorbed gas upon the internal surfaces (e.g. inmicropores) or within the molecular structure of the coal (Rightmire, 1984; Murray et al., 1991;Ertekin, 1991; Rice, 1993). The third mechanism, sorbed gas, is the primary mechanism formethane gas retention in coal (Rightmire, 1984).The amount of gas that may be sorbed appears to be dependant upon pore size (Gan et al.,1972). In meso- (pore diameters between 2 and 50 nm) and macroporous (pore diameters> 50nm) materials, the pores are thought to be filled with adsorbate by multilayer adsorption upon theinternal pore surface (Chapter 2). Total internal surface area therefore appears to be the primarycontrolling factor upon gas sorption in such materials.Greater gas sorption has been shown to occur in microporous substances (pore diameters<2 nm) such as activated carbons, zeolites (Dubinin, 1966), and coals (Gan et al., 1972) than inmesoporous and macroporous solids of similar composition. Micropores are believed to be filledby volume filling (Dubinin, 1975; Jaroniec and Choma, 1989) as opposed to layer by layeradsorption on the internal surface of the pores, therefore micropore volume, not surface area,appears to be the main control on gas sorption for microporous materials. The proportion thatmicroporosity contributes to the total pore volume is thus an important parameter in evaluatingthe gas sorption characteristics of a solid.49The portion of total pore volume occupied by the various pore size fractions (micro-,meso-, macroporosity) was determined for coals ofvarying rank by Gan et al. (1972). Forvitrinite-rich coals with carbon contents between 76 and 84 %, micropores and mesopores makeup the bulk of the porosity but, for coals of similar maceral composition but with carbon contentsless than 75 %, the porosity is mainly macroporosity. It should be noted that Gan et al. definedmicropores as pores with diameters between 0.4 and 1.2 nm, transitional (meso-) pores as poreswith diameters between 1.2 and 30 nm, and macropores as pores with diameters between 30 and2,960 nm. This pore size classification differs from the current IUPAC classification (Chapter 2)which is used in this thesis.Carbon dioxide surface areas of the coals studied by Gan et al. (1972) were found to varyin the following way with rank: coals ofmedium-volatile bituminous to anthracite (> 85 % C)rank yielded surface areas between 196 - 426m2Ig, high volatile bituminous coals (75 - 85 % C)yielded surface areas between 96 - 228m2Ig and low rank coals (< 75 % C) had surface areas2between 225 and 359 m 1g.Another important factor affecting pore size distributions and surface areas in coal iscomposition. Considerable scatter exists in the surface area values given by Gan et al. (1972) forcoals of similar rank which may be due to the variability in coal composition (Lamberson andBustin, 1993). Harris and Yust (1976;1979) studied high volatile bituminous coal using atransmission electron microscope to determine the pore size and porosity distributions associatedwith the three major maceral groups (vitrinite, inertinite, and liptinite). Vitrinite was found to bemainly micro- and mesoporous; inertinite, the most porous maceral group, was found to be mainlymesoporous; and liptinite, the least porous maceral group, was found to be mainly macroporous.Detailed studies of how coal composition affects gas adsorption characteristics have occurredonly recently (Clarkson et al., 1993; Lamberson and Bustin, 1993).Coal composition may also have an impact upon the producibility of the coalbed methaneresource.Ertekin et al. (1991) have shown that a typical coalbed methane production well has twocharacteristic production rate peaks. The first peak occurs as entrained water is flushed from thefracture system, leading to an increase in the permeability of the reservoir to gas. The second50peak occurs as gas is desorbed and diffuses through the micropore network and ultimately intothe fracture system. Ertekin et al. (1991) demonstrated that the magnitude and time ofoccurrence of the first and second production peaks are affected by several reservoir propertiesincluding coal seam thickness, porosity, permeability, and sorption characteristics. Studiesinvolving the determination of gas adsorption characteristics of coals of varying compositionshould therefore provide valuable production information.3.2.1 Research ObjectivesFactors affecting the natural gas content of and producibility from coal such as thicknessand continuity of the coal seam, rank, pressure, fracture permeability, amount ofmineral matter,and hydrologic conditions have been investigated in some detail (Kim, 1977; Meissner, 1984;Fassett, J.E., 1987; Dawson and Clow, 1992; and Ayers and Kaiser, 1992). The pore structureand resulting pore volume and associated effective internal surface area of a coal is an importantcontrol upon methane gas adsorption and hence retention within coalbeds. The effect of rankupon the pore structure and internal surface area of coals as well as methane gas retention hasbeen investigated thoroughly (Gan et al., 1972; Meissner, 1984) but little attention until now hasbeen focused upon the effect of coal composition, particularly the organic fraction, on the porestructure and gas adsorption characteristics of coal.Coal composition (organic and inorganic) has an important control upon gas sorptioncharacteristics, and, hence, total gas content. It is the objective of the current study to evaluatethe effect of coal composition upon the gas sorption characteristics of coal suites from tworegions of the WCSB. In this study, the micropore capacity (monolayer capacity), surface area,and micropore size distribution of coals of varying maceral and mineral composition areinvestigated.513.3 BACKGROUND3.3.1 Dubinin Theory of Volume FillingforMicroporesMicropores are thought to fill by the mechanism ofvolume filling (Dubinin, 1966, 1975;Jaroniec and Choma, 1989; Stoeckli, 1990) as opposed to multilayer formation and capillarycondensation which occurs in the larger mesopores. The adsorption capacity in micropores islarge due to the availability of the total (accessible) micropore volume as adsorption space(Jaroniec and Choma, 1989). Enhanced adsorption in microporous materials occurs over mesoand macroporous materials of similar composition due to the overlapping potentials between porewalls of pores commensurate in size with the adsorbate molecule.Two basic equations are derived from Dubinin’s theory of volume filling ofmicropores(TVFM) (Dubinin 1965; 66; 75; 82; 83, 85; 89). The first is the Dubinin-Radushkevich (D-R)equation (Gregg and Sing, 1982; McEnaney, 1987) which may be written for plotting purposesas:1) logW = logW0 -B(T/I3)21og(P0/ )where W is the volume filled by adsorbate at equilibrium pressure; W0 is the limiting microporevolume; B is a structural parameter related to the pore size (energy) distribution of the adsorbent;R is the Universal Gas Constant, T is the temperature in K, 3 is the relative differential molarwork of adsorption or affinity coefficient of the adsorbate relative to benzene or nitrogen(standard adsorbate); P is the equilibrium adsorbate vapour pressure; and P0 is the adsorbatevapour saturation pressure. A critical parameter from which the D-R equation was derived is thedifferential molar work of adsorption, A, where A = RTIn(P0/P). In the D-R equation, thedistribution of pore sizes (or more correctly, the distribution of the differential molar works ofadsorption, A) is assumed to be Gaussian. Further, characteristic curves, which are plots of e =W/W vs A are assumed to be invariant with temperature.52A plot of logy, where V is the volume of gas adsorbed at equilibrium (cm3/g, stp) versus1og2(P/P), referred to as the Dubinin transformed isotherm plot, should yield a straight line if thetheory of volume filling ofmicropores is obeyed. The micropore capacity, V0, may be obtainedfrom the Y-intercept of the Dubinin transformed isotherm plot. The micropore (monolayer)capacity is related to W0, the limiting micropore volume, through the relation: W0 =V0XD,where D is a density conversion factor (cm3 liquid/cm3stp), if it can be assumed that theadsorbate density is equal to the density of the bulk liquid at the adsorption temperature. Themicropore surface area may be obtained from the the monolayer capacity by multiplying themonolayer capacity by the cross-sectional area of the adsorbate molecule.The second equation which is the outcome of TVFM is the Dubinin-Astakhov (D-A)equation (Dubinin and Astakhov, 1971; Dubinin, 1975; Jarionec et al., 1990). The D-A equationis a generalized form of the D-R equation, developed to account for broader pore sizedistributions than the D-R equation, and may be written as:2) W/W0 = 9 exp [-(RT/E)T’lnlL(P0IP)]where E is the “characteristic free energy of adsorption” which is equal to the differential freeenergy of adsorption for 0 = .368. The free energy of adsorption is believed to be an inversethnction of the average micropore size (Stoeckli et al., 1989), and has been related to the half-width of slit-shaped micropores using the radius of gyration obtained from small-angle scatteringX-ray techniques (Dubinin and Stoeckli, 1980; Dubinin, 1982; Jaroniec and Choma, 1989).The energy distribution in the case of the D-A equation is assumed to obey a Weibulldistribution (Weibull, 1951; Gregg and Sing, 1982), and the exponent n is optimized to obtain abest fit to the linear regression obtained from the transformed Dubinin plot (plot of logW vslog”(P0IP)). For the case of n =2, the D-A equation reduces to the D-R equation. The value of ncan give some indication of the nature of the pore size distribution, as discussed in Stoeckli et al.(1989).53In order to obtain a distribution of pore sizes from the Weibull distribution of adsorptionenergies assumed in Dubinin-Astakhov Theory, certain assumptions must be made (Medek,1977). Firstly, the total adsorption potential (0) is assumed to obey the following relationship:3) Økz3where k is referred to as the interaction constant and z is the distance from the adsorbate moleculeto the adsorbent surface. Secondly, if the adsorbate molecule is assumed to be adsorbed in aconfined space and interacts with adsorbent walls in all directions, then z may be thought of as theaverage distance to the pore walls and equated to an equivalent pore radius (re) in the followingequation:z = re = 2Q/Pwhere Q is the area and P is the perimeter of the pore in cross-section. Finally, equation 3 isthought to be obeyed over the size range in which volume filling is thought to occur. Equation 3may then be written as:4) A=O=kre3The cumulative distribution function for pore sizes may be obtained by substitutingrelation 4 into equation 2 and the pore size distribution curve may be obtained differentiating theresulting equation with respect to the equivalent radius (Medek, 1977). Parameters such as themean equivalent pore diameter may be obtained from the parameters of the Weibull distribution(Medek, 1977).543.3.2 Langmuir andBET TheoryThe classic theory used to describe the Type I isotherm for microporous materials withsmall external surface area is that based on the Langmuir equation (1916). The Type I isothermdisplays a steep increase in adsorption at low relative pressures due to enhanced adsorptioncaused by the overlapping adsorption potentials between walls of pores whose diameters arecommensurate in size with the adsorbate molecule. The Type I isotherm then flattens out into aplateau region at higher relative pressure, which is believed to be due to the completion of amonolayer of adsorbed gas. The micropore volume is then thought to be filled by only a fewmolecular layers of adsorbate, and further uptake is limited by the dimensions of the micropores.The Langmuir model assumes that a state of dynamic equilibrium is established betweenthe adsorbate vapour and the adsorbent surface and that adsorption is restricted to a singlemonolayer (Gregg and Sing, 1982). The adsorbent surface is thought to be composed of aregular array of energetically homogeneous adsorption sites upon which an adsorbed monolayer isassumed to form. The rate of condensation is assumed to be equal to the rate of evaporation fromthe adsorbed monolayer at a given relative pressure and constant temperature. The Langmuirequation was developed with these assumptions and takes the following form:P= 1+PV BVmVmwhere P is the equilibrium pressure, V is the volume of gas adsorbed at equilibrium, Vm is thevolume of adsorbate occupying a monolayer, and B is an empirical constant. A plot ofP/V vsrelative pressure should yield a straight line whose slope will yield Vm from which the surface areamay be obtained. As discussed by Gregg and Sing (1982), variance from linearity often occurs.For example, the heat of adsorption of carbon dioxide gas appears to vary with degree ofuptakewhich is contrary to the assumption of constant heat of adsorption with surface coverage made byLangmuir.55The Brunauer, Emmett and Teller (BET) treatment (1938) is simply an extension of theLangmuir kinetic theory of adsorption whereby the dynamic equilibrium discussed above isextended to second and higher layers (multilayer adsorption). Some simplifying assumptions aremade in developing the BET equation: the heat of adsorption in second and higher layers isassumed to be equal to the adsorptive heat of liquefaction; condensation occurs only on sitesoccupied by molecules in a previously adsorbed layer; the number of adsorbed layers at saturationis infinite; and no lateral interaction occurs between adsorbate molecules. The BET equation is:__= 1+ C-1(PfP0)V(1P1P0) VmC VmCwhere C is the BET constant which is a function of the net heat of adsorption (Chapter 2). A plotof the left side of the equation versus relative pressure will yield a straight line. The values of Cand Vm may be obtained from the slope and the intercept. The monolayer volume may beconverted to surface area if the adsorbate cross-sectional area is known.563.4 METHODSTwo sample suites of coal of slightly different rank are used in the current study. The firstsuite consists of seven samples of the Lower Cretaceous Gates Formation ofNortheastern BritishColumbia (Lamberson and Bustin, 1993) obtained from the Bulimoose Mine C seam and onesample from the Al seam, approximately 50 m stratigraphically below the C seam. The secondsuite of eight samples was obtained from a drill core of Cretaceous coals from a locality inAlberta. Each suite represents a wide variation in lithotype composition.Petrography (maceral and mineral), proximate, sulphur, and low pressure carbon dioxideanalyses were performed on both sample suites. In addition, low-temperature ash (LTA), X-raydiffraction (XRD), equilibrium moisture, and high pressure methane adsorption analyses wereperformed for the Gates sample suite (Lamberson and Bustin, 1993). Petrographic, sulphur,LTA, XRD, and random reflectance procedures used are described in Lamberson and Bustin(1993). Samples were crushed to less than 250 .tm screen size for all analyses.Carbon dioxide adsorption analysis was performed at The University ofBritish Columbiausing a Micromeritics ASAP 2000® surface area analyzer. Samples were evacuated at 70°C forat least 16 hours prior to analysis to remove residual volatiles. Each sample (with sample tube)was then transferred to an analysis port on the instrument, back-filled with helium to remove anyremaining vapours, and re-evacuated. A preliminary leak test was then performed; this consistedof opening the sample tube to a pressure transducer which monitored pressure buildup due to therelease of volatiles from the sample. If a critical pressure increase was not achieved over a 60second interval, then the analysis was continued. Upon passing the preliminary leak test andfurther evacuation, a free space analysis was performed using helium gas at the analysistemperature, followed by a more stringent leak test. After the secondary leak test was passed, thesample was cooled to analysis temperature, exposed to fixed doses ofResearch Grade (99.999%)carbon dioxide, allowed to come to equilibrium, and the adsorbed volume of carbon dioxide gaswas measured. The analyses were performed over a relative pressure range of about 0.0006 to0.01 at 298 K, and 0.0006 to 0.032 at 273 K. No thermal transpiration correction or non-ideality57gas correction was used for carbon dioxide at either temperature. Carbon dioxide adsorptionisotherms were obtained for all samples at both 298 and 273 K. A saturation pressure of—6.4196MPa (48,151 mmHg) and 3.4853 MPa (26,142 mmHg) was used for carbon dioxide at298 and 273 K, respectively.The Dubinin-Radushkevich equation was used to obtain carbon dioxide microporecapacities and micropore surface areas from the 298 and 273 K isotherm data. A molecularcross-sectional area of 17.0 x icf20 m2 was used for carbon dioxide to obtain the surface areasfrom the monolayer capacities.The Dubinin-Astakhov equation was utilized to obtain pore size distribution data. Ananalysis bath temperature of 273 K was used instead of 298 K due to the larger relative pressurerange obtained at the lower temperature (0.0006-0.032 versus 0.0006-0.0 1). The upper limit forabsolute pressure measurements on the ASAP 2000 instrument was 120 kPa (900 mmHg). Anaffinity coefficient (J3) equal to 0.44 (Stoeckli et al., 1993) was used in this study for carbondioxide at 273 K.High pressure methane adsorption analyses were performed by Core Laboratories(Calgary, Alberta) using a high pressure volumetric adsorption technique similar to that ofMayorCt al. (1990). The procedure for sample preparation is described in Lamberson and Bustin (1993).583.5 RESULTS3.5.1 Gates suite3.5.1.1 Proximate, rank, and petrographic dataProximate, sulphur, equilibrium moisture, and ash yield data for the Gates suite arepresented in Table 3-1. Lithotype classification, LTA and XRD results are presented in Table 3-2. For a discussion of these results, see Lamberson and Bustin (1993).Random reflectance measurements on samples LTC-1, LTC-15, and LTA1-6 yieldedvalues of 0.97, 0.96, and 1.0 %, respectively. These values indicate a high-volatile A - medium-volatile bituminous rank for the Gates samples (Lamberson and Bustin, 1993). Rank, followingAmerican Society of Testing and Materials (ASTM, 1980) procedure, is medium to low volatilebituminous. The ASTM rank determination may be inappropriate for some western Canadiancoals, as discussed in Lamberson and Bustin (1993).Petrographic composition data for the Gates Formation coals is given in Table 3-3 andpresented graphically in Figure 3-1. The maceral percentages were calculated on a volumepercent, mineral matter-free basis (mmf), and were then recalculated to include mineral matterusing the Parr formula (Lamberson and Bustin, 1993). Liptinite content of the Gates suite is verylow (0-3 %, mmf), thus the coals are composed mainly of three components: vitrinite, inertinite,and mineral matter. Vitrinite content (mmf) ranges from 18 to 95%, and inertinite from 3 to 81%.On a raw coal basis (mineral matter-inclusive), vitrinite varies from 15 to 90%, inertinite from 3 to71%, and mineral matter from 2 to 22%. LTC-1 has the highest vitrinite and lowest mineralmatter and inertinite content; LTC-5 has the highest inertinite content; LTC-9 has the highestmineral matter content. LTC-5 is a unique sample containing 81% inertinite (mmf), which ismainly structured inertinite. The structured inertinite is interpreted as having been derived from59Table 3-1. Results of proximate, sulphur, and equilibriummoisture analyses.Sample Ash Moisture Volatile Fixed Total Equ.Yield (AR) Matter Carbon Sulphur Mois.(w%) (w%) (w%, dmf) (w%, dmf) (w%) (w%)LTC-1 8.0 1.1 28.0 72.1 0.7 2.1LTC-7 6.8 1.0 24.0 76.0 0.7 1.7LTC-15 9.6 0.7 25.9 74.1 0.8 1.7Gates LTA1-6 3.6 0.9 23.7 76.3 0.4 2.0Suite LTC-11 20.5 0.9 21.5 78.5 0.4 1.6LTC-14 17.6 0.7 21.3 78.8 0.5 1.4LTC-9 33.8 0.8 17.4 82.6 0.3 1.5LTC-5 22.2 0.5 30.5 69.5 0.2 1.4ACCC-27 4.3 0.1 35.1 64.9 0.6ACCC-29 1.6 0.2 35.1 64.8 0.5ACCC-1 6.3 0.4 33.1 66.9 3.0Alberta Accc-5 4.4 0.3 34.3 65.6 1.5Suite ACCC-6 2.9 0.1 35.5 64.5 1.3ACCC-35 10.6 0.5 33.5 66.5 0.5ACCC-13 1.2 0.4 23.1 76.9 1.0*ACCC 4.6 0.1 33.8 66.1 0.9w % = weight percent dmf= dry, mineral matter free (ASTM)AR = As received Equ. Mo/s. = equilibrium moisture60Table 3-2. Lithotype classification, low-temperature ash and x-raydiffraction results for the Gates suite. Modified fromLamberson and Bustin (1993).Sample * Lithotype Low-temp Quartz Kaolinite Dolomite Ferroanash yield Dolomite(weightLTC-l Bright 7.20 major major - minorLTC-7 Banded bright 7.51 major major - minorLTC-15 Banded coal 11.28 major major minor -LTA1-6 Banded dull 2.83 dominant minor - minorLTC-1 1 Banded dull 18.57 major major - minorLTC-14 Dull 19.15 major minor minor -LTC-9 Dull 37.59 major minor - -LTC-5 Fibrous 33.43 minor - dominant -Dominant: essentially monomineralic.Major: strong peak intensity (15 -40% ?).Minor: weak peak intensity (5-15% ?).* Lithotype classification is a modified Australian classification (Diessel, 1965).MaceralTable 3-3. Gates and Alberta suite petrography data.GATES SUITELTC-1 LTC-7 LTC-15 LTA1-6 LTC-11 LTC-14 LTC-9 LTC-6ALBERTA SUITEACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 *ACCC* Structured Vitrinite: Degraded Vitrinite** Structured Inertinite : Degraded Inertinite19 1361Structured 79 50 46 41 22 13 6 11VitriniteDesmocollinite 6 12 9 17 7 2 0 6Vitrodetrinite 11 1 5 1 7 14 13 1Semifusinite 1 24 21 26 38 37 23 20Fusinite 1 9 8 4 3 4 1 53Other Inertinite 2 4 10 11 23 29 54 8Total Liptinite 1 0 0 0 1 1 3 1Total Vitrinite 95 63 61 59 35 29 19 18Total Inertinite 3 37 39 41 64 70 78 81Struct:DegVit* 5 4 3 2 2 1 1 2StructOeg Inerr* 1 9 3 3 2 11 0 9Structured 75 48 43 40 20 12 5 10VitriniteDesmocolliriite 5 11 9 16 6 2 0 5Vitrodetrinite 10 1 5 1 6 13 10 1Semifusinite 1 23 20 25 34 33 18 18Fusinite 1 9 7 4 3 4 1 46Other Inertinite 2 4 9 10 21 26 43 7Total Liptinite 1 0 0 0 1 1 2 1Total Vitrinite 90 61 57 58 32 26 15 16Total Inertinite 3 35 37 40 57 63 61 71Ash Yield (vol.%) 5 4 6 2 10 10 22 12U)U)(Ua)a)E(U0ciU)Eci)U)LçCUU)Cci)ECU0ciU)EStructured 77 73 49 49 34 34VitriniteDesmocollinite 9 9 14 23 19 17 30 22Vitrodetrinite 2 0 6 3 9 5 3 1Semifusinite 3 5 19 14 23 25 31 34Fusinite 5 3 8 6 11 16 10 3Other Inertinite 2 8 2 3 3 2 5 25Total Uptinite 2 2 2 2 1 1 2 1Total Vitrinite 88 83 69 75 63 55 52 37Total Inertinite 10 16 29 23 37 44 46 62Struct:Deg Vit 7 8 2 2 1 2 1 1Struct:Deg Inert 4 1 13 7 13 18 8 1Structured 75 73 47 47 33 32 18 13VitriniteDesmocofllnite 8 9 13 23 19 16 30 21Vitrodetrinite 2 0 6 3 9 4 3 1Semifusinite 3 5 18 13 22 24 31 33Fusinite 5 3 7 5 11 15 10 3Other Inertinite 2 8 2 3 3 2 5 25Total Liptinite 2 2 2 2 1 1 2 1Total Vitrinite 85 82 66 73 61 52 51 36Total lnertinite 10 15 27 22 36 41 46 60Ash Yield (vol.%) 2 1 4 3 2 6 1 362Figure 3-1. Gates coal petrography data. Samples analysed on amineral-free (a) and raw-coal (b) basis. Maceral andmineral contents expressed as volume %. Modifiedfrom Lamberson and Bustin (1993).Mineral-Free100 -8060KEY4020(a)ciE(b)100ci)Ec 400Cu202zC.)0Raw CoalMineralMatterLipti n iteOtherlnertiniteFusiniteSemifusiniteVitrodetriniteDesmocolliniteStructuredVitriniteI I I I I I ILTC-1 LTC-7 LTC-15 LTA1-6 LTC-11 LTC-14 LTC-9 LTC-563fires (charcoal) (Lamberson, 1993). The Gates therefore contains a broad range of compositionswith respect to the three main components vitrinite, inertinite, and mineral matter.Ratios of structured to unstructured vitrinite and inertinite are presented in Table 3-3.Structured vitrinite (telocollinite, telinite, pseudovitrinite) increases with decreasing in mineralmatter content. The higher ash yield samples are enriched in degraded (unstructured) vitrinite(i.e., vitrodetrinite) and inertinite. In general (with exception ofLTC-5), the structured vitrinitecontent decreases with decreases in total vitrinite content. Structured inertinite (semifiisinite andfusinite) appears to decrease with an increase in total inertinite, with the exception ofLTC- 1 andLTC-5. For a discussion of the impact of depositional environment and original vegetation onthese compositional trends, see Lamberson and Bustin (1993).3.5.1.2 Gas AdsorptionPlots of carbon dioxide monolayer capacities (calculated from D-R equation) versusvitrinite content on a mineral matter-free (calculated from Parr formula) and raw coal basis areshown in Figure 3-2. Micropore capacities are corrected to a volume percent, mineral matter-freebasis using the Parr mineral formula (Lamberson, 1993). Both the 298 K and 273 K data isdisplayed in Table 3-4.The 298 K carbon dioxide surface areas (Table 3-5) and micropore capacities (raw coalbasis) vary from 87.1 to 176 m2/g and 19.1 to 38.6 cm3/g, respectively. The 273 K carbondioxide values range from 94.9 to 192m2/g and 20.8 to 40.1 cm3/g, respectively. Experimentalerror associated with these values is +1- 10%. The correlation between the micropore capacitiesand total vitrinite content appears to be logarithmic. For the raw coal data, the correlationbetween mineral matter and micropore capacities appears to be linear but a logarithmic correlationalso yields a high correlation coefficient.b)840Cr 0) C) C) 0 C) 0 C) 0 C) 0 0.0 0a)C) C) C C)C) 0)0 0 0.(C 0 C) 0 0 0 040 30 20TotalVitriniteContent(Vol%,MineralMatter-Free)30020406080TotalVitriniteContent(Vol%,MineralMatter-Free)c)I 1) C 0) 0 0 0 C) 0 0 0 C)d))40Cr 0)C) C) 300 C) 0 a, 0200100Figure3-2.PlotsofcarbondioxidemicroporecapacitiesversusGatessuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemicroporecapacitiescalculatedusingtheParrFormula.10020020406080ComponentVolume(%, RawCoal)020406080ComponentVolume(%,RawCoal)100Table3-4.CarbondioxideDubinin-Radushkevich(D-R)microporecapacities,LangmuirandBETmonolayervolumespresentedonarawcoalandmineralmatter-free(seetext)basis.Dataobtainedfrom298and273Kelvinisotherms.D-RMicroporeCapacityLangmuirMonolayerVolumeBETMonolayerVolume(cclg)(cclg)(cc/g)RawCoal**Mm.FreeRawCoal**Mm.FreeRawCoal**Mm.Free298K273K298K273K298K273K298K273K298K273K298K273K37.738.632.734.032.230.623.919.141.242.135.737.534.032.225.220.839.840.234.834.835.834.030.621.743.443.937.938.437.735.732.223.724.327.621.124.024.220.016.112.830.632.229.829.829.729.321.921.325.628.722.424.626.822.220.614.632.333.531.730.533.032.528.024.322.925.919.922.622.418.815.112.127.829.126.626.925.825.419.518.624.127.021.123.124.920.919.313.829.330.328.327.628.728.224.921.2ACCC-27ACCC-29ACCC-1ACCC-5ACCC-6ACCC-35ACCC-13*ACCC46.449.236.236.237.835.633.833.4**Mm.Free=MineralFreeGatesSuiteLTC-1LTC-7LTC-15LTA1-6LTC-11LTC-14LTC-9LTC-5AlbertaSuite50.254.240.340.342.438.838.236.047.649.737.837.238.537.934.134.351.554.742.141.443.241.338.537.027.330.520.523.522.220.020.519.440.243.233.436.635.832.833.832.828.030.821.424.122.621.320.719.941.243.634.937.636.534.934.133.725.828.919.622.420.819.219.718.735.738.229.832.732.029.630.629.626.529.220.423.021.220.419.919.236.638.631.133.632.631.530.930.4UITable3-5.CarbondioxideDubinin-Radushkevich(D-R),LangmuirandBETequivalentLTC-1LTC-7LTC-15LTA1-6LTC-11LTC-14LTC.9LTC-5surfaceareaspresentedonarawcoalandmineralmatter-free(seetext)basis.Dataobtainedfrom298and273Kelvinisotherms.AlsoincludedareBETCvaluesmesuredat298Kelvin.134.1138.4129.2126.1131.1128.8113.797.0167.3176.5142.0153.7149.0143.7140.9138.946 42 45 43 35 44 44 53 53 50 71 67 69 76 87 80D-RSurfaceAreaLangmuirSurfaceAreaBETSurfaceAreaBETC(sq.m/g)(sq.mlg)(sq.m/g)ValueRawCoal**Mm.FreeRawCoal**Mm.FreeRawCoal**Mm.Free298K273K298K273K298K273K298K273K298K273K298K273K298KGatesSuiteAlbertaSuite172.2188.2181.7198.5110.9139.9117.0147.6176.3192.3183.6200.3125.9147.1131.1153.2149.6163.3158.9173.596.5136.1102.5144.6155.1171.4158.9175.6109.7136.3112.4139.6147.2155.6163.5172.9110.7135.7123.0150.8139.8147.0155.2163.191.6133.7101.7148.3109.3115.0139.6146.973.7100.194.1127.887.194.999.2108.258.397.566.5111.1ACCC-27211.8229.3217.2235.2124.5183.4127.7188.1ACCC-29224.6247.7226.9250.2139.3197.3140.7199.3ACCC-1165.4184.3172.6192.393.6152.597.7159.2ACCC-5165.2184.3169.8189.4107.4167.0110.4171.6ACCC-6172.9193.7176.2197.4101.3163.7103.3166.9ACCC-35162.5177.3172.9188.691.4149.697.2159.1ACCC-13154.5174.4155.9176.093.6154.494.4155.8*ACCC152.5164.4156.7169.088.6149.791.0153.8**Mm.Free=MineralFree104.7127.1110.4118.2132.9123.190.8121.696.5103.0123.1105.5102.4118.0113.885.8116.195.269.289.088.455.385.163.0118.0163.1121.0131.9174.7133.289.7136.093.6102.5149.6105.396.9146.298.887.8135.193.488.7139.689.585.3135.287.767Plots ofLangmuir and BET monolayer volumes (calculated from the Langmuir and BETequations, respectively) versus vitrinite content on a raw coal and mineral matter-free basis (at298 and 273 K) are shown in Figures 3-3 and 3-4 and are presented in Table 3-4. Thecorresponding surface areas and BET C values are given in Table 3-5. The relationship betweenthe BET and Langmuir monolayer volumes and vitrinite content is similar to that observed for theD-R monolayer volume.High pressure methane monolayer capacities (Table 3-6), determined at 295 K using theLangmuir equation (Lamberson and Bustin, 1993), and plotted versus vitrinite content (mineralmatter-free and raw coal basis) are given in Figure 3-5. Like the carbon dioxide data, alogarithmic correlation occurs. The low pressure carbon dioxide monolayer capacities are plottedversus methane monolayer capacities in Figure 3-6 and are given an exponential correlation but alinear correlation also applies.3.5.2 Alberta suite3.5.2.1 Proximate, rank, and petrographic dataProximate and sulphur analysis results are summarized in Table 3-1. Sulphur contentsvary from 0.5 (ACCC-29) to 3.1 wt% (ACCC-1). Volatile matter content (weight %, dmmf)varies from 23.1% (ACCC-13) to 35.5% (ACCC-6). Ash yields (wt %) range from 1.2%(ACCC-13) to 10.6% (ACCC-35). The average ash content of the Alberta suite (4.4%) is muchlower than the average for the Gates suite (15.3%). The moisture contents (as-received) of theAlberta suite vary from 0.1% (ACCC-27) to 0.5 % (ACCC-35) and are lower on average than theGates suite (0.2 % vs 0.8 %, respectively).Random reflectance values for the Alberta suite range from 0.50 to 0.65, which span thesub-bituminous A/high volatile bituminous C boundary. The reflectance values may besuppressed by the high degree of resinite impregnation in the vitrinite macerals, leading to a lowerrank determination. The ASTM ranking for the coals is high volatile A bituminous, with thea,c)Jd) a:300)0 () a) E a, 0 C 0100a,aflH 20 E D a, 0 C 010E 0) C2020406080TotalVitriniteContent(%,MineralMatter-Free)100020406080TotalVitriniteContent(%,MineralMatter-Free)100020204060ComponentVolume(%,RawCoal)80020406080ComponentVolume(%,RawCoal)Figure3-3.PlotsofcarbondioxideLangmuirmonolayervolumesversusGatessuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.100C’0030 20 10 30 20 1020406080TotalVitriniteContent(%,MineralMatter-Free)100020406080TotalVitriniteContent(%,MineralMatter-Free)c)30a C 0) 0 0 a E20 Id)300)0 0 a E 0 >2011080100Figure3-4.PlotsofcarbondioxideBETmonolayervolumesversusGatessuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.1000204060ComponentIume(%,RawCoal)4060ComponentVolume(%,RawCoal)70Table 3-6. High pressure methane monolayer volumesmeasured for the Gates suite.Monolayer Volume (cc/g)Sample Raw Coal Ash-Free* Mineral-FreeLTC-l 19.3 21.3 20.4LTC-7 22.0 23.7 22.9LTC-15 20.5 22.9 21.8LTA1-6 18.7 19.5 19.1LTC-l1 12.3 14.8 13.6LTC-14 15.5 18.6 17.2LTC-9 7.9 11.8 10.0LTC-5 8.1 10.2 9.2* Corrected using ash yield (weight %)0 C-) 20 100 C 0(30200 (‘ 0 Cu30020406080100TotalVitriniteContent(VoI%,MineralMatter-Free)0020406080ComponentVolume(%,RawCoal)100Figure3-5.PlotsofmethaneLangmuirmonolayercapacitiesversusGatessuitecoalcompositiononamineral-free(a)andrawcoal(b)basis.ModifiedfromLambersonandBustin,(1993).PlotsofGatessuitecarbondioxideD-Rmicroporecapacitiesversusmethanemonolayercapacitiesonarawcoalandmineral-freebasis.D-Rmicroporecapacitiescalculatedfromthe273Kisotherm(a),andthe298Kisotherm(b).Anexponentialregressionwasusedineachcase.a)b)252520-200.Cu oC)15i515o o01010551020304050CarbonDioxideMicroporeCapacity(cclg)CarbonDioxideMicroporeCapacity(cc/g)Figure3-6.102030405073exception ofACCC-13, which is medium volatile bituminous in rank. The rank of the Albertancoals is therefore between high volatile bituminous C and A in rank.Petrographic composition data for the Alberta suite are shown in Table 3-3 and presentedgraphically in Figure 3-7. Maceral percentages were calculated in the same fashion as for theGates suite. Liptinite content is slightly higher than the Gates suite (1 to 2% vs 0-3% mmf).Mineral matter-free (mmf) vitrinite composition (volume %) varies from 37 to 88%, and inertinitefrom 10 to 62%. The average mmfvitrinite content is higher for the Alberta coals (67%) than theGates (47%), whereas the total inertinite content is lower (31% vs 52%). On a raw coal basis,vitrinite varies from 36 to 85%, and inertinite from 10 to 60%. ACCC-27 has the highest vitriniteand lowest inertinite content, and *ACCC has the lowest vitrinite and highest inertinite content.Samples ACCC-35 and ACCC-13 have the highest and lowest ash contents, respectively.Ratios of structured vitrinite and inertinite to degraded vitrinite and inertinite, respectively,are given in Table 3-3. With one exception (ACCC-5), the total vitrinite content decreases withdeclining structured vitrinite abundance. The structured vitrinite also decreases with increases intotal inertinite content, with the exception of ACCC-5. As stated by Lamberson and Bustin(1993), the relationship between the structured vitrinite and total inertinite contents appears to bedue to the original depositional environment and vegetation. The structured vitrinite-rich coalsprobably formed from woody peats, whereas the duller coal (inertinite-rich) formed in moreherbaceous (less resistant) wetlands or wetlands subjected to higher fire frequency.The semifusinite (mostly high reflecting) contents of the Alberta coals decreases with anincrease in structured vitrinite content, which may also be related to fire frequency in wetlands.The partial burning and charring of the semifhsinite precursor, i.e. structured vitrinite, would leadto an increase in abundance of high reflecting semifusinite. The sample with the highestsemifusinite content (*ACCC) also has the highest inertodetrinite content, which is probably dueto brittleness of inertinite macerals derived from burning and charring of their precursors.(a)ci)E0>04-(I)00.E0C.)L..a)C)CuIa)ED0>C04-U)00.E0C-)I—0)C.)IFigure 3-7. Alberta suite petrography data. Samples analysedon a mineral-free (a) and raw-coal (b) basis. Maceraland mineral contents expressed as volume %.Mineral-FreeKEY(b)100806040200100806040200ACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 *ACCC74MineralMatterLiptiniteOtherInertiniteFusiniteSemifusiniteVitrodetriniteDesmocolliniteStructuredVitriniteRaw CoalACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 *ACCC753.5.2.2 Gas adsorptionPlots of carbon dioxide micropore capacities versus vitrinite content on a mineral matter-free and raw coal basis at 298 and 273 K are given in Figure 3-8. The 298 K surface areas (Table3-5) and micropore capacities (raw coal basis) vary from 152 to 224m2/g and 33.4 to 49.2 cm3/g,respectively. The 273 K surface areas and micropore volumes (raw coal basis) for the Albertasuite range from 164 to 247m2/g and 36.0 to 50.2 cm3/g, respectively.A linear correlation best fits the relation between carbon dioxide micropore capacities andtotal vitrinite content as opposed to a logarithmic correlation found for the Gates suite. For theGates suite, the correlation approaches linearity for values of total vitrinite greater than 30% (rawcoal basis). The total vitrinite contents (raw coal basis) of all samples in the Alberta suite aregreater than 30%, therefore it is no surprise that the correlation is linear. A better correlation,however, is obtained if the carbon dioxide micropore capacities are plotted against structuredvitrinite content as opposed to total vitrinite content (Figure 3-9).Plots ofLangmuir and BET monolayer volumes versus total vitrinite content at 298 and273 K are shown in Figures 3-10 and 3-1 1. A relationship similar to the plots ofD-R microporecapacities versus vitrinite (total and structured) content was again achieved. Plots ofBET andLangmuir monolayer volumes versus structured vitrinite also yield better correlations thanmonolayer volume versus total vitrinite (Figures 3-12, 3-13).60 50 40 30 20 10a)! C) a, C C) C) 0 C) C) 0 0. 0 0b)C) C) C) C) C 0)0 0 0 C) 0 0.0 C)c) d)50 40 30 20 10 0 50 40 30 20 10020406080TotalVitriniteContent(%,MineralMatter-Free)=0.72273K10020406080TotalVitriniteContent(%, MineralMatter-Free)100060C) 0 o50 40 300 C) 0 C) o200,010100Figure3-8.PlotsofcarbondioxideD-RmicroporecapacitiesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemicroporecapacitiescalculatedusingtheParrFormula.020406080ComponentVolume(%,RawCoal)020406080ComponentVolume(%, RawCoal)1000Figure3-9.PlotsofcarbondioxideD-RmicroporecapacitiesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemicroporecapacitiescalculatedusingtheParrFormula.r=0.84298K0C)160 10U,0 0 0d)50 40 30 20 10 0 50 40 30 20 10r2=0.85273KII20406080StructuredVitriniteContent(%,MineralMatter-Free)1000204060StructuredVitriniteContent(%,MineralMatter-Free)0CU 0 0 0 CU 0.CU 0 a) 0 0. 2 060 50 40 30 20 100204060ComponentVolume(%,RawCoal)800204060ComponentVoume(%,RawCoal)80b10z E 0) C C) -JFigure3-10.PlotsofcarbondioxideLangmuirmonolayervolumesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.0)C)CIL:100E 0) C273K020406080TotalVitriniteContent(%,MneralMatter-Free)20406080100TotalVitriniteContent(%,MineralMatter-Free)d)0204060ComponentVolume(%,RCoal)80020406080ComponentVolume(%,RawCoal)100-.10030 20 10 0a, a, Ca a, C 0) 0 0 a, E ICa a, Cbc) d)30 20 10 0r2=0.70273°K020406080TotalVitriniteContent(%,MineralMatter-Free)10020406080TotalVitriniteContent(%,MineralMatter-Free)100020406080ComponentVolume(%,RvCoal)Figure3-11.400 0300 0 a, 2010100PlotsofcarbondioxideBETmonolayervolumesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.020406080100ComponentVolume(%,RCoal)ca)(a30a)ra)CC30220a)a)20E2Er=0.81r=0.8210a)100298KCCo273K•-0I0II0204060801000204060800)0)CStructuredVitriniteContent(%,MineralMatter-Free)CStructuredVitriniteContent(%,MineralMatter-Free)d)b)2O40•r079030f*nitea)‘4r2=0.75IjteIa)EI MineralI204I..-.-.-.I20II444a)44>2°r2=0.22‘4Cr=0.27Co10298K410273K444E 0)I2II0)020406080020406080ComponentVolume(%, RawCoal)ComponentVolume(%,RawCoal)Figure3-12.PlotsofcarbondioxideLangmuirmonolayervolumesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.C0 0 c’ 0, E Ir2=0.81298K030 20 10 0 30 20 10a)ibr2=0.79273K20406080StructuredVitriniteContent(%,MineralMatter-Free)10030 20 10 040 30 20 10II020406080StructuredVitriniteContent(%,MineralMatter-Free) IteIa=0.72I MineralIaI.-.....IaIISaSa0.29Saaa273Kaaa0204060ComponentVo’ume(%,RawCoal)800204060ComponentVolume(%,RawCoal)Figure3-13.PlotsofcarbondioxideBETmonolayervolumesversusAlbertasuitecoalcompositiononarawcoal(b,d)andmineral-free(a,c)basisat298K(a,b)and273K(c,d).Mineral-freemonolayervolumescalculatedusingtheParrFormula.8000823.6 DISCUSSIONThe adsorption characteristics of the two coal suites studied are determined by theirpetrographic compositions. The composition of the coals affects the pore structure and resultingmicropore capacity, which in turn determines the ultimate gas capacity. Details about themicropore structure and adsorbate-adsorbent interactions may be obtained through the study ofthe Dubinin parameters and plots.In the following discussion, factors affecting the gas adsorption characteristics of the coalsuites are discussed, and the Dubinin plots examined in detail in an attempt to elucidate the effectof coal composition upon pore structure. Langmuir and BET plots for both suites are used toobtain further information concerning the effect of coal composition upon gas adsorptioncharacteristics of the coals. The variation of equilibrium moisture content with composition forthe Gates coals is examined and the implications for methane gas adsorption discussed. Finally,the origin ofmicroporosity with respect to coal structure will be discussed.3.6.1 Gates SuiteFor the Gates suite, the low pressure carbon dioxide (Figure 3-2) and high pressuremethane monolayer (Figure 3-5) capacities show a general decrease with increased mineral matteror inertinite content. Conversely carbon dioxide and methane monolayer capacities increase withincreased total vitrinite content. The high pressure methane monolayer capacities are smaller butcorrelatable with the carbon dioxide micropore capacities (Figure 3-6). The smaller carbondioxide micropore capacities may be due to the quadrupolar nature of the carbon dioxidemolecule which may allow it to assume a more closely packed arrangement within the microporescompared to methane (Lamberson and Bustin, 1993). The polar carbon dioxide molecule mayalso interact more strongly with polar groups at the micropore surface than methane. Finally, thehigh pressure methane analyses were performed at equilibrium moisture whereas the carbondioxide analyses were performed on evacuated samples. Previous studies (Joubert et al., 1973)83have shown that sorption ofmethane decreases with an increase in moisture up to a criticalmoisture content; this may also explain the lower high pressure methane monolayer capacities.The sample with the smallest carbon dioxide micropore capacity is the sample with thehighest total inertinite content. The most abundant inertinite maceral in LTC-5 is fusinite, whichappears to suppress the amount of gas adsorbed.The increase in carbon dioxide micropore capacity (and micropore surface area) withvitrinite content is due to an associated increase in the total amount ofmicroporosity (Figure 3-15). These results support earlier conclusions that vitrinite is essentially microporous whereasinertinite is mainly meso-macroporous (Harris and Yust, 1979). The coals in the Gates suite havesimilar mean micropore sizes, but differing micropore capacities, which are dependant uponvitrinite content.The sample with the largest carbon dioxide and methane monolayer capacity, however, isLTC-7 which does not have the highest total vitrinite content, but a mixture ofvitrinite andinertinite. The inertinite in LTC-7 is mostly in the form of semifusinite, a submaceral interpretedto be created by the partial burning (charring) of vitrinite precursors. It is possible that theburning of vitrinite precursors creates microporosity. The loss of volatile matter as a result ofcharring may open up the pore structure, allowing additional adsorption. The process may beanalogous to that described by Dubinin (1982) for strongly or overactivated carbons, wherebymicroporosity is thought to be created by the removal ofwalls between adjacent microporesthrough burning (Dubinin and Stoeckli, 1980). Dubinin (1966) proposed a two-term D-Requation to account for two linear segments of the transformed plot in active carbons subjected tovarying degrees of burnout. The steeper sloped linear segment was believed to be due to theexistance of supermicroporosity (< 1.4 - 3.2 nm diameter) created by burnout of the activatedcarbon and the shallower segment due to inherent microporosity. In the D-R transformedisotherms given for the Gates samples (Figure 3-14), only one linear segment is observed,however. That burning creates microporosity in semifusinite is supported by the fact that LTC-784aI0aI-0Ia)2 4 6 8 10TRANSFORMED PRESSURE, (LOG(P0IP) “ 2.0000)1.20.80.60.40.20(0.2)b)1.41.20.80.60.40.20(0.2)12Figure 3-14. Dubinin-Radushkevich transformed isotherm plots forthe Gates (a) and Alberta (b) suites. Calculations weremade using the 273 K carbon dioxide isotherm.2 4 6 8 10TRANSFORMED PRESSURE, (LOG(PoIP) “ 20000)850.12co.iLI 0.08D0.060.040.160.140.12LL01D-J0.080.02EQUIVALENT PORE DIAMETER, (nm)a) 0.14LTC-1LTC-7LTC-15-.-LTA1-6pLTC-11LTC-14LTC-9isLTC-50.02b) 0.180.9 1 1.1 1.2 1.3 1.4 1.5ACCC-27ACCC-29ACCC-1-.-ACCC-5pACCC-6pACCC-35ACCC-13x*ACCC0.9 1 1.1 1.2 1.3 1.4 1.5EQUIVALENT PORE DIAMETER, (nm)Figure 3-15. Dubinin-Astakhov differential pore volume plots for theGates (a) and Alberta (b) suites. Calculations were madeusing the 273 Kelvin carbon dioxide isotherm.86has the greatest total number ofmicropores (Figure 3-15) even though it does not have thehighest total vitrinite content.The increased adsorption within the semifUsinite and vitrinite-rich coal could be also beattributed to the swelling properties of semifusinite. Semifusinite has been demonstrated to swellin water to a greater extent than vitrinite and may swell more than vitrinite when carbon dioxide isadsorbed, creating more accessable surface area (Unsworth et al., 1989).Semifiisinite content appears to be an important factor controlling gas yield determinedfrom canister desorption studies ofwestern Canadian coals. Potter (1993) noted that gas yield isgreatest in coals of the Mist Mountain Formation (Southeastern Alberta) with semifusinite as thedominant inertinite maceral. Gas desorption results from this same formation are correlated withmicropore capacities later in this chapter.3.6.1.1 Dubinin-Radushkevich PlotsD-R plots (Figure 3-14) reveal information about the nature of porosity and adsorbateadsorbent interactions in the Gates coals. The plots have similar slopes but differing intercepts.The coals have differing micropore capacities (obtained from Y-intercept) depending oncomposition as discussed above.The linearity of these plots suggests that the assumption that adsorption energies inmicropores of the Gates coals obey a Gaussian distribution is satisfactory. Further, deviationsfrom linearity due to the presence of polar-polar adsorbate-adsorbent interactions does not occuras might be expected for the interaction of the quadrupolar carbon dioxide molecule withfunctional groups of the coal surface. Marsh states (1987) however, that with polar adsorbates,the gradient of the D-R plot should decrease with increasing temperature due to the temperaturedependance of polar adsorbate adsorption. The Gates 273 K D-R plots do have a higher averagegradient than the 298 K D-R plots, suggesting that the quadrupolar carbon dioxide adsorbate isnot strictly being adsorbed to the coal surface through temperature invariant van der Waals87forces. The increased gradient at lower temperatures accounts for the higher micropore capacitiesat 273 K.The consistancy of the gradient at each temperature of the D-R plots suggests that theaverage pore size of the Gates coals is similar, as revealed by the pore size distribution plots(Figure 3-15). According to Marsh (1987), for adsorbents of similar type, lower gradientsindicate narrow pores, and higher gradients represent wider pores. The mean equivalent porediameter as determined from the Dubinin-Astakhov treatment, however, does decrease slightlywith total vitrinite content and increase with mineral matter content (Figure 3-16).3.6.1.2 Dubinin-Astakhov D(ferential Pore Volume PlotsThe Dubinin-Astakhov differential pore volume plots for the Gates suite are shown inFigure 3-15. The pore size distributions are fit to a Weibull distribution as opposed to a Gaussiandistribution assumed for the Dubinin-Radushkevich equation. In the Dubinin-Astakhov treatment,the exponent n’ is optimized, whereas for the Dubinin-Radushkevich treatment the exponent isassumed to be equal to 2. The value of’n’ is believed to reflect the nature of the pore sizedistribution.The exponent ‘n’ for the Gates suite appears to increase linearly with vitrinite and decreasewith inertinite and mineral matter content (Figure 3-16), although the value does not vary muchfrom 2 (Table 3-7). Dubinin (1966) states that a value of n = 2 is indicative of a carbonaceousadsorbent with a homogeneous micropore distribution. Stoeckli (1989), however, states thathomogeneous active carbons should have an exponent ‘n’ equal to 3, and that the degree ofheterogeneity of the micropore system increases as ‘n’ decreases. The value of ‘n’ for the Gatescoals is for the most part less than 2, so they do not qualify as truly homogeneous molecular sievematerials. The decreasing value of ‘n’ with decreasing vitrinite content and increasing inertinitecontent suggests that the degree of heterogeneity of the micropore system increases withdecreasing vitrinite and increasing inertinite content. This follows because the two maceralgroups have different ranges of pore sizes.881.602.1.10a1•9uJ1.84-< 1.7a)b)40 60 80Component Volume (%, Raw Coal)100__8.500w2’0Lu074-C)6.50c)? 1.5414-.G)$ 1.481.46. 1.441.421.40Figure 3-16.20 40 60 80Component Volume (%, Raw Coal)10020 40 60 80Component Volume (%, Raw Coal)100Plots of Astakhov exponent (a), characteristic energy (b) andmean equivalent pore diamter (c) versus Gates coal compositionon a raw coal basis. The 273 K isotherm was used.Table3-7.Astakhovexponents,characteristicenergies,andmeanequivalentporediameterscalculatedfrom273Kcarbondioxideisotherm.SampleAstakhovExponentCharacteristicEnergyMeanEquivalentPoreDiameternE(KJ/mole)(nm)LTC-1LTC-7LTC-15LTA1-6LTC-11LTC-14LTC-9LTC-52.021.991.821.861.831.761.661.708.077.907.247.527.036.946.617.181.411.421.471.451.491.501.531.48AlbertaSuiteACCC-27ACCC-29ACCC-1ACCC-5ACCC-6ACCC-35ACCC-13*ACCCGatesSuite2.032.072.001.892.011.851.941.858.148.138.338.448.598.089.008.381.411.411.401.401.381.421.361.4000903.6.1.3 Langmuir andBETAnalysisCarbon dioxide monolayer volumes were obtained in the low relative pressure range, usingLangmuir and BET theories, to determine the relationship between these values and carbondioxide D-R micropore capacities. Langmuir and BET monolayer volumes were obtained byextrapolation from the linear portion of the plot, at relative pressures greater than .004, to the Yaxis. At lower relative pressures (< .004) the plot deviates upward from linearity. The cause forthis deviation from linearity for carbon dioxide at relative pressures lower than 0.004 is uncertain,but may be related to the polarity of the carbon dioxide molecule and interaction with surfacegroups of the coal.The BET equation is normally assumed to be valid over the relative pressure range 0.05 to0.30, although Dubinin (1969) obtained a linear plot for nitrogen on carbon black at relativepressures from .005 to 0.15 which is similar to the range of linearity obtained in the current study.The cause for linearity of the obtained plots at relative pressures below 0.1 is believed to be dueto enhanced adsorption potential in micropores which may lead to condensation at lower relativepressures. BET C values for the Gates suite are relatively small compared to values recorded formicroporous materials (Table 3-5), but low values of C are not uncommon with carbon dioxide asan adsorbent (Gregg and Sing, 1982).A linear relationship exists between the Langmuir monolayer volumes found in the aboverange of relative pressures and the D-R micropore capacities for the Gates suite. Kobayashi et al.(1993) have shown that for several adsorbates on a variety of carbons V0 and Vm obey therelationship:V0 = KVmwhere K is a constant. For carbon dioxide (at 298 K) on a variety of carbons, the constant K wasfound to be equal to 1.145. For the Gates coal suite, this constant was found to be equal to1.470.913.6.1.4 EquilibriumMoistureJoubert et al. (1973) noted that adsorption ofmethane is a function of moisture contentup to a ‘critical value’ ofmoisture content. Equilibrium moisture (EM) contents (wt %) of theGates coals appear to increase linearly with total vitrinite content (vol %, raw coal) (Figure 3-17).Unsworth et al. (1989) suggest that the difference in EM of inertinites and vitrinites found byprevious studies is due to the fact that inertinite is mainly meso-macroporous and vitrinite ismainly microporous. Unsworth et al., however found that there is no clear dependance ofEMand total porosity upon vitrinite content in coals.In addition to differences in pore structure between inertinite and vitrinite-rich coal,differences in surface chemistry such as a lack of primary sites for adsorption at low relativepressures for the inertinite-rich coals relative to vitrinite-rich coal may also account for thevariation in adsorbed water (Evans, 1986). Joubert et al. (1973) found that the moisture contentof coal increases with oxygen content of coal in a general way. Vitrinite has a larger number ofoxygen-containing surface complexes that act as primary adsorption sites for the polar watermolecule. Because vitrinite usually has a higher average oxygen content than inertinite (Greene etal., 1982), it follows that that vitrinite-rich coals should also have a higher equilibrium moisturecontent than inertinite-rich coals. The vitrinite-rich coals have higher equilibrium moisturecontents despite the fact that inertinite is more hydrophilic than vitrinite on a macroscopic surface(Arnold and Aplan, 1989). The increase in equilibrium moisture with vitrinite content thereforeappears to be due to both the pore structure and surface chemistry of the vitrinite maceral group.Although equilibrium moisture varies with total vitrinite content of the Gates, increase inmethane adsorption measured at equilibrium moisture with vitrinite content still occurs.922.2 -.2-•0o 2r=.43/VitrinfteE Mineral•••••--.1.6-2ff r=.821.4- • •1.2 -0 20 40 60 80 100Component Volume (%, Raw Coal)Figure 3-17. Plot of equilibrium moisture versus Gates coal compositionon a raw coal basis.933.6.2 Alberta SuiteThe Alberta suite coals yield a linear relationship between carbon dioxide microporecapacity and total and structured vitrinite (Figures 3-8, 3-9). The micropore capacities increasewith an increase in vitrinite content and decrease with inertinite content. A poor correlation,however, is achieved between micropore capacity and mineral matter content for the Albertasuite. No XRD analysis was performed on the Alberta suite so it is difficult to assess thecontribution of the mineral matter to the total surface area of the coals. The mineral mattercontent of the Alberta suite is less variable than the Gates, which might explain the poorcorrelation with monolayer capacities.Samples ACCC-27 and ACCC-29 yield the largest micropore capacities due to their highvitrinite content. ACCC-29 has a larger micropore capacity than ACCC-27 even though theformer has a lower vitrinite content.Semifhsinite does not appear to be a significant contributor to the surface area of theAlberta suite. Semifusinite content is greatest in the coals (*ACCC and ACCC-13) with thelowest carbon dioxide micropore capacity. These two samples, however, have a high totalinertinite content which suppresses adsorption.Plots of the carbon dioxide micropore capacity versus vitrinite content for both the Gatesand Alberta suites is given in Figure 3-18. Because the two sample suites are similar in rank, thevariation between suites is mainly due to composition. The Alberta suite has a higher averagevitrinite content and lower mineral matter content, which may explain the higher average carbondioxide micropore capacities for the Alberta suite.The effects of rank cannot be excluded, however. The Alberta suite is of lower rank andmay contain a higher amount ofmicropore surface polarity or smaller average micropore meansize (see later) due to this fact. Both factors would lead to greater apparent micropore capacitiesfor the Alberta suite.b)50L400)0 0 C) 0 C) a 0 0a)50120 10C)020406080ComponentVolume(%,RCoal)10030ComponentVolume(%,RCoal)20c)60 50 I:: 20 0. 0 a 10 0 0. 0 0d)60 50 40C) 03020406080100020406080ComponentVolume(%,RawCoal)ComponentVolume(%,RvCoal)Figure3-18.PlotsofcarbondioxideD-RmicroporecapacitiesversusGatesandAlbertasuitecoalcomposition,onarawcoalbasis,calculatedfroma298Kisotherm(a,b)anda273Kisotherm(c,d).Totalvitriniteversusmicroporecapacityisplottedinb)andd);Structuredvitriniteversusmicroporecapacityina)andc).100100953.6.2.1 D-RPlotsD-R plots for the Alberta suite (Figure 3-14) are linear, which shows that the adsorptionenergies obey a Gaussian distribution. Gradients increase with a decrease in temperature (as withthe Gates suite) due to the polar interaction of the carbon dioxide molecule with the coal surface.The gradients of the plots, unlike the Gates suite, are not uniform, which may be due to a greatervariation in mean pore size of the samples. The gradients for the low total vitrinite contentsamples are slightly lower than those for higher vitrinite samples. The mean equivalent porediameter, as determined from D-A treatment, does not decrease in a consistant manner withvitrinite content as with the Gates suite, however.3.6.2.2 D-A Dqferential Pore Volume PlotsDubinin-Astakhov differential pore volume plots for the Alberta suite are given in Figure3-15. The two samples that have the largest carbon dioxide monolayer capacities and totalvitrinite contents (ACCC-29 and ACCC-27) have the largest number of micropores. The numberofmicropores, like the Gates suite, appear to increase, in a general way, with vitrinite content.The exponent ‘n’ for the Alberta suite does not vary much from the value of 2, but doesdecrease slightly with a decrease in vitrinite content. An increase in degree of heterogeneity ofpore size with decrease in vitrinite content thus occurs in both suites. The higher average value of‘n’ for the Alberta suite (1.9 versus 1.8) is indicative of smaller pore sizes.The mean equivalent pore diameter of the Alberta suite is slightly smaller than that of theGates suite (1.40 nm and 1.47 nm, respectively) which may be due to: a) higher average vitrinitecontent of the Alberta suite; b) the difference in rank c); resinite impregnation in structuredvitrinite of the Alberta suite; or d) mineral matter composition. Resinite impregnation mayconstrict pore access analogous the situation of active carbon impregnation with cobalt and nitratesolutions (Alvim Ferraz, 1989). Pore constriction due to resin impregnation may decrease theaverage micropore size in vitrinite. This may exlain why ACCC-27 and ACCC-29 have higher96micropore surface areas than LTC-7 or LTC- 1 of the Gates suite, despite the fact the two Albertasuite coals have lower total vitrinite contents.Another possible cause of the difference in mean pore sizes between suites may be due tothe type ofmineral matter present in the coals. The Alberta suite may contain more mixed layerclay which could increase the micropore volume. The Gates suite contains little or no mixed layerclays, whereas the mixed layer clay content of the Alberta suite is not known. The interlayer ofsuch clays could possibly provide additional adsorption space for carbon dioxide. For example,the basal (001) spacing of the phyllosilicate vermiculite is 1.4 nm, which is close to the meanequivalent pore diameter of the Alberta suite. The accessibility of the adsorption space betweeninterlayers depends on the amount of adsorbed water left in the interlayer after degassing underthe conditions specified earlier. Under the conditions of degassing used, most of the adsorbedwater in the interlayer would likely still be present and the interlayer may not be accessible for thephysisorption of carbon dioxide, although some carbon dioxide gas may be dissolved in interlayerwater. Further research is required on this point.3.6.2.3 Langmuir andBETAnalysisA linear relationship between the Langmuir monolayer volume (Vm) and the D-Rmicropore capacity (Vo) also occurs for the Alberta suite. The value ofK is 1.68 compared to1.47 for the Gates suite.BET C values range from 87 to 50 for the Alberta suite and are higher on average than theGates suite, which is consistant with the higher average micropore capacity of the Alberta suite.973.6.3 Comment on the Origin andNature ofMicroporosity in CoalsA generally accepted view of coal structure is that it consists of a three-dimensional cross-linked macromolecular framework (Greene et al., 1982). Derbyshire et al. (1989) have suggestedthat there may, in fact, be two components of coal including a three-dimensional cross-linkedmacromolecular structure and a molecular phase trapped within this structureSome authors (Dryden, 1963; Fuller, 1981; and Given, 1984) suggest that microporosityis not necessarily a fixed property of coal structure and argue that gas sorption in coal may eitherbe modeled as adsorption within the molecular structure of the coal or as dissolution of thesorbate within the molecular structure. Marsh (1987) states that microporosity exists as ‘space oflow electron density between the macromolecules of the cross-linked entanglements°. The cross-link density has been shown to change with rank: the initial predominantly oxygen cross-links(primarily ethers?) decrease in density to a minimum at Ca. 86% carbon (medium volatilebituminous stage) after which carbon-carbon cross-links are formed. The trend in carbon dioxidesurface areas and micropore volumes of vitrinite-rich coals appears to mimic the trend in cross-link density. These values have been found to decrease with rank to about 85% carbon contentand increase for higher carbon content (Mahajan, 1982). As the cross-link density decreases withrank below 85% carbon content, the adsorption capacity of the coal would decrease due to adecrease in adsorption potential in the pore space. An additional consideration is that the oxygencontent of the whole coal decreases with coalification, and that the interaction of the polar carbondioxide molecule with the oxygen containing functional groups may decrease to the minimum at85% carbon. Polar interactions of the carbon dioxide molecule in lignites is not consideredsignificant.Increase in carbon dioxide surface areas at carbon contents greater than 85% may berelated to the formation of carbon-carbon cross-links. Toda et al. (1971) have shown, however,that the carbon dioxide micropore volumes and apparent size increase with the amount ofaliphatic, alicyclic, and aromatic CH hydrogen.98The trend in carbon dioxide micropore surface areas and volumes with organiccomposition, demonstrated in the current study, must similarly be linked to the structure andsurface chemistry of the macromolecular network of the coal. Vitrinite-rich coals have beenshown to have a greater amount ofmicroporosity, and hence, larger monolayer capacities thanvitrinite-poor coals of the same rank. The vitrinitic components may possess a greater density ofcross-linking and possess a different structural orientation than inertinitic components. Further,differences in surface fhnctionality of the respective maceral groups may account for differences ingas adsorption characteristics. For example, vitrinite possesses a greater oxygen content thaninertinite macerals, and this may lead to a greater interaction of polar adsorbates such as carbondioxide with the micropore surface in coals rich in vitrinite. Functional group type, density andorientation may similarly account for differences in gas adsorption characteristics of coals ofvarying composition.Ultimately a true understanding of the trends in gas adsorption characteristics of coals ofvarying composition is dependant upon the understanding of coal structure and chemistry whichcontinue to be debated (Derbyshire et al., 1989).993.7 CONCLUSIONSThe current study has focused on the gas adsorption, particularly low pressurecarbon dioxide adsorption, characteristics of two coals suites representing a large range incomposition. Several important conclusions have been reached through the current study:1) For both suites, Dubinin-Radushkevich carbon dioxide micropore (monolayer) capacitiesincrease with total vitrinite content and decrease in a general way with mineral matter content. Abetter correlation was acheived between structured vitrinite composition and micropore capacityfor the Alberta suite than for total vitrinite. For the Gates suite, high pressure methane monolayercapacities, as determined from the Langmuir equation, display a similar relationship to coalcomposition as carbon dioxide micropore capacities and are correlative but smaller than thecarbon dioxide monolayer capacities.2) For both suites, a general increase in the total number of micropores occurs with an increase invitrinite content which, in turn, leads to an overall increase in carbon dioxide micropore capacitieswith vitrinite content. Microporosity correspondingly decreases with an increase in total(structured, unstructured and degraded) inertinite and mineral matter content.3) Carbon dioxide micropore capacities of the Alberta suite are larger on average than the Gatesmicropore capacities. Such differences are attributed to the higher average vitrinite content of theAlberta suite and differences in rank.4) For the Gates suite, a sample (LTC-7) with a high vitrinite and semifusinite content has thelargest carbon dioxide monolayer capacity and total number ofmicropores. Semiflisinite maycontribute to the large number ofmicropores in this sample due to the creation of microporositythrough charring ofvitrinitic precursors. Swelling due to the adsorption of carbon dioxide gasupon vitrinite and semifusinite-rich coals may contribute to the large micropore capacities of suchcoals, but this effect requires investigation.1005) Carbon dioxide BET and Langmuir monolayer volumes show a similar relationship to coalcomposition as Dubinin-Radushkevich micropore capacities. Langmuir monolayer volumes arelarger than the corresponding Dubinin-Radushkevich micropore capacities but the two arecorrelative.6) Dubinin micropore capacities and Langmuir and BET monolayer capacities obtained at 273 Kare larger than the corresponding values at 298 K which appears to indicate that polar interactionbetween the quadrupolar carbon dioxide molecule and polar surface groups is occurring.7) For both suites, micropore heterogeneity appears to increase with total inertinite and mineralcontent, as indicated by a general increase in the Astakhov exponent ‘n’.8) For both suites, adsorption energies and hence micropore diameters appear to obey a Gaussianand Weibull distribution.9) For the Gates suite, mean equivalent pore diameter decreases slightly with an increase in totalvitrinite content and decrease in total inertinite and mineral matter content.10) For the Gates suite, equilbrium moisture content generally increases with total vitrinitecontent due to an increase in microporosity with vitrinite. High pressure Langmuir methanemonolayer capacities do generally increase with vitrinite content despite this fact.From the above conclusions it is obvious that composition, as well as rank, has a definitecontrol upon the pore structure and adsorption capacity of coal. In fact, the variation ofmethaneadsorption capacities within one suite of compositionally variable coals may be just as large as thevariation observed between coals of varying rank.1013.8 REFERENCESAlvim Ferraz, M.C., 1989. Micropore volume determination in activated carbon. Fuel, 68:635-640.Arnold, B.J., and Aplan, F.F., 1989. The hydrophobicity of coal macerals. Fuel, 68: 65 1-658.Ayers, W.B., Jr., and Kaiser, W.R., 1992. Coalbed methane occurrence and producibility,Fruitland Formation, Navajo Lake Area, San Juan Basin, New Mexico. In: Proceedingsof the Canadian Coal and Coalbed Methane Geoscience Forum, Parksville, BritishColumbia, February 2-5th, 1992. pp. 3-20.Brunauer, S., Emmett, P.H., and Teller, E., 1938. Adsorption of gases in multimolecular layers.The Journal of the American Chemical Society., 60: 309-319.Dawson, F.M., and Clow, J.T., 1992. Coalbed methane research Elk Valley Coalfield. In:Proceedings of the Canadian Coal and Coalbed Methane Geoscience Forum, Parksville,British Columbia, February 2-5th, 1992. pp. 57-7 1.Deissel, C.F.K., 1965. Correlation of macro- and micropetrography of some New South Walescoals. In: J.T. Woodcock, R.T. Madigan and R.G. Thomas (Editors), Proceedings 8thCommonwealth IVlining and Metallurgical Congress, 6: 669-667.Derbyshire, F., Marzec, A., Schulten, H., Wilson, M., Davis, A., Tekely, P., Delpuech, J.,Jurkiewicz, A., Bronnimann, C.E., Wind, R.A., Maciel, G.E., Narayan, R., Bartle, K.,Snape, C., 1989. Molecular structure of coals: a debate. Fuel, 68: 1091-1106.Dryden, I.G.C., 1963. Chemical constitution and reaction of coal. In: H.H. Lowry (Editor),Chemistry of Coal Utilization Supplementary Volume. New York, John & Sons, pp232-295.Dubinin, M.M., 1965. Theory of the bulk saturation ofmicroporous activated charcoals duringadsorption of gases and vapours. Russian Journal ofPhysical Chemistry, 39: 697-704.Dubinin, M.M., 1966. Porous structure and adsorption properties of active carbons. In: P.L.Walker, Jr. (Editor), Chemistry and Physics of Carbon, Volume 2. Edward Arnold, Ltd.,New York. pp. 51-120.Dubinin, M.M., 1975. Physical adsorption of gases and vapors in micropores. In:D.A.Cadenhead, J.F. Danielli, and M.D. Rosenberg (Editors), Progress in surface andmembrane science, Volume 9. Academic press, New York. pp. 1-70.Dubinin, M.M., 1982. Microporous structures of carbonaceous adsorbents. Carbon, 20: 195-200.Dubinin, M.M., 1983. Microporous structures and absorption properties of carbonaceousadsorbents. Carbon, 21: 359-366.102Dubinin, M.M., 1985. Generalization of the Theory ofVolume Filling ofMicropores tononhomogeneous microporous structures. Carbon, 23: 373-380.Dubinin, M.M., 1989. Fundamentals of the theory of adsorption in micropores of carbonadsorbents: characteristics of their adsorption properties and microporous structures.Carbon, 27: 457-467.Dubinin, M.M., and Astakhov, V.A., 1971. Advances in Chemistry Series. No. 102: p102.Dubinin, M.M., and Stoeckli, H.F., 1980. Homogeneous and heterogeneous micropore structuresin carbonaceous adsorbents. Journal ofColloid and Interface Science, 75: pp. 34-42.Ertekin, T., Sung, W., and Bilgesu, H.I., 1991. Structural properties of coal that control coalbedmethane production. In: D.C. Peters (Editor), Geology in Coal Resource Utilization.pp. 105-124.Evans, M.J.B., 1987. The adsorption ofwater by oxidized microporous carbon. Carbon, 25:81-83.Fassett, J.E., 1987. Geometry and depositional environments ofFruitland Formation coalbeds,San Juan Basin, New Mexico and Colorado: anatomy of a giant coalbed methanedeposit. In: Proceedings of the 1987 Coalbed Methane Symposium, Tuscaloosa,Alabama, Nov. 16-19, 1987. pp. 19-3 5.Fuller, E.L., Jr., 1981. Physical and chemical structure of coals: sorption studies. In: M.L.Gobarty, and K. Ouchy (Editors), Coal structure: Advances in Chemistry Series 192,Washington, D.C., American Chemical Society, pp. 293-209.Gan, H., Nandi, S.P., and Walker, P.L., Jr., 1972. Nature of the porosity in American coals.Fuel, 51: 272-277.Given, P.H., 1984. An essay on the organic geochemistry of coal. In: M.L. Gobarty et al.(Editors), Coal Science, v.3 New York Academic Press, pp. 63-252, 339-341.Greene, T., Kovac, J., Brenner, D., and Larsen, J.W. The macromolecular structure of coals. In:R.A. Meyers (Editor), Coal Structure. Academic Press. pp. 199-282.Gregg, S.J., and Sing, K.S.W., 1982. Adsorption, Surface Area and Porosity, Second Edition.Academic Press, New York. 303 pp.Harris, L.A., and Yust, C.S., 1976. Transmission electron microscope observations of porosityin coal. Fuel, 55: 233-236.Harris, L.A., and Yust, C.S., 1979. Ultrafine structure of coal determined by electronmicroscopy. Preprint paper, American Chemical Society, Division ofFuel Chemicals24: 210-217.103Jaroniec, M., and Choma, J., 1989. Theory of gas adsorption on structurally heterogeneoussolids and its application for characterizing activated carbons. In: P.A. Thrower(Editor), Chemistry and Physics of Carbon, Volume 22. Marcel Dekker, Inc., NewYork and Base!. pp. 197-243.Jaroniec, M., Lu, X., Madey, R., and Choma, J., 1990. Comparitive studies of the overalladsorption isotherm associated with the Dubinin-Astakhov equation. Carbon, 28:243-245.Joubert, J.I., Grein, C.T., and Bienstock, D., 1973. Sorption of methane on moist coal. Fuel, 52:181-185.Kim, A.G., 1977. Estimating methane content ofbituminous coalbeds from adsorption data.U.S. Bureau ofMines Report of Investigations 8245, 22p.Kobayashi, M., Ishikawa, E., and Toda, Y., 1993. Experimental relation between the DubininRadushkevich equation and Langmuir equation for various adsorbates on many carbons.Carbon, 31: 990-992.Lamberson, M.N., and Bustin, R.M., 1993. Coalbed methane characteristics of the GatesFormation coals, northeastern British Columbia: effect ofmaceral composition. In press.Langmuir, I., 1916. The constitution and fundamental properties of solids and liquids. Journal ofthe American Chemical Society, 38: 222 1-2295.Levine, J.R., 1993. Coalification: the evolution of coal as source rock and reservoir rock for oiland gas. In : B.E. Law and D.D. Rice (Editors), Hydrocarbons from coal, AAPG Studiesin Geology # 38, pp. 3 9-77.Lowell, S., and Shields, J.E., 1984. Powder Surface Area and Porosity, Second Edition.Chapman and Hall, London, 234 pp.Mahajan, O.P., 1982. Coal porosity. In: R.A. Meyers (Editor), Coal Structure. Academic Press.pp. 5 1-86.Mahajan, O.P., 1991. CO2 surface area of coals: the 25 year paradox. Carbon, 29: 735-742.Mahajan, O.P., and Walker, P.L., Jr. 1978. Porosity of coal and coal products. In: C. Karr, Jr(Editor), Analytical Methods for Coal and Coal Products, Volume I. Academic Press,New York. pp. 125-162.Marsh, H., 1987. Adsorption methods to study microporosity in coals and carbon - a critique.Carbon, 25: 49-58.104Mayor, M.J., Owen, L.B., and Pratt, T.J., 1990. Measurement and evaluation of isotherm data;Proceedings of 65th Annual Technical Conference and Exhibition of the Society ofPetroleum Engineers, SPE 20728: 157-170.McEnaney, B., 1987. Estimation of the dimensions ofmicropores in active carbons using theDubinin-Radushkevich equation. Carbon, 25: 69-75.Medek, J., 1977. Possibility ofmicropore analysis of coal and coke from the carbon dioxideisotherm. Fuel, 56: 131-133.Meissner, F.F., 1984. Cretaceous and lower Tertiary coals as sources for gas accumulations inthe Rocky Mountain area. Source rocks of the Rocky Mountain Region, 1984Guidebook, Rocky Mountain Association of Geologists, pp. 401-431.Potter, J., 1993. Coalbed methane potential and the effects of coal composition and fractures inmedium-volatile bituminous coals from the Mist Mountain Formation, southwesternAlberta (abstract). Geological Association of Canada/Mineralogical Association ofCanada, Joint Annual Meeting, Program and Abstracts, p. A-84.Rice, D.D., 1993. Composition and origin of coalbed gas. In: B.E. Law and D.D. Rice (Editors),Hydrocarbons from Coal, AAPG Studies in Geology # 38, pp. 159-184.Sato, T., 1981. Methane recovery from coal beds: surface and physical properties ofwesternUnited States coals; M. Sci thesis, The University ofNew Mexico, 78 pp., unpublished.Sing, K.S.W., 1989. The use of physisorption for the characterization ofmicroporous carbons.Carbon, 27: 5-11.Stoeckli, H.F., 1990. Microporous carbons and their characterization: the present state of the art.Carbon, 28: 1-6.Stoeckli, F., Huguenin, D., Greppi, A., Jakubov, T., Priblyov, A., Kalashnikov, S., Fomkin, A.,Pulin, A., Regent, N., Serpinski, V., 1993. On the adsorption of C02 by active carbons.Chimia, 47: 213-214.Stoeckli, H.F., Kraehenbuehl, Ballerini, L., and De Bernardini, S., 1989. Recent developments inthe Dubinin equation. Carbon, 27: 125-128.Toda, Y., Hatami, M., Toyoda, S., Yoshida, Y. and Honda, H., 1971. Micropore structure ofcoal. Fuel, 50: 187-200.Unsworth, J.F., Fowler, C.S., Jones, L.F., 1989. Moisture in coal 2. Maceral effects on porestructure. Fuel, 68: 18-26.Weibull, W., 1951. A statistical distribution ofwide applicability. Journal of Applied Mechanics,18: 293-297.1051980 Annual Book ofASTM Standards. Part 26 Gaseous Fuels, Coal and Coke, section D38877.106CHAPTER 4VARIATION IN MESOPORE VOLUME AND SIZE DISTRIBUTIONWITH COMPOSITION IN A HIGH-VOLATILE COALOF THE WESTERN CANADIAN SEDIMENTARY BASiN:IMPLICATIONS FOR COALBED METHANE TRANSMISSIBILITY4.1 ABSTRACTThe influence of composition upon mesopore volumes and surface areas of high-volatile bituminous coals is investigated in the current study. BET surface areas rangefrom 1.1 to 5.3 m2/g on a raw coal basis and generally increase with an increase in totalinertinite content and decrease with an increase in total and structured vitrinite content.Mineral content appears to have little control upon the surface areas. Cumulativemesopore volumes obtained from the adsorption branch ofnitrogen isotherms alsoincrease with inertinite content. Isotherm hysteresis loops indicate a slit shape for themesopores. Gas yields obtained from desorption canister testing generally increase withmesopore volumes obtained from subset samples. Mesopore volumes, which aredependent upon rank and composition, should be considered in methane diflhsionmodeling through coal seams.1074.2 INTRODUCTION AND RESEARCH OBJECTWESCoalbed gas within a coal reservoir is primarily retained as gas adsorbed within thematrix porosity of the coal. Matrix porosity in coals consists ofmicro-, meso- andmacroporosity which represent pore diameters of less than 2 nm, between 2 and 50 nm,and greater than 50 nm, respectively (Orr, 1977). The distribution of pore sizes in coal isprimarily a function of two properties: rank and composition. The control of rank uponpore size distribution and surface area ofmainly vitrinite-rich coals has been investigatedin detail by Gan et al. (1972). The effect of coal composition upon the pore structure ofcoals, particularly mesoporosity, has only received cursory investigation.Coal composition has been shown to be an important control upon the macro- andmicrostructures of coal, and hence, may have an important control upon gastransmissibility (Close, 1993; Gamson et al., 1993). A popular model (Ertekin et al.,1991; Gamson et al., 1993) of how methane gas travels from the micropore network tothe cleat system and ultimately to the borehole is as follows: gas is desorbed from themicropore network due to a decrease in pressure associated with the drilling of the holeinto the seam; diffi.ision ofmethane gas, governed by Fick’s law, through the coal matrixto the macrofracture system (cleat); and flow, governed by Darcy’s law, through the cleatsystem to the borehole. The process may be more complex than this, however (Gamson etal., 1993). Gamson et al. have concluded that microstructures in coal, ranging in size from0.05 - 20 p.m and consisting of fractures and cavities, have an important control uponmethane gas transmissibility of the coal seam. Although the microstructures as defined byGamson et al. fall in the upper mesopore - macropore range of pore sizes, smallermesopores (if present) would surely also affect the diflhsion of gas from the microporenetwork to the microfracture system.108Harris and Yust (1976; 1979) utilized transmission electron microscopy todetermine the pore structure of the maceral groups vitrinite, inertinite and liptinite andfound the inertinite maceral group to be mainly meso- and macroporous, whereas thevitrinite group was found to be mainly microporous. In a recent gas sorption study byFaiz et a!. (1993), it was postulated that an increase in mineral matter content causes adecrease in the meso- and macropore volume of coal and hence a decrease in the totalvolume of adsorbed gas. The effect of the organic composition of coals upon the porestructure was not addressed. In addition, Langmuir volumes obtained from gravimetricgas sorption of carbon dioxide and methane were found in the Faiz et a!. study to show avague negative correlation with inertinite content, but the relationship was masked by theeffect ofvarying rank among the coals. No detailed gas sorption study has beenperformed to determine the effect of coal organic composition upon mesoporosity.The objective of the current study is to determine the effect of coal composition,particularly the maceral fraction, upon mesopore volume and size distribution andassociated BET surface area. In addition, pore shapes are inferred from isothermhysteresis loop shapes (or types). Total mesopore volume and the shape and sizedistribution ofmesopores may prove to be an important control upon coalbed gastransmissibility from the micropore network to the microfracture network, and to a lesserextent, coal gas content. It is therefore important to understand the origin ofmesoporosity.1094.3 BACKGROUND4.3.1 Barret, Joyner, andHalenda (BJH) TheozyMesopores are generally considered to be filled by the duel mechanisms ofmultilayer formation, described by the Brunauer, Emmett, and Teller (BET) equation(Brunauer et al., 1938), and capillary condensation, described by the Kelvin equation(Barrett et al., 1951). BJH theory, (Barrett et al., 1951), which was developed to describemesopore distributions, makes two fundamental assumptions: the pores of the adsorbentare cylindrical in shape; and the pores are filled by multilayer formation and capillarycondensation. The cylindrical pore would contain adsorbate in two forms: an adsorbedifim on the pore wall; and a core of capillary condensate at the center of the pore (Figure4-1).BJH theory does not fit the pore size distribution to a known statistical distribution(i.e. a Gaussian distribution). Further, the adsorbed film is assumed to change thicknessduring adsorption or desorption in the absence of a capillary condensed core of adsorbateliquid.The BJH theory (Barrett et al., 1951) is based on the Wheeler equation, which maybe written as:V0 - V = itJ (r-t)2 L(r) drwhere the integration is carried from rp, the radius of the largest pore filled withadsorbate at a given pressure, to infinity; V0 is the volume of adsorbate adsorbed atsaturation vapour pressure; V is the volume adsorbed at equilibrium pressure; L(r) is thelength of pores with radii lying between r and r + dr; t is the multilayer thickness,110A)B)Figure4-1. Diagram A) shows the location of the adsorbed film andpore core in a cylindrical capillary; B) illustrates thedifference between the Kelvin (rk) and pore (rp) radii.111described by the Halsey equation (as used in this thesis), at equilibrium pressure. Theform of the Halsey equation used in this thesis is as follows:1/3t = .354 x [5/ln(P/P)]where t is the thickness of the adsorbed layer; P0 is the (measured) saturation vapourpressure for nitrogen; and P is the equilibrium vapour pressure. A monolayer thickness of.354 mn is assumed for adsorbed nitrogen in the equation. The step by step description ofhow BJH theory calculates pore size distributions, volumes and surface areas is discussedby Barrett et al. (1951) and Gregg and Sing (1982).4.4 METHODSA sample suite ofWestern Canadian Sedimentary Basin coal was utilized in thisstudy. The suite, which consists of eight samples, was obtained from drill core ofCretaceous coals from a locality in Alberta. The Alberta coals represent a wide range inlithotype composition.Petrography (maceral and mineral), proximate, sulphur, random reflectance, andnitrogen adsorption analyses were performed. Petrography, sulphur, and randomreflectance procedures used are described in Lamberson and Bustin (1993). Samples werecrushed to less than 250 jim screen size for all analyses.Nitrogen adsorption isotherms were collected using a Micromeritics ASAP2000® surface area analyzer. Samples were first evacuated at 100°C for at least 16 hoursprior to analysis to remove residual volatiles. Each sample (with sample tube) was thentransferred to the analysis port on the instrument, back-filled with helium, and reevacuated. A leak test was then performed. During a leak test, the sample tube is openedup to a pressure transducer, and the rate of increase in pressure, due to loss of volatiles112from the sample, is monitored. If a critical pressure is not reached over a set period oftime, analysis is continued. Following the leak test, free space analysis was perfonnedusing helium gas.Nitrogen isotherms at 77 K were then collected. Both adsorption and desorptiondata was collected with a maximum and minimum relative pressure of about .9995 and.0660, respectively. Only the 2 nm to 50 nm pore diameter range is discussed in thischapter as this range represents the mesopore range. Problems associated with using theKelvin equation outside this range are discussed in Chapter 2.Nitrogen gas was the choice of adsorbate for the following reasons (Gregg andSing, 1982): nitrogen gas is inert; the saturation pressure of the gas is large enough so thata large range of relative pressures may be obtained accurately; the cross-sectional area ofthe gas is well established and is relatively small; liquid nitrogen is a readily availablecommon refrigerant and the saturation pressure may be monitored throughout analysis.The following parameters were utilized for nitrogen (at 77 K) in this study: across-sectional area of .162 nm2, a non-ideality gas correction of 6.6 x i0, and a densityconversion factor of 1.5468 x i0. Ultra High Purity (99.999 %) nitrogen gas was usedas an adsorbate.4.5 RESULTS4.5.1 Proximate, ranlç andpetrographic dataProximate and sulphur analysis results are summarized in Table 4-1. Sulphurcontents range from 0.50 (ACCC-29) to 3.1 (ACCC-1) weight %. Volatile mattercontent, on a weight %, dry, mineral matter-free (dmmf) basis, varies from 23 % (ACCC13) to 35 % (ACCC-6). Ash yields (weight %) range from 1.2 % (ACCC-13) to 11 %(ACCC-3 5).113Table 4-1. Results of proximate and sulphur analyses.Sample Ash Moisture Volatile Fixed TotalYield (AR) Matter Carbon Sulphur(w%) (w%) (w%,dmf) (w%,dmf) (w%)ACCC-27 4.3 0.1 35.1 64.9 0.6ACCC-29 1.6 0.2 35.1 64.8 0.5ACCC-1 6.3 0.4 33.1 66.9 3.0ACCC-5 4.4 0.3 34.3 65.6 1.5ACCC-6 2.9 0.1 35.5 64.5 1.3ACCC-35 10.6 0.5 33.5 66.5 0.5ACCC-13 1.2 0.4 23.1 76.9 1.0*ACCC 4.6 0.1 33.8 66.1 0.9w % = weight percent dmf= dry, mineral matter free (ASTM)AR = As received Equ. Mois. = equillbrium moisture114Random reflectance values for the Alberta suite vary from 0.5 to 0.6, which spanthe sub-bituminous Alhigh-volatile bituminous C boundary. The random reflectancevalues of the Alberta suite may be somewhat suppressed by abundant resiniteimpregnation within the cell structure of the vitrinite group maceral, telinite. The ASTMrank (ASTM, 1980), based on proximate and sulphur data, assigned to the coals is high-volatile A bituminous, with the exception ofACCC-13, which is medium-volatilebituminous in rank. The ASTM rank classification may give artificially high rank values toinertinite-rich coals (Lamberson and Bustin, 1993) and therefore the rank of the Albertasuite coals is most likely between high-volatile bituminous C and A.Petrographic composition data is presented in Table 4-2 and shown graphically inFigure 4-2. Maceral percentages were calculated on a volume percent, mineral matter-free (mmf) basis, and were then recalculated to include mineral matter using the Parrformula (Lamberson and Bustin, 1993). The mineral matter-free vitrinite composition(volume %) varies from 37 to 88 %, and the inertinite from 10 to 62 %. On a raw coalbasis, vitrinite composition ranges from 36 to 86 %, and inertinite from 10 to 60 %.ACCC-27 has the highest vitrinite and lowest inertinite content, and *ACCC has thelowest vitrinite and highest inertinite content. The coals have a low liptinite content (1-2%, raw coal), and is thus composed mainly of the two organic components vitrinite andinertinite as well as mineral matter.115Table 4-2. Alberta suite petrography data.Macerai ACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 *ACCCLL Structured 77 73 49 49 34 34 19 13- VltriniteDesmocollinite 9 9 14 23 19 17 30 22. Vitrodetrinite 2 0 6 3 9 5 3 1Semifusinite 3 5 19 14 23 25 31 34Fusinite 5 3 8 6 11 16 10 3Other Inertinite 2 8 2 3 3 2 5 25E Total Liptinite 2 2 2 2 1 1 2 1- Total Vitrinite 88 83 69 75 63 55 52 37> Total Inertinite 10 16 29 23 37 44 46 62Struct:DegVit 7 8 2 2 1 2 1 1Struct:Deg Inert 4 1 13 7 13 18 8 1Structured 75 73 47 47 33 32 18 13o VitriniteDesmocollinite 8 9 13 23 19 16 30 21Vitrodetrinite 2 0 6 3 9 4 3 1Semifusinite 3 5 18 13 22 24 31 33Fusinite 5 3 7 5 11 15 10 3Other Inertinite 2 8 2 3 3 2 5 25D Total Liptinite 2 2 2 2 1 1 2 1Total Vitrinite 85 82 66 73 61 52 51 36Total Inertinite 10 15 27 22 36 41 46 60Ash Yield (vol.%) 2 1 4 3 2 6 1 3* Structured Vitrinite: Degraded VitriniteStructured Inertinite: Degraded Inertinite116Figure 4-2. Alberta suite petrography data. Samples analysedon a mineral-free (a) and raw-coal (b) basis. Maceraland mineral contents expressed as volume %.Mineral-FreeKEY100806040200(a)24(b)a)EACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 ACCCRaw CoalMineralMatterLiptiniteOtherInertiniteFusiniteSemifusiniteVitrodetriniteDesmocolliniteStructuredVitriniteACCC-27 ACCC-29 ACCC-1 ACCC-5 ACCC-6 ACCC-35 ACCC-13 *ACCC1174.5.2 Isotherms andHysteresis loopsIsotherms, obtained using nitrogen at 77°K, for the Alberta suite, are presented inFigure 4-3. The samples shown are *ACCC and ACCC-27. *ACCC has the highest totalinertinite and lowest total vitrinite content (raw coal basis) and ACCC-27 has lowest totalinertinite and highest total vitrinite. These two samples thus represent the range in organiccomposition of the Alberta suite.The isotherms of the Alberta suite are Type IV, according to the Brunauer,Deming, Deming and Teller (1940) classification. These isotherms are associated withmesoporous solids. A wide hysteresis loop initiates for all samples at relative pressuresbetween 0.4 and 0.5, above the relative pressure at which the first monolayer is believed tobe completed (— 0.3), and closes only at saturation. This hysteresis loop is referred tohere as the high-pressure hysteresis ioop and is coincident with the onset of capillarycondensation in mesopores. In all samples the high-pressure loop, as illustrated in Figure4-3, is a deBoer Type B hysteresis loop which corresponds to slit-shaped pores. The poreshape is believed to correspond to the mesopore shape in the organic fraction of the coals,as very little clay, which might cause a Type B hystersis, is observed in the samples. Asdiscussed earlier, the Alberta suite coals are generally very low in mineral matter content.Gan et al. (1972) have also postulated that fine mineral particulates entrained in the coalmatrix may not be accessible to the nitrogen adsorbate at 77 K and therefore it is unlikelythat mineral matter is affecting the hysteresis loop shape.Some of the sample isotherms display low-pressure hysteresis in which case thehysteresis loop does not close at relative pressures between 0.4 and 0.5 (Figure 4-3).Low-pressure hysteresis described by Gregg and Sing (1982) refers to a lack of closure ofthe high-pressure loop and is thought to be due to swelling of the coal structure or due toadsorption in materials that contain microporosity. Either of these explanations may betrue for the Alberta suite, but since all the samples are microporous, and only some display100..I- (I) I8 6 4 2 0 00.20.40.60.8RELATIVEPRESSUREFigure4-3.Nitrogenisothermsobtainedforsamples*ACCCandACCC-27.00119low-pressure hysteresis, the first cause appears more likely. This problem requires fl.irtherinvestigation.Gas adsorption increases with total inertinite content. The high-pressure hysteresisloop also becomes wider with total inertinite content (Figure 4-3). The total mesoporevolume thus appears to increase with inertinite content. The shape of the high-pressurehysteresis loop is the same for all samples. Because the samples of the Alberta suite varyconsiderably in organic (maceral) composition and all have similar high-pressure hysteresisloop shapes, it is likely that mesopore shape is not affected by the organic composition ofthe coals. A slit-shaped mesopore structure is common to all the coals in this suite.4.5.3 BETandBJH surface areasBET surface areas, Bill surface areas for pores between 2 and 50 nm diameter,and C values are given for the Alberta suite in Table 4-3. The five-point BET surfaceareas were at the relative pressures 0.068, 0.091, 0.14, 0.18, and 0.22. The range ofrelative pressures that the BET equation is applicable is generally assumed to be from 0.05to 0.35, so all calculation points were taken in this range. BET C values are greater than20 (average 68) and thus estimation of monolayer capacities from the BET equation forthe Alberta suite is assumed to be valid (Chapter 2). The BET equation has beensuccessfully applied to other adsorbent-adsorbate systems yielding Type IV isotherms,because monolayer formation on pore walls in mesopores is thought not to be affected byneighbouring surfaces (Gregg and Sing, 1982, p. 168).BETBJH*BJH**BJH**CumulativeBETSampleSurfaceDesorptionAdsorptionAdsorptionAdsorptionCAreaSurfaceSurfaceSurfaceMesoporeValueAreaAreaAreaVolume(sq.mlg)(sq.mlg)(sq.mlg)(sq.m/g)(cclg)ACCC-271.111.451.260.990.0021182ACCC-292.473.272.782.490.0037260ACCC-12.523.363.312.630.0051662ACCC-54.806.796.395.120.0086166ACCC-62.223.002.842.220.0047768ACCC-354.115.524.664.070.0078976ACCC-135.267.226.695.450.0098065*ACCC5.106.916.195.170.0098365*100%poresopenatbothends.**0%poresopenatbothends.Table4-3.BETandBJHsurfaceareasandmesoporevolumesfortheAlbertasuite(rawcoalbasis).121The average value of the BET surface areas of the Alberta suite (— 3.5 m2/g, rawcoal basis) are much lower than the corresponding average carbon dioxide surface areasdetermined from the Dubinin-Radushkevich equation from the 298 K isotherm (- 176m2/g, raw coal basis) (Chapter 3). This is not unique as Gan et al. (1972) also foundnitrogen BET surface areas of less than 1 m2/g for coals that exhibited greater than 200m2/g of carbon dioxide surface area (at 298 K). The reason for the smaller BET surfaceareas is that the BET equation is essentially only determining the internal surface area ofmesopores (and external surface area), whereas the D-R carbon dioxide surface areas areessentially the surface areas associated with microporosity, Gan et al. (1972) noted thatcoals with high carbon dioxide surface areas have smaller nitrogen BET surface areaswhich is also the case here. The cause of this, apart from differences in rank between thecoals, may in part be attributed to compositional variation in the samples.In an attempt to determine the effect of organic composition upon BET surfacearea of the coals, BET surface areas (5-point) versus total and structured vitrinite contentare plotted in Figure 4-4. Plots ofBET surface area versus total inertinite content andsemifhsinite content are also given (Figure 4-5). Results are presented on a raw coal andmineral matter-free basis (Chapter 3). The BET surface areas decrease with structuredand total vitrinite content, although some scatter in the data exists. The sample with thehighest structured vitrinite content (ACCC-27) has lowest BET surface area (1.1m2/g, 5-point, raw coal basis). ACCC-13 has close to the same structured vitrinite content as*ACCC but has a lower mineral-matter content, which may cause the slightly higher BETsurface area (5.3 m2/g, 5-point, raw coal basis) of ACCC-13. There is no apparentcorrelation between surface area and mineral-matter content in these samples; the organiccomponent of the coals appears to be the main control upon the BET surface area.c)C a ‘I) (Ua)(U a C 0) E30 (U C) 0 C!) wo30 6ED): (U (-I) i—1w0 3040506070TotalVitriniteContent(Vol%,MineralMatter-Free)809010d)2030405060StructuredVitriniteContent(Vol%,MineralMatter-Free)70806(U 0 053 U)U)40506070TotalVitriniteContent(Vol%,RawCoal)080900102030405060StructuredVitrinttoContent(Vol%,RawCoal)Figure4-4.Plotsof5-pointBETsurfaceareasversustotalvitrinite(a,b)andstructuredvitrinite(c,d).Mineralmatter-free(a,c)andrawcoal(b,d)valuesareplotted.Mineralmatter-freevaluescalculatedusingtheParrFormula.7080t’36c)a)E C) E4& U) (U w3ci) 0 z Cl)I— w006b)d)(U (I) -1w0oio203040506070TotalInertiniteContent(Vol%,RawCoal)Figure4-5.Plotsof5-pointBETsurfaceareasversustotalinertinite(a,b)andsemifusinitecontent(c,d).Mineralmatter-free(a,c)andrawcoal(b,d)valuesareplotted.Mineralmatter-freevaluescalculatedusingtheParrFormula.10203040TotalInertiniteContent(Vol%,MineralMatterFree)506070SemifusiniteContent(Vol%,MineralMatter-Free)1020SemifusiniteContent(Vol%,RCoal)124Gan et al. (1972) showed that a possible way of determining which branch of theisotherm, adsorption or desorption is best for acquiring pore size distributions from theCranston and Inkley model, which assumes cylindrical pores, is to compare the Cranstonand Inkley adsorption and desorption surface areas to the BET surface areas, becauseBET theory does not assume a geometry for pore shapes. A similar approach is appliedhere for BJH theory. Adsorption and desorption BJH cumulative mesopore surface areasare plotted against BET surface areas in Figure 4-6. In determining the adsorption BJHsurface areas, the percentage of cylindrical pores which are open at both ends wereconsidered; BJH cumulative adsorption surface areas were calculated assuming: 1) that100% of the pores were open at both ends; and 2) 0% of the pores were open at bothends. Good correlations are achieved between the BET surface areas and the BJH surfaceareas (desorption and adsorption). The BiTT adsorption surface areas, with theassumption that 0% of the pores are open at both ends, appear to agree most closely withthe BET surface areas. The adsorption branch should thus be used for pore sizedistribution analyses, but both the adsorption and desorption branch results will bestudied.The BET surface areas probably measure the surface areas of a larger range ofpores than just mesoporosity. In addition, the external surface area of the coal particles ismeasured by BET. The BET surface areas should theoretically be larger than thecumulative surface area of the mesopores measured by the BJH analysis. Five of theeight Alberta samples have BJH cumulative adsorption mesopore surface areas that aregreater than the corresponding BET surface areas. The non-conformity of the mesoporeshapes to that of cylinders is likely the cause of this discrepancy. As indicated by theisotherm hysteresis loops, the pores are probably more slit-shaped than cylindrical.125Plot of BJH cumulative surface area for pores between 2 and 50 nmdiameter versus BET surface area. Plot a) is obtained using the desorptionbranch of the isotherm; b) is obtained using the adsorption branch withthe assumption of 100% pores with both ends open; c) is obtained fromthe adsorption branch with the assumption of 0% pores with both ends open.2 3 4 c 6BET Surface Area (sq. m/g, Raw Coal)a)’7o-50(67b’) 6E.400(6C)U)0C)!Figure 4-6.2 3 4BET Surface Area (sq. mlg, Raw Coal)62 3 4 5BET Surface Area (sq. mlg, Raw Coal)1264.5.4 Mesopore size distributions and volumePlots ofmesopore (2 - 50 nm pore diameter) distributions obtained using the BJHmethod and both adsorption and desorption isotherm branches are given in Figure 4-7.The two samples chosen for these plots are again samples *ACCC and ACCC-27.Sample *ACCC, the sample with the highest inertinite content and lowest vitrinitecontent, has the greatest amount ofmesoporosity (*ACCC). For the adsorption branch,the mesoporosity declines from 2 nm pore diameters to 50 nm. For the desorption branch,mesoporosity declines in a general way from about 3 nm pore diameter to 50 nm, with apeak at about 3 - 3.5 nm. Caution must be exercised in interpreting this peak, however.As mentioned by Gregg and Sing (1982), the surface tension and molar volume of theadsorbate may vary significantly from that of the bulk liquid. In very fine pores, the Kelvinequation, which is the basis ofBJH theory, thus breaks down. The absolute magnitude ofthe 3 nm peak must therefore be viewed with caution. There is, however, a relativeincrease in the 3 nm peak with increase in inertinite content (Figure 4-7).Cumulative pore volume plots, obtained from the adsorption and desorptionbranches of the isotherm, for samples *ACCC and ACCC-27 samples are given in Figure4-8. For the adsorption branch, cumulative pore volumes decrease in a steady fashionfrom 2 nm pore diameters to 50 nm. For the desorption branch, the samples with thehighest inertinite content show a steep inflection at around 3.5 nm (corresponding to thepeak in figure). This inflection decreases in magnitude for the low inertinite contentsamples.127a) 0.001_______0.00051 0.00030.0002aSC6 0.00010.000050.000030.0000250b) 0.0050.0020.0010.0005SC0.0002a00.000100.000050.000020 50Figure 4-7. Pore volume distribution curves for a) adsorptionbranch and b) desorption branch of the isotherm.Samples are *ACCC and ACCC-27.•CC ACCC-27—.- p0 10 20 30PORE DIAMETER (nm)4010 20 30 40PORE DIAMETER (nm)1280.010.0080.0062 0.004a)0.010.0080.006D-IUI20.0020b)2 3 5 10 20 30 50PORE DIAMETER (nm)0.00202 3 5 10 20 30PORE DIAMETER (nm)50Figure 4-8. Cumulative a) adsorption and b) desorption pore volume plotsfor samples *ACCC and ACCC-27.129A plot of total mesopore volume, obtained from integrating the pore volumesfrom 2 to 50 nm, versus vitrinite content is given in Figure 4-9. The adsorption branch ofthe isotherm was used to obtain total pore volumes, but the desorption branch yieldssimilar results. The cumulative adsorption mesopore volume decreases with an increase invitrinite content in a linear fashion (Figure 4-9). Slightly better correlations are achieved ifthe mesopore volumes are plotted against structured vitrinite content. The sample withthe highest structured vitrinite content (ACCC-27) has the lowest cumulative adsorptionmesopore volume (0.0021 cm3/g, raw coal basis) and the sample with the loweststructured inertinite content (*ACCC) has the highest cumulative mesopore volume(0.010 cm3/g, raw coal basis). Plots ofmesopore volume versus semifusinite and totalinertinite content are also given (Figure 4-10).The total amount ofmesoporosity in coals is therefore governed by composition.Coals enriched in vitrinite, in particular structured vitrinite, lack significant mesoporosity.Coals enriched in inertinite have a greater amount ofmesoporosity than vitrinite-rich coalsof the same rank. Mineral matter content varies little in this suite of samples (Table 4-2),and therefore the affect ofmineral matter content cannot be ascertained.4.6 DISCUSSIONComposition, particularly the organic fraction of coal, has an important controlupon the adsorption of nitrogen gas in coals. In particular, the mesopore volume and BETsurface areas using nitrogen gas as an adsorbate are affected by modal abundances of thevarious maceral groups. A decrease in structured vitrinite and coincident increase in totalinertinite leads to an overall increase in mesopore volume and increase in BET surfacearea. This study confirms the Harris and Yust (1979) TEM study which showed thatvitrinite is essentially microporous and inertinite is essentially meso- and macroporous. Asshown here, the mesopore shape changes little with composition.0.012c)40E E0.010 00.0080.0060 0 00.0040.002E C-)0 300.012b°0.012.0.008C) E0.006C) 0 0.0.0040) C) > (U S D C.)0 305060TotalVitriniteContent(Vol%,MineralMatter-Free708090StructuredVitriniteContent(MineralMatter-Free)0.012,_l0I_i,00.010 2.0.008a S0.006a 0 0.0.004C) I..405060708090TotalVitriniteContent(Vot%,RawCoal)Figure4-9.0.72.0102030405060StructuredVitrlnlteContent(VoI%,FlawCoal)70Plotsofcumulativeadsorptionmesoporevolumesversustotalvitrinite(a,b)andstructuredvitrinite(c,d).Mineralmatter-free(a,c)andrawcoal(b,d)valuesareplotted.Mineralmatter-freevaluescalculatedusingtheParrFormula.80C0.012a) E0.0180.00880.0060 0.00.0040.0028 D C)000012b° 1)0.01 2-0.008a, S D0.006a) 0 0. 80.004a> ;o.oo: 01020304050TotalInertinitoContent(Vol%,RawCoal)0.012,-J\0U)00.01C.)2.0.008a, S0.006a) 0 0. 80.004a)6070SemifusiniteContent(Vol%,RawCoal)Figure4-10.Plotsofcumulativeadsorptionmesoporevolumesversustotalinertinite(a,b)andsemifusinitecontent(c,d).Mineralmatter-free(a,c)andrawcoal(b,d)valuesareplotted.Mineralmatter-freevaluescalculatedusingtheParrFormula.1020304050TotalInertiniteContent(Vol%,MineralMatterFree)6070SemifusiniteContent(Vol%,MineralMatter-Free)10203040132The cause(s) for the difference in pore structure of the two maceral groups,vitrinite and inertinite, is unclear, but must certainly include physical, chemical, andbiological affects prior to and during diagenesis. Charring in particular may have animportant control upon the ultimate pore structure of the maceral groups. The mostcommon inertinite maceral subgroup in most of the coal samples is semifusinite. Thesemifhsinite content of the Alberta suite decreases with an increase in structured vitrinite(telinite, telocollinite, and pseudovitrinite submacerals) which is possibly related to the firefrequency in wetlands (Lamberson and Bustin, 1993). The burning or charring ofsemifusinite precursors would lead to an increase in semifusinite and a correspondingdecrease in structured vitrinite. It is possible then that the process ofburning, which leadsto a loss of volatiles, may lead to an increase in mesoporosity. In the previous chapter, itwas indicated that this process may lead to an increase in microporosity of the semifusinitemacerals. Dubinin and Stoeckli (1980) demonstrated that over activated or stronglyactivated carbons possess a more heterogeneous pore structure than that of less activatedcarbons. In particular, supermicropores, pores with diameters between about 1.4 and 3.2nm, are created through the process of activation. No mention was made about the affectupon mesoporosity, however. The cause of the increase in heterogeneity in microporositywith activation was thought to be due to the burning-out of pore walls between adjacentmicropores. It is possible that an analogous process has lead to the increase ofmesoporosity in inertinite macerals, in particular semifusinite, over that of vitrinite.1334.6.1 Relationship between mesopore volume andgasyieldsfrom desorption testsA study by Potter (1993) ofmedium volatile bituminous coals from the MistMountain Formation showed that methane gas yields are greatest for high inertinite coals.The high gas yields of the inertinite rich coal, were thought to be, in part, due to increasedtransmissibility afforded by the presence of open cell lumen in semifusinite and fusinite.Further, Faiz and Cook (1993) found that in situ gas contents, or total desorbed gas frommineral matter-including coal, increased with inertinite content. Conversely, in bothstudies, gas contents were found to decrease with ash content.Gas yields (raw coal basis) were obtained from desorption canister testing of theAlberta suite used in the current study and then plotted against cumulative adsorptionmesopore volumes (raw coal basis) (Figure 4-11). The Smith and Williams UniporeModel (1984) was used to perform lost gas calculations for the canister data. The gasyields appear to increase very generally with mesopore volume.4.6.2 Implicationsfor coalbed methane transmissibilityGamson et al. (1993) classify and discuss the control of microstructures in coals ofthe Permo-Triassic Bowen Basin of Queensland, Australia upon methane transmissibility.These microstructures, which include fracture, matrix, and phyteral porosity have widthsbetween 0.05 and 20 p.m. Gamson et al. also proposed a four-stage model of gastransmission through coal seams: the first stage involves diffi.ision from the microporenetwork; the second involves diffusional and/or laminar flow through the microstucturenetwork which may contain entrained mineral matter; the third stage involves strictly1347-. .. .. . .V 5.5 -G)>-0 5-4.5- I I I I I0 0.002 0.004 0.006 0.008 0.01 0.012Cumulative Mesopore Volume (cclg, Raw Coal)Figure 4-1 1. Plot of gas yields from desorption canister testing versusmesopore volumes (raw coal basis). See text.135laminar flow through ‘open’ microstructures; and the fourth stage includes laminar flowthrough the open cleat system. The microstructure density, orientation and connectivity,shape, size and degree ofmineralization among other factors were suggested to beimportant controls upon diffusional andJor laminar flow through the coal seam on route tothe macrofracture or cleat system. Throughout the study, pores intermediate betweenmicroporosity (< 2 nm pore diameter) and microstructures (.05 - 20 Im in width) wereignored, and it was suggested that diffusional flow ofmethane starts and finishes in themicropore network after which flow is governed by the microstructure network. Thepores that were not included in the four-stage model include the entire realm ofmesoporosity (2 - 50 nm), which, if present, even in minute amounts, must surely have aneffect upon the transmission of methane.The current study has shown that coal composition, particularly the organicfraction, has an important control upon the amount ofmesoporosity. In the Gamsonstudy, microstructures were similarly shown to be controlled by composition, whereby acontinuous microcleat system was associated with bright bands of coal (vitrinite-rich) andphyteral and matrix porosity was associated with dull bands of coal (less vitrinite-rich,more enriched in inertinite and mineral matter). Ertekin (1991) has shown that the timingand magnitude of the first and second coalbed methane production peak is determined byseveral reservoir properties such as coal seam thickness, permeability, sorptioncharacteristics and porosity. The increase in mesoporosity with inertinite content of coals,at least for high-volatile bituminous coals, should then be an additional consideration inmodeling methane gas transmission through coal seams.1364.7 CONCLUSIONSComposition, particularly the organic (maceral) constituents, has been determinedto be an important control upon the mesopore volume and BET surface areas of anisorank coal. The following observations and conclusions have been made:1) Nitrogen adsorption isotherms, determined at 77 K, are all Type IV (Brunauer,Demming, Demming and Teller, classification) for the Alberta suite. Prominent high-pressure hysteresis loops are displayed for all samples. The hysteresis loops are Type B(deBoer classification), which are associated with slit-shaped pores. Low-pressurehysteresis, attributed to pore swelling also occurs for some samples.2) BET surface areas decrease generally with an increase in total and structured vitrinitecontent and conversely increase with an increase in inertinite content. BJH-derivedcumulative surface areas for the mesopore range (2 - 50 nm) are correlative with the BETsurface areas but are generally larger. This is believed to be do to the nonconformity ofthe mesopore shape in these samples to the cylindrical shape assumed in BJH theory. Thepores are probably more slit-shaped, as indicated by the obtained isotherms, whichexplains the discrepancy between the two surface areas.3) Mesopore volumes decrease in a linear fashion with an increase in total vitrinitecontent. A better correlation was achieved with structured vitrinite content. Themesopore volumes conversely increase with total inertinite content. Since the inertinite ismainly semifusinite in these samples, increased mesoporosity associated with an increasesemifusinite might be the result of burning of vitrinite precursors.1374) Gas yields from desorption canister testing appear to increase generally with mesoporevolume.5) Mesopore volumes, which are dependent upon rank and composition, should beconsidered in methane diffusion modeling through coal seams.The importance ofmicrostructures (0.05 - 20 i.tm) in coals in determining methanetransmissibility has been discussed by Gamson et al. (1993). The bulk of these structuresfall into the macropore range of pore sizes. Future studies will be aimed at determiningand quantifying the effect of coal composition upon macroporosity, presumably usingtechniques such as mercury porosimetry.1384.7 REFERENCESBarrett, E.P., Joyner, L.G., and Halenda, P.P., 1951. The determination of pore volume and areadistributions in porous substances. I. Computations from nitrogen isotherms. TheJournal of the Ammerican Chemical Society, 73: 373-380.Brunauer, S., Deming, L.S., Deming, W.S., Teller, E., 1940. On a theory ofvan der waalsadsorption of gases. The Journal of the American Chemical Society, 62: 1723.Brunauer, S., Emmett, P.H., and Teller, E., 1938. Adsorption ofgases in multimolecular layers.The Journal of the American Chemical Society, 60: 309-319.Close, J.C., 1993. Natural fractures in coal. In: B.E. Law and D.D. Rice (Editors),Hydrocarbons from Coal, AAPG Studies in Geology # 38, pp. 119-132.Dubinin, M.M., and Stoeckli, H.F., 1980. Homogeneous and heterogeneous micropore structuresin carbonaceous adsorbents. Journal of Colloidal Interface Science, 75: pp. 34-42.Ertekin, T., Sung, W., and Bilgesu, H.I., 1991. Structural properties of coal that control coalbedmethane production. In: D.C. Peters (Editor), Geology in Coal Resource Utilization.pp. 105-124.Faiz, M.M., and Cook, A.C., 1993. Influence of coal type, rank and depth on the gas retentioncapacity of coals in the southern coalfield, N.S.W.Faiz, M.M., Aziz, N.J., Hutton, A.C., and Jones, B.G., 1992. Porosity and gas sorption capacityof some eastern Australian coals in relation to coal rank and composition. Symposiumon Coalbed Methane Research and Development in Australia, 19-2 1 November, 1992,Townsville 4, 9-20.Gamson, P.D., Beamish, B.B., and Johnson, D.P., 1993. Coal microstructure andmicropermeability and their effects on natural gas recovery. Fuel, 72: 87-99.Gan, H., Nandi, S.P., and Walker, P.L., Jr., 1972. Nature of the porosity in American coals.Fuel, 51: 272-277.Gregg, S.J., and Sing, K.S.W., 1982. Adsorption, Surface Area and Porosity, Second Edition.Academic Press, New York. 303 pp.Harris, L.A., and Yust, C.S., 1976. Transmission electron microscope observations of porosityin coal. Fuel, 55: 233-23 6.Harris, L.A., and Yust, C.S., 1979. Ultrafine structure of coal determined by electronmicroscopy. Preprint paper, American Chemical Society, Division ofFuel Chemicals24: 210-217.139Lamberson, M.N., and Bustin, R.M., 1993. Coalbed methane characteristics of the GatesFormation coals, northeastern British Columbia: effect ofmaceral composition.American Association ofPetroleum Geologists, 77: 2062-2076.Off, C., 1977. Pore size and volume measurement. In: I.M. Kolthofl P.J. Elving, and F.H.Stross (Editors), Treatise on Analytical Chemistry Part III, Volume 4. John Wileyand Sons, New York. pp. 3 59-402.Potter, J., 1993. Claobed methane potential and the effects of coal composition and fracturesin medium-volatile bituminous coals from the Mist Mountain Formation, southwesternAlberta (abstract). Geological Association of Canada/Minerological Association ofCanada, Joint Annual Meeting, Program and Abstracts, p. A-84.1980 Annual Book ofASTM Standards. Part 26 Gaseous Fuels, Coal and Coke, section D38877.140CHAPTER 5VARIATION IN PRESSURE-DECAY PROFILE PERMEAMETERDERIVED PERMEABILITIES WITH LITHOTYPE AND MACERALCOMPOSITION OF COALS5.1 ABSTRACTCoal beds are markedly heterogeneous with respect to composition and fabric,which imparts significant vertical and lateral variation in permeability, and thus may beimportant in making production decisions in the extraction of hydrocarbons from coal.The current study, utilizing a pressure-decay permeameter, quantifies changes inpermeability of coal at the lithotype (megascopic) and maceral (microscopic) scale. Theorder of decreasing permeability with lithotype is: bright > banded > fibrous > banded dull> dull, for the coal samples used. Bright coals are the most permeable because ofassociated macro- (cleat) fracturing. For a banded dull sample, permeability generallyincreased with increasing vitrinite content. The lowest permeabilities measured occur indull coals with a high mineral and inertinite content. Fibrous coal has a higherpermeability than dull coal of the same rank due to the abundance ofmacroporous fusinitein the former. Dull coal permeability decreases with an increase in rank, but these resultsare obscured by compositional variability between samples. Pressure-decay measurementsare more reliable for dull lithotypes as these lithotypes do not fracture as easily duringsample preparation. In addition, measured permeabilities are optimistic due to therelaxation of stress upon exposure of coal to atmospheric pressure.1415.2 INTRODUCTION AND RESEARCH OBJECTIVESPermeability is an important parameter in the prediction of reservoir performance.In conventional reservoirs, the average permeability and permeability heterogeneitycontrol production rate and efficiency, respectively (Georgi et a!., 1993).Coal beds, which are unconventional hydrocarbon reservoirs, are typicallyheterogeneous with respect to composition. An understanding of the effect ofbothmegascopic (lithotype) and microscopic (macera!) composition upon permeability is thuscritical in making completion and production decisions in the extraction of hydrocarbonsfrom coal.Among the most important factors affecting permeability in coalbeds is the fracturesystem which, in turn, is largely controlled by composition. At several producing regionsof the San Juan Basin, for example, fracture permeability is considered to be the singlegreatest control upon production (Close et al., 1990).The current study attempts to quantify the change in permeabilities of coal withlithotype (megascopic) and maceral (microscopic) composition. In addition, the change inpermeabilities with coal rank are documented. Permeabilities are measured using a newtype of permeameter, referred to as a Pressure-Decay Profile PermeameterTht (PDPK -300, patent pending, Jones, 1992). The device can measure permeabilities on a bed-bybed scale, and may thus be used to document permeability variations on the lithotype scalein coal. A detailed account of the permeability variation in the dull components of coalmay be important in the accurate prediction of gas producibility in seams rich in dull coal.1425.3 EFFECT OF COAL STRUCTURE ON PERMEABILITY5.3.1 Cleat systemsCleats are (natural) fractures in coal which are formed through a variety ofdifferent processes including dessication, coalification, lithification, and paleotectonicstress (i.e. Close, 1993). Typically cleat comprises two (usually mutually orthoganal) sets:the continuous face cleat and the less continuous butt cleat which terminates into the facecleat. These two sets are generally perpendicular, or nearly so, to bedding in the coal. Insome coals a third cleat set is developed which is also perpendicular to bedding but whichis curviplanar and intersects the face and butt cleat (Gamson et al., 1993).The cleat system is important in controlling gas production in that the cleat systemis the principal permeability pathway for water and gas during production(depressurization). The most important properties of the cleat system that affect coalpermeability are; cleat spacing and height, aperture width, connectivity, and degree ofinfilling and closure. Cleat spacing and height are affected by lithotype thickness. Cleatsare generally most abundant in bright bands of coal and their height is mainly restricted bythe widths ofbright bands, Cleat spacing and height appear to decrease with decreasinglithotype thickness (Close, 1993). In addition, rank has an effect upon cleat spacing(Close, 1993)1435.3.2 MicrostructuresIn addition to the megascopic fractures, coal seams possess microstructures whichcontribute to the overall permeability of the seam. Microstructures include microfracturesand cavities which are micrometre in scale (0.05 - 20 urn in width). Gamson et al. (1993)state that porosity represented by the microstructure system is of three types, includingfracture, phyteral, and matrix porosity, and that the porosity lies within the realm ofmesoto macroporosity. Figure 5-1 illustrates the microfractures of coal. Because of thecontinuity of the microstructure system of coals, the microstructures are thought to be animportant control upon gas transmissibility.In the Gamson et al. (1993) study, it was assumed that the microstructuredistribution as well as size, shape, and continuity is affected by coal lithotype. Thephyteral and matrix porosity is generally associated with duller coals whereas fractureporosity is more typical of brighter lithotypes. The microcleat system in brighter coalsoften forms a continuous network with the larger cleats.The transmissibility ofmethane through a coal seam is dependant upon themegascopic and microscopic fracture and pore systems and their degree of connectivity.The orientation, continuity, and density of these structures, in addition to the coal rankand composition (organic and mineral) are important considerations in the production ofcoalbed gas.144Butt CleatFace CleatBlockyFracturesFigure 5-1. Diagram illustrating microstructures in coal. (a) relationshipbetween face and butt cleat; (b) various microstructures inbright coal and their relationship with the larger cleat; (c)cell lumen in dull coal. Modified from Gamson et al., 1993.A. Face Cleat B. VerticalmicrocleatHorizontalmicrocleatButt CleatConchoidalfracture1455.4 THE PRESSURE-DECAY PROFiLE PERMEAMETERThe PDPK - 300 is a steady-state permeameter developed by Core Laboratorieswhich is capable ofmeasuring permeabilities in the range of 0.001 to 20,000 md within 2to 35 seconds ofmeasurement time (Jones, 1992; Georgi and Jones, 1992). Lowpermeability samples require longer measurement times (i.e. 30 seconds for .001 md rock).Measurements can be corrected for gas-slippage (Klinkenburg) and inertial flow resistanceeffects (Georgi and Jones, 1992).The instrument consists of a manifold and probe (Figure 5-2) which togethercomprise four volume-calibrated tanks of varying volumes (Georgi and Jones, 1992).Nitrogen gas is bled into one of the chambers and then injected through a probe tip, whichis flush against the flat surface of the sample, and into the sample. A practise blowdown isperformed prior to the actual measurement in order to determine which manifold volumeto use for the sample point. A larger volume is chosen for high permeability samples, andthe probe tip volume itself maybe used for very low permeability samples.During the measurement, once the gas is bled through the probe tip and into thesample, the pressure decay time is monitored. From this, the instantaneous volumetricflow rate is obtained and, through the use of the Forchheimer equation, permeabilities andinertial resistivity coefficients () are calculated (Georgi et al., 1993; Jones, 1992). Dataprecision is about ± 2%. The rate of change of pressure with time is a reflection of thepermeability of the sample; the higher the permeability of the sample, the greater the rateof change of pressure with time.The probe tip seal may be changed according to the depth of sample that is to beinvestigated; the smaller the seal, the shallower the depth of penetration. For small coreplug measurements, 0-rings (— 5 mm in diameter) maybe used.LaserPositionIndicatorProbeOperatorPressureTransducerF1’ :•H_______I————iH’•.111111___Ht-J—tcHI___1UU_______HL——Ill1tIiiiJEE1T11!111L“.‘:.:.:_ProbeTip.SlabbedWhole-CoreSample100pslg25-4Pr.esFigure5-2.SchematicdiagramillustratingthePressure-DecayProfilePermeameter(afterGeorgietal.,1993).147The PDPK-300 may be used to obtain closely spaced rapid permeabilitydeterminations. Such an instrument is ideally suited to obtain a lithotype permeabilityprofile for coals.5.5 METHODSSeven coal samples were used in this study. Four samples were obtained from theUpper Jurassic-Lower Cretaceous Mist Mountain Formation; one (LC-8) from the southElk Valley coalfields (SEVC) and three from the north Elk Valley coalfields (NEVC) ofsoutheastern British Columbia. Three samples were also obtained from the LowerCretaceous Gates Formation of northeastern British Columbia.The coal samples were cut into rectangular blocks with a water-lubricated diamondrocksaw. Care was taken to make sure the cut surfaces were flat and as free fromirregularities as possible.Pressure-decay profile permeabilities were measured at Core Laboratories inCalgary, Alberta. The instrument used was a PDPK-300. Profiles parallel andperpendicular to face cleat (if present) were performed for each sample. Sample pointswere spaced at least one centimeter apart. Portions of the coal surface that exhibitedsurface irregularities or artificially (sawcut)-induced fracturing were avoided. For allsamples, with the exception ofLTC-1 1, a Gates Formation sample, points were taken onat least two cut surfaces. Profiles were taken on all four cut surfaces of the SEVC (LC-8)sample, the largest sample of the seven used.Sample points were located with the use of a laser sight. A shot was then firedwhereby the probe tip was neumatically projected against the coal surface at a pressure ofabout 173 kPa (25 Psi). A practise blowdown was performed to determine whichcalibrated volume was to be used in the analysis. The reservoir chosen was then filledwith nitrogen gas to a pressure of about 69 kPa (10 Psi) and computer-operated valves148opened to bleed gas into the sample. Pressure-decay with time was then recorded tocalculate sample permeabilities. Both slip-corrected (liquid-equivalent) and conventionalpermeabilities were calculated. Measurement times were generally less than 33 secondsand varied depending on the permeability of the sample. An 0-ring probe tip with adiameter of 0.5 cm was used in the analyses.Lithotype descriptions of the samples were performed following standardconventions (Diessel, 1965; Marchioni, 1980; Lamberson and Bustin, 1993). The coalsurface within about 0.5 cm of the sample measurement site were observed using abinocular petrographic microscope at 60 X magnification to determine if any irregularitiesor microfractures existed.A representative sample of each permeability point was obtained for petrographicanalysis. Cubes of about 0.125 cm3 of coal were cut, with the measured point at thecenter of the top face, using a gem saw. With an 0-ring seal of 0.5 cm diameter on theprobe tip, the depth of measurement was about 0.5 cm, thus 0.125 cm3 is believed to berepresentative of the volume measured in the permeability analyses. About 90% of thepoints measured were recovered during the cube-cutting procedure.The cubes of representative sample were then crushed to less than 250 p.m screensize and made into 2.54 cm pellets for standard petrographic analysis. Standardpetrographic analysis was then performed for each point (Chapter 3). Because very littlesample was utilized for each pellet, some samples had to be discarded due to the loss ofcoal during the polishing procedure.Random (vitrinite) reflectances were also performed for each coal sample usingstandard techniques (Bustin et al., 1985). Mean random reflectances (R0)for at least twomeasured permeability points of the hand sample were obtained, with a minimum of 25reflectance measurements per pellet.1495.6 RESULTS5.6.1 Lithotype, Megascopic Structure, andMeasurement Surface DescriptionsPhotos of all coal sample measurement surfaces and points are shown in Figures5-3 to 5-11. In addition, lithotypes are labeled for samples with permeability profiles. Thelithotype classification used in this study is a modification of the Australian classificationsystem (Diessel, 1965; Marchioni, 1980; Lamberson and Bustin, 1993) (Table 5-1).The SEVCF sample (LC-8) is shown in Figures 5-3 to 5-5. All four cut surfacesare displayed (Figure 5-4 and 5-5) plus the two uncut upper and lower surfaces (Figure 5-3). LC-8 contains a banded dull segment, an upper bright band (—‘ 1.5 cm thick) and alower bright band (- 1cm thick) (Figure 5-4). The upper bright band surface (Figure 5-3),which is parallel to bedding, is sheared (parallel to bedding) and has a prominant face cleatwith a regular spacing of about 1 - 2 mm. The lower bright band surface also displaysregular face cleating with a 1 - 2 mm spacing. The measurement surfaces of sample LC-8are both perpendicular (faces lA-i and 1A-2) and parallel (faces lB-i and 1B-2) to theface cleat of the upper and lower bright bands (Figure 5-4 and 5-5). The banded dullsegment of faces A and B has minor pitting associated with very thin bright bands (<1mm) and some small fractures associated with cutting, which were avoided duringmeasurement. The upper bright band has large pits and fractures which are due to thebrittleness of bright coal. The lB-i and 1B-2 faces are slightly more pitted than the Afaces possibly due to the fact they are parallel to face cleat.Sample 2, a NEVCF coal is a dull coal (Figure 5-6). This sample has a largeamount of artificially-induced fracturing and pitting, and thus there are limited number ofpoints measured on this sample.150Table 5-1. Lithotype classification used in current study. Modifiedfrom Lamberson and Bustin (1993).Stopes-Heerlen Nomenclature Description(ICCP) used in this studyClassificationvitrain bright coal (B) subvitreous to vitreous lustre,conchoidal fracture, less than10% dullbanded bright coal predominantly bright coal,(B B) 10-40% dullclairain banded coal (B C) interbedded dull and bright inapproximately equal proportionsbanded dull coal predominantly dull coal, 10-40%(B D) brightdurain dull coal (D) matte lustre, uneven fracture, lessthan 10% bright coal, hardfusain fibrous (F) satin lustre, friable, sooty to touch151a)b)Figure 5-3. Sample 1 showing top (a) and bottom (b) faces.a) b)B BD BB BDF---.).0010.0030.010.030.10.313Permeability(md)B).0010.0030.010.030.10.3Permeability(md)Figure5-4.PermeabilityprofilesforfacesiA-i(a)and1A-2(b).Seetextforexplanation._7I— vi3a) b)).0010.0030.010.030.10.3Permeability(md)33VII..)10010.0030.010.000.10.3Permeability(md)Figure5-5.Permeabilityprofilesforfaces1B-i(a)and1B-2(b).‘1C01(-I)p)CDCl,0DCDCD0p)p)DaC)C-‘CDj;:.3:1N33c3-155Sample 3, from NEVCF, is a banded to banded dull coal (Figure 5-7). The topface of the sample (not shown) displays face cleating with a spacing of-- 2 mm. Face 3Ahas a bright band with a thickness of about 5 mm at the bottom of the face. Allmeasurement points are located in the banded dull portion of the coal. Both faces displaya large amount of artificially-induced fracturing and pitting.Sample 4 (LTC-2), from the Gates Formation, is a dull coal with a 2 mm thickbright band at the top of the sample (Figure 5-8). Face cleat is visible on the top andbottom surfaces (not shown) of the sample and has a spacing of- 1 mm. All points arelocated within the dull portion of the coal and each surface is essentially free from pittingand fracturing. Some very fine laminations (fibrous coal or mineral) occur within the dullsection.Sample 5 (Figure 5-9), from NEVCF is similar to sample 3 but is slightly brighter.A 6 mm thick bright band occurs at the bottom of face 5A and a 4 mm thick bright bandoccurs at the top. Face cleat in the top and bottom bright bands has a spacing of about 1 -2 mm and is oriented at high angle (not quite orthogonal) to face 5A. Sample 5, likesample 3, has alot of artificially-induced fracturing and pitting in the surfaces.Sample 6 (LTC-5), from the Gates Formation, is a fibrous coal (silky luster)(Figure 5-10). The sample displays no cleat. Some banding does occur in the sample,which may be attributed to fire cycles.Sample 7 (LTC-11), also from the Gates Formation, is a banded dull - dull coalwith poorly developed face cleating in the thin bright bands (Figure 5-1 1).156C)CC)C’,a)C.)CCU)C)aECl)LC)Dc,)LLCENT TRE.0001.0001.001•1•Permeability(md)Figure5-8.Permeabilityprofilesoffaces4A(a)and4B(b).Seetextforexplanation.a) b).0001.001.01Permeability(md)I.-.‘S.-1,—ICD(1(J)aCDC,’Cl)INT1MEtIc3-sci159Figure 5-10. Sample 6 showing all faces on which points were measured.160I I.01 .02 .03 .05 .07Permeabibty (md)BD-DFigure 5-11. Permeability profile of sample 7. See text for explanation.1615.6.2 Permeability Variation with Lithotype CompositionIn order to determine the effect of lithotype (megascopic) composition of coalupon measured permeabilities, (vertical) profiles were taken on some of the coal samplesat right angles to bedding. The spacing ofmeasured points was generally around 1 cm,but because of surface irregularities such as pitting or artificially-induced microfracturing,spacing varied. Both non-slip corrected permeabilities (Ka) (solid circles in profile) andliquid equivalent permeabilities (Kl) (open circles in profile) are given.Vertical profiles are plotted adjacent to coal photos in Figures 5-4, 5-5, 5-8, and 5-11 for coal samples 1, 4, and 7. The other samples were not chosen, as it was difficult toobtain vertical profiles due the irregularities on the measurement surfaces.Profiles of faces lA-i and 1A-2 (cut perpendicular to face cleat) are plotted inFigure 5-4. The profile of face lA-i shows a slight increase in permeabilities from thebottom to the top of the sample. Permeabilties in the bright bands at the top and bottomare higher than those in the banded dull lithotype. The permeabilities of face 1A-2 show asimilar trend as in face lA-i; the highest permeability is for point 11, located in the lowerbright band, and the permeabilities in the banded dull lithotype are considerably lower.Profiles of faces lB-i and 1B-2 (cut parallel to face cleat) are shown in Figure 5-5.The profiles show similar trends as the 1A faces. The highest permeabilities are associatedwith bright coal. The banded dull band of face lB-i is remarkably uniform inpermeability, and is considerably tighter than the bright bands. The dull banded band offace 1B-2 a greater variability in permeability than face lB-i.162Profile permeabilities for sample 4 are shown in Figure 5-8. The permeabilitieswithin the dull band of faces A and B are fairly uniform and are quite low (Kl generally <0.02 md). The slight variation in permeabilities may be due to compositional variability,but the dull band appears to be quite uniform. Point 2 of face B has a very lowpermeability (Ki = .00006 md), but there is no visual compositional difference betweenthis point and the rest of the dull band.A permeability profile of sample 7 is presented in Figure 5-11. The Kl permeabilityis less than 0.02 md and is quite uniform throughout the profile.In general, the brighter lithotypes have the greatest permeabilities. The brightbands of sample 1 have the highest average permeability (average = 4.1 md, range = 2 - 7md) of the samples, and the dull bands of samp’es 4 and 7 have the lowest averagepermeabilities (average = .016 md, range = 0..00006 - .12 md). Sample 2 is a dulllithotype and has permeabilities ranging from 0.03 to 1 md (average = .13 md), which ishigher than for samples 4 and 7, but this may be attributable to the high amount of pittingand artificial fracturing in this sample. The banded dull band of sample 1 has a range inliquid permeabilities from .01 - 1.5 md (average = .14 md) which is intermediate inpermeability to the dull and bright lithotypes. The banded to banded dull coals (3 and 5)have permeabilities ranging from .07 - 4 md (average = .79 md). The range in KI for thefibrous sample (6) is 0.2 - 1 md (average = .5 md). In the above averages, points that arenot thought to be representative of the sample and are anomolously high, due to surfacepitting, fracturing, proximity to an edge, or loss of seal, are not used. The followingrelationship of decreasing permeabilities with lithotype thus occurs for the samples in thisstudy: bright> banded > fibrous> banded dull > dull.1635.6.3 Permeability Variation with Macera! CompositionPetrography was performed upon 0.125 cm3 volumes of coal for each point in thepermeability analysis. The results are presented in the Appendix.Samples of similar rank are grouped in the following analyses in order to eliminatethe effects of rank upon measured profile permeabilities. The sample 1 (LC-8) has thehighest rank, with a random vitrinite reflectance of 1.25 % (medium-volatile bituminous),and will be considered seperately. Samples 2, 3 and 5 have random reflectances of 0.90,0.91, and 0.92 %, respectively, are of similar rank (high-volatile bituminous A) and will bediscussed together. Samples 4, 6 and 7 have random reflectances of 1.08, 1.03, and 1.06%, respectively, are of similar rank (high-volatile bituminous A) and will be consideredtogether.Sample 1 (LC-8)A plot of total vitrinite versus liquid-equivalent profile permeability is shown inFigure 5-12. Both mineral-matter free (mmf) and raw (mineral-matter including) vitrinitecontents are plotted.The profile permeabilities appear to increase in a general way with total vitrinitecontent and the two parameters can be correlated linearly.No apparent correlation occurs between the permeability and mineral mattercontent. This may in part be due to the way in which mineral matter content wasdetermined. The mineral matter content was obtained petrographically through visualpoint count, instead of through ASTM proximate and sulphur analysis and the Parrformula (Chapters 3 and 4), which is believed to be a more reliable method for mineralmatter content determination. Not enough material was recovered for each point to164. d.• .0.3-. ... .0.03-0.01- • S0.003 -0.001- I I I20 40 60 80 100Total Vitrinite Content (volume %, mmf)b).• .0.3-..0.1- ••E0.03-0.01- • •• •0.003 -0.001- I I20 40 60 80 100Total Vitrinite Content (volume %, raw coal basis)Figure 5-12. Plots of KI versus total vitrinite content of sample 1 on a;a) mineral matter-free, and ; b) raw coal basis.165facilitate the use of the ASTM method. A bulk sample mineral matter analysis was notthought to be relative because it would not have been representative of the individualvolumes measured in the profile permeability analysis. Mineral matter content was quitelow in this sample (0 - 7 %) and thus does not appear to be an important controllingfactor on the permeability measurements.Scatter in the plot ofFigure 5-12 is attributable to several factors. Firstly,artificially-induced (sawcut) fractures or pitting near or at the point may cause variationsin the measured permeabilities for points of similar composition. Secondly, the samplevolume cut for petrographic analysis may not be completely representative of the volumemeasured either compositionally or structurally. Thirdly, some sample points that have ahigh vitrinite content but low permeabilities may lack macroscopic fracturing which wouldlead to higher permeabilities. Although great care was taken to avoid any areas of samplewith artificial fracturing or choose points where the coal appeared homogeneous incomposition, this task proved to be very difficult. The permeability data presented heremust be interpreted with care.Samples 3 and 5 (SEVCF)Samples 3 and 5 are grouped together because of similar reflectance values (rank).Sample 2, although of similar rank is not used in this analysis because it is highly fractured(artificially-induced).No significant correlation is obtained between maceral composition andpermeability. This is due mainly to the highly fractured nature of samples 3 and 5 (Figures5-6 and 5-8). The fracturing is caused by the frequent occurrence of brittle bright bandswhich “fall apart” during the cutting procedure. Even the duller material between brightbands is fractured, and the obtained permeabilities are suspect.166Samples 4, 6 and 7 (Gates Formation)Samples 4 and 7 are dull and dull to banded dull coals, respectively, and sample 6is a fibrous coal. Of these samples, sample 4 has the lowest average permeability andsample 6 has the highest.Samples 4 and 7 are similar in appearance and texture. Sample 4 has an averagetotal vitrinite content (volume percent, mmf) of 21 % and sample 7 an average vitrinitecontent of 23% (mmf), which is consistent with the latter sample’s higher averagepermeability. The high average mineral matter content for samples 4 and 7 of 7.4 and 5.0volume %, respectively, may also be a cause of low permeability.Sample 6 is unique in that it has a very high fhsinite content (69 %, mmf). Theaverage permeability of this sample is higher than samples 4 and 7 despite the high averagemineral matter content of sample 6 (7.5 volume %).5.6.4 Effect ofRank upon Profile PermeabilityAlthough it is difficult to determine the effect of rank upon permeability of thechosen coal suite because of compositional variability, a few general statements can bemade. Typical banded dull permeabilities of sample 1, the highest rank coal, are between0.01 - 0.02 md, and are often less than 0.01 md. The dull band permeabilities of samples2, 3, and 5, which are of lower rank are never less than about 0.03 md. It appears thatdull coal permeabilities decrease slightly with rank, but the above result is not conclusive.Pore size distribution studies of coal of varying rank (Gan et al., 1972) have shown thatmacroporosity generally decreases with increasing coal rank such that medium volatilecoals (ie. sample 6) typically contain less macroporosity than a high volatile A bituminouscoal (ie. samples 2, 3, and 5). Because macroporosity may be an important contributor to167permeability, it is consistant that sample 1 has lower dull coal permeabilities than samples2, 3, and 5.5.7 DISCUSSIONPermeabilities vary with lithotype composition of coals. The order of decreasingpermeabilities of the lithotypes is as follows: bright> banded > fibrous > banded dull>dull.The abundance and orientation of macroscopic fracturing as well as the type, size,shape, density, orientation, distribution, and connectivity ofmicrostructures is dependantupon lithotype. The well-defined macroscopic fracturing (cleating) of bright coal probablycontribute to its high permeability. Too few measurements were taken on bright bands todetermine the effect cleat orientation, however. Bright coals also have a continuousmicrocleat network which may contribute to the overall permeability. Duller lithotypes,on the other hand, typically lack macroscopic fracturing and have a greater abundance ofphyteral porosity, such as cell lumens in füsinite.On the microscopic level, an increase in total vitrinite content of sample 1correlates with an increase in permeability. The thick bright bands have a very highvitrinite content and associated permeabilities. In the banded dull lithotype, variation invitrinite content is probably due to the presence or absence of thin bright bands in thesampled volume used for petrographic analysis.Permeabilities are generally less variable for samples 4 and 7 than for sample 1,which is due to the relative compositional homogeneity of the former samples. The lackofmacrofracturing associated with these dull lithotypes is a cause of their overall lowpermeabilities. In both samples 4 and 7, high mineral matter contents are an additionalcause of low permeabilities. Sample 6 has a higher permeability than samples 4 and 7,which may be caused by the abundance of fhsinite, some ofwhich has open cell lumen.168Bright coals may provide the most permeable pathways for methane transmission.As discussed in Chapter 3, high rank coals rich in vitrinite tend to have a greater amountofmicroporosity than those rich in inertinite. This leads to a higher capacity for methanegas storage. The microcleat network, along with conchoidal fracturing and striae, maycreate a continous permeable pathway from the micropore network through to themacrofracture network. Brighter lithotypes thus not only provide a high gas storagecapacity for high rank coals, but also have the potential to provide continuoustransmission ofmethane gas to the borehole.Although profile permeabilties reflect to a certain degree the compositionalvariation of the samples measured, caution must be taken when interpreting the data if asubstantial amount of fracturing induced by sampling, core slabbing, or sample cutting ispresent. The duller lithotypes are less prone to artificial fracturing and the profilepermeameter technique appears to be most useful to the study of such lithotypes. Samples3 and 5 include highly fractured bright bands and hence the measurements taken on thesesamples are suspect.It is also important to note that the samples utilitized in this study were sampledfrom outcrop and hence do not represent subcrop coal permeabilities. The resulting stressrelaxation from surface exposure (McElhiney et al., 1993) and brittleness during cuttingprocedures probably accounts for higher permeabilities of the coal than would be obtainedduring well testing.1695.8 CONCLUSIONSA pressure-decay permeameter is useful in determining variation in measuredpermeabilities with lithotype and maceral composition of coals of the Western CanadianSedimentary Basin. The technique is reliable for dull lithotypes, which due not fracture aseasily during core slabbing, sample cutting or handling techniques. The followingconclusions are obtained from this study:1) The order of decreasing profile permeabilities with lithotype is as follows: Bright>banded > fibrous > banded dull > dull. The increased permeabilities with increasedbrightness of the coals is due to the presence of abundant macrofracturing (cleating) inbright coals.2) A general increase in permeability is associated with an increase in total vitrinite. Thehighest permeabilities are associated with bright bands with a high vitrinite content.3) The lowest permeability dull coals lack macroscopic fracturing and have a high mineraland inertinite content. Fibrous coal has a higher permeability than dull coal of the samerank. This is possibly due to the high content of fusinite in fibrous coal, which is highlymacroporous.5) Dull coal permeabilities appear to decrease with an increase in rank. These results areobscured by compositional differences between samples of the same rank.1706) Pressure-decay permeability measurements are more reliable for dull lithotypes.7) Measured permeabilites are optimistic due to the relaxation of stress upon exposure tothe atmosphere which opens up fractures in the sample.If care is taken in measuring profile permeabilities, the technique may provide avaluable method of predicting methane recoverability from coals ofvarious lithotypecompositions and rank.1715.9 REFERENCESClose, J.C., 1993. Natural fractures in coal. In: B.E. Law and D.D. Rice (Editors),Hydrocarbons from Coal, AAPG Studies in Geology # 38, PP. 119-132.Close, J.C., Mayor, M.J., and McBane, R.A., 1990. Importance, genesis, andrecognition of fracture permeability in Fruitland coalbed methane reservoirsof the northern San Juan Basin, Colorado and New Mexico. PetroleumSociety of CIM/Society ofPetroleum Engineers International Technical Meeting,Calgary, Alberta, Canada, Paper no. CTMJSPE 90-106.Diessel, C.F.K., 1965. Correlation ofmacro- and micropetrography ofNew SouthWales coals. In: J.T. Woodcock, R.T. Madigan and R.G. Thomas (Editors),Proceedings-General, 8th Commonwealth Mineralogy and Metallurgy Congress,Melbourne, 6: 669-677.Gamson, P.D., Beamish, B.B., and Johnson, D.P., 1993. Coal microstructure andtheir effects on natural gas recovery. Fuel, 72: 87-99.Gan, H., Nandi, S.P., and Walker, P.L., Jr., 1972. Nature of porosity in Americancoals. Fuel, 51: 272-277.Georgi, D.T., and Jones, S.C., 1992. Application of pressure-decay profile permeametryto reservoir description. Society ofPetroleum Engineers, 9212: 1-12.Jones, S.C., 1992. The profile permeameter: a new, fast, accurate minipermeameter.Society ofPetroleum Engineers, 24757: 973-983.Lamberson, M.N., and Bustin, R.M., 1993. Coalbed methane characteristics of the GatesFormation coals, northeastern British Columbia: effect of maceral composition.American Association ofPetroleum Geologists, 77: 2062-2076.Marchioni, D.L., 1980. Petrography and depositional environment of the Liddell Seam,Upper Hunter Valley, New South Wales. International Journal of Coal Geology,1: 35-61.McElhiney, J.E., Paul, G.W., Young, G.B.C., and McCartney, J.A. Reservoir engineeringaspects of coalbed methane. In BE. Law and D.D. Rice (Editors), Hydrocarbonsfrom coal, AAPG Studies in Geology # 38: pp. 361-372.1980 Annual Book ofASTM Standards. Part 26 Gaseous Fuels, Coal and Coke,sections D3 174-73 and D3 177-75.172CHAPTER 6SUMMARY AND CONCLUSIONS6.1 EFFECT OF COAL COMPOSITION UPON GAS SORPTION CAPACITY ANDTRANSMISSIBILITYCoal is a compositionally complex material containing both organic and inorganicconstituents. The proportion of these two constituents can vary widely within a seam. Coal typeultimately controls its utilization potential and hence lateral and vertical variations in coal seamsshould be accounted for in exploration and development strategies.Similarly, coal composition variability should be considered in exploration programs fornatural gas from coal seams. The current thesis has demonstrated that coal gas capacity is affected bymaceral and mineral content. Specifically, gas capacity of bituminous coals increases with vitrinitecontent. Conversely, gas capacity generally decreases with increasing inertinite and mineral mattercontent. The sorption capacity of coal is a function of the pore size distribution which is in turnaffected by maceral content: vitrinite is more microporous than inertinite whereas inertinite has agreater amount ofmesoporosity. The ultimate gas content of coal is hence intimately related to therelative proportion of the maceral groups.The permeability of coal is greatly affected by lithotype (megascopic) and maceral(microscopic) composition. Brighter coals typically have a higher permeability due to abundance ofassociated macro- (cleat) fracturing. Permeability also generally increases with vitrinite content,although further work is required to confirm this.Ignoring rank and other factors affecting gas content, a bright coal with low mineral mattercontent and abundant macrofracturing should have a high gas content and transmissibility.1736.2 FUTURE WORKThe current study has focused upon the controls of coal composition upon single componentgas sorption. Previous studies (Greaves et al., 1993; Harpalani and Pariti, 1993) have alreadydocumented the relative sorption capacities of single and multicomponent gases. The effect ofmacera! composition upon the sorption ofmixtures require study.The effect of coal composition upon gas permeability requires further investigation. Thepressure-decay permeametry technique in conjunction with more conventional laboratory techniquesshould yield further insight into the effects of both maceral and mineral composition upon gastransmission.1746.3 REFERENCESGreaves, K.H., Owen, L.B., McLennan, J.D., and Olszewski, A., 1993. Multi-component gasadsorption-desorption behaviour of coal. In: Proceedings of the 1993 InternationalCoalbed Methane Symposium, The University of Alabama/Tuscaloosa, May 17-21, 1993:197-205.Harpalani, S., and Pariti, U.M., 1993. Study of coal sorption isotherms using a multicomponent gasmixture. In: Proceedings of the 1993 International Coalbed Methane Symposium, TheUniversity ofAlabama/Tuscaloosa, May 17-21, 1993: 151-156.cU176AppendixPetrography dataMineral matter-free data - volume %SAMPLE POINT# SV DESMO VITDET SFUS FUS OTHERI TOTLIP TOTVIT TOTINlA-i 44 6 0 4 0 46 0 50 501A-2 27 12 0 21 0 40 0 39 611A-3 38 13 0 8 0 41 0 51 491A-4A 30 9 0 7 0 54 0 39 611A-5 23 19 0 21 0 38 0 42 581A-6 29 4 0 13 0 53 0 33 671A-7 56 1 0 40 1 2 0 56 431A-8 99 0 0 0 0 1 0 99 11A-8A 99 0 0 1 0 0 0 99 11A-9 58 10 0 6 0 26 0 68 32lA-li 60 9 0 2 0 29 0 69 311A-12 17 6 0 33 0 44 0 23 771A-13 28 2 0 13 0 58 0 29 711A-14 36 8 0 6 0 51 0 44 561A-15 23 7 0 11 0 60 0 30 701A-16 24 3 0 12 0 61 0 27 73lB-i 37 15 0 16 1 30 0 53 4718-2 29 17 0 12 0 43 0 45 55iB-3 30 9 0 22 0 38 0 39 611B-5 29 10 0 20 0 40 0 40 601B-6 32 18 0 27 0 24 0 50 50lB-b 29 10 0 12 0 49 1 39 611B-12 31 22 0 15 0 33 0 53 471B-13 22 4 0 20 0 55 0 26 741B-14 51 9 0 15 1 25 0 59 412 2A-2 23 7 0 17 6 47 1 30 692A-3 15 2 0 31 10 42 1 17 822A-4 23 5 0 23 0 46 3 28 692B-1 29 11 0 14 2 40 4 41 552B-3 22 7 0 12 5 53 1 29 712B-5 26 3 0 36 0 35 1 29 703 38-1 17 10 0 22 2 50 0 26 743B-2 14 6 0 24 2 50 3 20 773B-3 22 7 0 39 0 30 1 30 693B-5 7 2 0 63 0 27 0 9 9038-6 26 8 0 32 0 34 1 34 653A-1 23 14 0 25 0 38 0 36 633A-2 23 2 0 30 0 45 0 25 753A-3 ii 11 0 41 0 36 0 22 783A-4 21 3 0 42 0 34 0 24 76177SAMPLE POINT# SV DESMO VITDET SFUS FUS OTHERI TOTLIP TOTVIT TOTIN4 4A-1 20 5 0 15 0 61 0 25 754A-2 14 7 0 18 1 62 0 21 794A-3 12 2 0 11 0 74 0 14 854A-4 19 7 0 15 1 57 0 26 734A-5 12 3 0 13 0 73 0 14 864B-1 19 8 0 26 0 48 0 27 734B-2 23 11 0 21 0 44 1 34 654B-3 13 0 0 23 0 64 0 13 874B-5 13 5 0 23 1 58 0 18 824B-6 11 7 0 16 0 65 1 18 815 5A-1 16 3 0 71 1 11 0 19 825A-2 24 2 2 60 1 11 1 28 725A-4 22 6 0 35 1 35 1 28 725A-5 12 14 2 61 0 11 0 28 725A-6 25 3 1 47 1 23 1 29 715A-7 28 11 0 34 1 25 1 39 605A-8 26 4 1 45 1 23 0 31 695A-9 32 6 0 32 0 27 3 38 595A-10 27 4 0 28 4 36 2 31 675A-11 35 4 0 35 5 20 1 39 605A-12 20 5 0 10 0 63 2 25 735A-13 27 11 0 40 0 22 0 37 625A-14 16 4 0 39 0 42 0 19 815B-3 23 7 0 19 0 50 2 30 695B-5 22 6 0 24 0 48 0 28 725B-6 20 3 0 38 0 38 1 23 765B-7 27 4 0 52 4 13 0 30 705B-8 27 0 1 55 1 14 2 28 705B-9 25 7 0 22 0 43 2 32 655B-10 27 7 0 50 1 15 1 33 666 6-3 10 3 0 12 73 2 0 13 876-6 13 11 2 13 57 5 0 25 756-7 8 5 0 11 59 18 0 13 876-8 5 1 0 2 89 4 0 6 947 7A-2 12 4 0 17 0 66 0 16 847A-3 16 4 0 21 0 60 1 19 807A-5 26 6 0 22 0 45 0 32 68178Raw coal data - volume %POINT# SV DESMO VITDET SFUS FUS OTHERI TOTLIP MM TOTVIT TOTINlA-i 41 6 0 3 0 43 0 7 47 461A-2 26 12 0 21 0 39 0 2 38 601A-3 36 12 0 8 0 39 0 5 48 471A-4A 28 9 0 6 0 52 0 5 37 581A-5 22 19 0 20 0 38 0 1 41 581A-6 28 4 0 13 0 51 0 3 32 651A-7 52 1 0 37 1 2 0 7 52 401A-8 98 0 0 0 0 1 0 1 98 11A-8A 99 0 0 1 0 0 0 0 99 11A-9 53 10 0 5 0 24 0 8 63 29lA-li 56 9 0 2 0 27 0 7 64 291A-12 17 6 0 33 0 43 0 2 23 761A-13 28 2 0 13 0 58 0 0 29 701A-14 35 8 0 6 0 50 0 1 43 561A-15 23 6 0 10 0 59 0 1 29 701 A--i 6 23 3 0 11 0 60 0 1 26 7218-1 35 14 0 15 1 28 0 7 49 441B-2 28 16 0 12 0 42 0 2 44 541B-3 30 9 0 22 0 38 0 0 39 6018-5 28 10 0 19 0 38 0 5 38 581B-6 32 18 0 27 0 24 0 0 50 50lB-b 27 10 0 11 0 47 1 4 37 58iB-12 30 21 0 14 0 32 0 3 51 4618-13 22 4 0 19 0 54 0 2 25 731B-14 48 9 0 14 0 24 0 4 57 392A-2 20 6 0 15 5 41 1 12 26 612A-3 14 2 0 29 9 39 1 6 16 772A-4 22 5 0 22 0 43 3 5 26 6528-1 27 10 0 12 1 37 4 8 37 512B-3 20 6 0 11 5 49 1 8 26 642B-5 25 3 0 35 0 34 0 2 28 693B-1 16 9 0 22 2 50 0 1 26 733B-2 14 6 0 23 2 48 3 5 19 733B-3 22 7 0 39 0 30 1 1 29 693B-5 7 2 0 61 0 27 0 3 9 883B-6 26 8 0 31 0 33 1 1 34 653A-1 23 14 0 25 0 38 0 0 36 633A-2 23 2 0 29 0 45 0 0 25 753A-3 11 ii 0 41 0 36 0 0 22 783A-4 21 3 0 42 0 33 0 1 24 75179POINT# SV DESMO VITDET SFUS FUS OTHERI TOTLIP MM TOTVIT TOTIN4A-1 19 5 0 14 0 58 0 3 24 734A-2 13 7 0 17 0 60 0 2 20 784A-3 11 2 0 10 0 69 0 6 13 804A-4 18 7 0 14 1 53 0 8 24 684A-5 11 2 0 12 0 67 0 9 13 7848-1 16 7 0 23 0 42 0 11 24 6548-2 20 10 0 18 0 37 1 14 29 564B-5 12 5 0 21 1 55 0 5 17 784B-6 10 6 0 15 0 58 1 10 16 735A-1 15 3 0 70 0 10 0 0 18 815A-2 23 2 2 58 1 10 1 4 27 695A-4 21 6 0 35 1 34 1 2 27 705A-5 12 13 2 58 0 11 0 4 27 695A-6 25 3 1 46 1 23 1 2 28 695A-7 27 11 0 34 1 25 1 2 38 595A-8 25. 4 1 44 1 22 0 3 30 675A-9 31 6 0 32 0 27 3 1 38 595A-10 27 4 0 28 4 36 2 0 31 675A-11 34 4 0 35 5 20 1 1 38 595A-12 19 5 0 10 0 62 2 1 24 725A-13 26 11 0 40 0 21 0 1 37 625A-14 15 3 0 39 0 42 0 1 19 8058-3 23 7 0 19 0 49 2 0 30 685B-5 22 6 0 24 0 48 0 0 28 7258-6 20 3 0 38 0 38 1 0 23 765B-7 26 4 0 52 4 13 0 1 30 6958-8 27 0 1 54 1 13 2 3 28 6858-9 25 7 0 22 0 43 2 1 32 645B-10 26 7 0 49 1 15 1 1 33 656-3 9 2 0 11 66 2 0 9 12 796-6 12 10 2 12 53 5 0 7 23 696-7 7 5 0 10 54 17 0 7 12 816-8 5 1 0 1 84 4 0 5 6 907A-2 12 4 0 17 0 64 0 3 16 817A-3 14 3 0 19 0 54 0 9 17 747A-5 26 6 0 22 0 44 0 3 31 66SV - STRUCTURED VITRINITEDESMO - DESMOCOLLINITEVITDET - VITRODETRINITESFUS - SEMIFUSINITEOTHERI - OTHER INERTINITETOTLIP - TOTAL LIPTINITETOTIN - TOTAL INERTINITEMM - MINERAL MATTER180Permeability dataSAMPLE POINT# Ka KI Comments(md) (md)lA-i 0.186 0.0941A-2 0.0166 0.003941A-3 0.0183 0.004521A-4A 0.0383 0.01251 A-4B 0.0240 0.006561A-4C 0.0449 0.01561 A-4C2 0.0157 0.003691 A-5 0.450 0.2671A-6 0.256 0.1411A-7 0.261 0.1441A-8 2.63 1.981A-8A 2.84 2.161A-8c 8.88 7.271A-9 0.677 0.435lA-b 0.0567 0.0211iA-li 0.869 0.579iA-12 0.0292 0.0087lA-is 0.0434 0.01461A-14 0.159 0.07741A-15 0.0553 0.01991A-16 0.0488 0.0171lB-i 0.265 0.1461 8-2 0.0314 0.009581 B-3 0.036 0.01151B-4 0.0345 0.01101B-5 0.0324 0.010118-6 0.638 0.4081B-7 4.06 3.17lB-BA 2.16 1.60iB-8B 0.0664 0.0259iB-9 5.74 4.59lB-b 0.0688 0.0271lB-il 0.781 0.510IB-12 1.87 1.361B-13 0.0447 0.015218-14 0.491 0.30018-15 7.03 5.672 2A-1 9.45 7.74 NOT USED IN AVERAGE2A-1B 0.126 0.05892A-2 0.0915 0.03942A-3 0.0791 0.03252A-4 0.153 0.0752B-l 0.0779 0.03142B-2 0.183 0.09232B-3 0.684 0.44228-4 0.162 0.080228-5 0.662 0.421181SAMPLE POINT# Ka KI Comments(md) (md)3 3B-1 1.19 0.8163B-2 0.171 0.08253-B3 0.704 0.4483B-4 3.09 2.33 NEAR SAWCUT IRREGULARITY3B-5 0.873 0.5753B-6 2.12 1.573A-1 0.956 0.6453A-2 1.66 1.193A-3 0.460 0.2783A-4 0.241 0.1293A-5 4.16 3.24346 5.07 4.014 4A-1 0.0146 0.003344A-2 0.00403 0.0004924A-3 0.00672 0.001074A-4 0.0118 0.002434A-5 0.232 0.1244A-6 0.0149 0.003444B-1 0.0829 0.03454B-2 0.00112 6.93E-054B-3 0.0400 0.01284B-4 0.0284 0.008344B-5 0.014 0.003124B-6 0.0143 0.003225 5A-1 1.07 0.7375A-2 0.383 0.2555A-3 15.4 12.9 CLOSE TO FRACTURE, NOT USED5A-4 1.23 0.8545A-5 34.2 30.2 CLOSE TO FRACTURE, NOT USED546 0.537 0.334547 0.247 0.1345A-8 0.390 0.2305A-9 0.255 0.1405A-10 1.130 0.7805A-11 0.278 0.1545A-12 0.241 0.1305A-13 1.00 0.6815A-14 0.150 0.07275B-1 94.4 86.5 NO SEAL, NOT USED5B-2 43.6 38.9 NO SEAL, NOT USED5B-3 1.27 0.8875B-4 3.11 2.385B-5 2.19 1.625B-6 0.0788 0.03235B-7 0.161 0.07925B-8 0.335 0.1925B-9 0.416 0.2485B-10 0.203 0.105182SAMPLE POINT# Ka KI Comments(md) (md)6 6-2 7.86 6.38 CLOSE TO FRACTURE, NOT USED6-3 0.986 0.6726-4 5.62 4.50 CLOSE TO EDGE, NOT USED6-5 0.534 0.3326-6 0.371 0.2176-7 1.44 1.026-8 0.396 0.2357 7A-2 0.0485 0.01727A-3 0.0379 0.01247A-4 0.0398 0.01317A-5 0.0497 0.0177All surface observations made using 60X binocular microscope and 16X hand lens.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0052727/manifest

Comment

Related Items