CO2 storage and gas diffusivity properties of coals from Sydney Basin, Australia

International Journal of Coal Geology 70 (2007) 240 – 254 www.elsevier.com/locate/ijcoalgeo

CO2 storage and gas diffusivity properties of coals from Sydney Basin, Australia A. Saghafi a,⁎, M. Faiz b , D. Roberts a a

CSIRO Energy Technology, P.O. Box 330, Newcastle, NSW 2300, Australia b CSIRO Petroleum, P.O. Box 136, North Ryde, NSW 1670, Australia Received 6 May 2005; accepted 16 March 2006 Available online 7 July 2006

Abstract Measurements of CO2 adsorption and diffusion properties of coals are reported for various coalfields within Sydney Basin, New South Wales (NSW), Australia. Adsorption measurements were undertaken using a gravimetric method. Measurements carried out on 27 coals show that Sydney Basin coals at CO2 sub-critical conditions, namely gas pressures below 6 MPa and temperatures below 39 °C, can adsorb a maximum volume (Langmuir volume) of 40 to 80 m3 of CO2 per tonne of coal on a dry ash free basis (daf). The coals used in this study are of sub-bituminous to bituminous rank, ranging from 0.66 to 1.45% mean maximum vitrinite reflectance, and are from depths ranging from about 27 m to 723 m. The highest adsorption capacity applies to the highest rank coal, which is also the deepest coal. The standard deviation between Langmuir modeled and measured values is less than 1.5 m3/t, corresponding to a relative error of less than 2.7% for all except one coal. Based on adsorption isotherms, the CO2 storage capacity for in-situ seam pressure conditions range from about 6 to 51 m3/t. CO2 diffusion properties of 15 of these coals, determined using a newly developed system capable of accurately measuring diffusivity of gases in solid coal indicate that CO2 diffusivity (diffusion coefficient) in the Sydney Basin coals varies from 1.2 × 10− 6 to 10.2 × 10− 6 cm2/s. The diffusivity does not show any discernable trend with the variation in depth and rank. Porosity measured by a mercury injection method varies from 4 to 10% and decreases with increase in coal depth and rank. For some of the coal samples adsorption measurements for pure CH4, CO2 and N2 indicate that the Sydney Basin coals can store twice as much CO2 as CH4 and six times more CO2 than N2 (volume basis). Also, measurement of diffusivity in solid coal samples shows that CO2 diffuses twice as quickly as CH4. The data obtained from this study and the estimated coal resources in the state of New South Wales, allow CO2 sequestration potentials to be calculated. © 2006 Elsevier B.V. All rights reserved. Keywords: Coal; Carbon dioxide; Methane; Diffusivity; Gas adsorption; Langmuir; Sequestration

1. Introduction About 300 Mt of coal is produced in Australia annually with almost half of the production from Sydney Basin coalfields. The Sydney Basin, which is located in New ⁎ Corresponding author. Tel.: +61 2 9490 8670; fax: +61 2 9490 8921. E-mail address: [email protected] (A. Saghafi). 0166-5162/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.coal.2006.03.006

South Wales (NSW), contains four major coalfields: Hunter, Newcastle, Western and Southern Coalfields. All coals are Permian and their rank is generally medium to high volatile bituminous, except for the Southern Coalfield coals which are generally low to high volatile bituminous. Coal in the Southern Coalfields of Sydney Basin is >300 m deep and is mined underground whereas most coal in northern coalfields is extracted from open-cut mines. The open-cut mining produces half of the total coal production of NSW.

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

The current study was undertaken on coals obtained from underground faces, in-seam boreholes and surface exploration boreholes in the Sydney Basin and involves investigations on factors affecting CO2 storage and injectability in coal, in particular the adsorption and diffusivity properties of coal. CO2 adsorption and diffusivity properties were determined for coals ranging in depth from about 27 to 723 m. The results from the deep coal seams are applicable to CO2 sequestration projects, whereas the data from the shallow seams are useful in assessing the extent of fugitive CO2 emissions into the atmosphere from surface coal mining. The maximum quantity of gas that can be stored in a given coal is mainly a function of its adsorption capacity, though at high pressures significant amounts of gas can also be stored in the pore system if the pores are not saturated with water. At a given set of P–T conditions, a coal can adsorb higher amounts of CO2 than CH4, depending on coal rank; high to medium volatile bituminous rank coals store more gas than sub-bituminous coal and anthracite. The desorption rates of CO2 and CH4 for coal are also considerably different (Williams et al., 1995). During the early stages of coalification, coal generates large amounts of carbon dioxide, but most of this is lost due to dissolution in mobile water and migration out of the coal seam. The origin of CO2 in some coalfields of the world (such as Australia, Poland and France) has been subject to comprehensive investigations by researchers worldwide, initially because of the safety hazard of this gas in underground mining operations, as evidenced by the tens of thousands of gas outbursts over the last century (Lama and Saghafi, 2002). Clayton (1998) reviewed the geochemistry of seam gas and based on the work of other researchers (Kotarba and Rice, 1995a,b; James and Burns, 1984; Smith et al., 1985; Whiticar et al., 1986; Smith and Pallasser, 1996) listed four sources for CO2 gas in coal seams: a) decarboxylation reactions of kerogen and soluble organic matter during burial heating of the coal, b) mineral reactions such as thermal decomposition or dissolution of carbonates or other metamorphic reactions, c) bacterial oxidation of organic matter and d) magmatic intrusion. One particularity of Australian coalfields is the occurrence of large amounts of CO2 in many coal seams; in some cases close to 100% of the seam gas. It is believed that CO2 in Sydney Basin coals is mainly derived from magmatic sources as the high CO2 content areas generally have isotopic compositions of δ13C of about −7‰ (Smith and Gould, 1980; Smith et al., 1985). Also high concentrations of CO2 are localised laterally along some major faults. Mining experience in Australia shows that CO2 content can vary significantly within short distances in the same seam and within the same coal mine (Lama and

241

Saghafi, 2002). Differences in the behaviour of CH4 and CO2 have been noted for more than a century from coal mining operations worldwide. It has been found that gas outbursts can occur at lower gas contents for CO2 than for CH4, and the presence of high CO2 content in coal seams has been the cause of numerous gas outbursts during underground coal mining. During the last 50 years, many outbursts occurred in Australian mines. In some instances where the dominant gas is CO2, outbursts happened more frequently. For instance, at Tahmoor, Metropolitan and Westcliff collieries in Illawarra coalfield in the southern part of the Sydney Basin, gas outbursts were caused mainly by CO2. However there were also some occurrences of CH4 outbursts in Australian coals. In some instances, the large number of CH4 outbursts has led to complete closure of the mine, for example the Leichhardt colliery, Bowen Basin, Queensland (Moore and Hanes, 1980). Due to the common occurrence of CO2 in Australian coal seams and its implications for coal mining, the mechanism of CO2 storage and flow in coal has been investigated during the last two decades (e.g. Lama and Bodzinoy, 1996; Saghafi and Williams, 1998). However most of these studies have concentrated on the behaviour of methane in coal, particularly for mine safety and coal bed methane (CBM) production. In the last decade, the impetus for enhanced coalbed methane recovery (ECBM) and increasing interest in the CO2 sequestration to abate greenhouse gas emissions has led to further research into physical chemistry of the interaction of gas with coal. The capacity of coal to store CO2 and its ability to allow movement of gas through its pore and fissure systems are two major factors that influence the release of gas during coal mining and storage of gas in coal seams (CO2 sequestration). Storage and migration properties for coal are represented by adsorption and diffusivity measurements and these can be considered as primary coal reservoir properties that are required for any meaningful evaluation and predictive modeling of coal gas reservoirs. They depend largely on coal type, rank, pore structure and any pore filling, present as a consequence of paleo-fluid flow and mineralisation. External factors such as mining methods and stress regimes, as well as the flow of meteoric water, also affect these properties, but are of secondary importance. The aim of the present investigation is to characterise Sydney Basin coals according to these two parameters, particularly for CO2. The measurement methodology and mathematical expressions were developed and applied to 27 coal samples. After a brief discussion on the nature of gas in coal, this paper presents the measuring systems, methodology and implications of the measurements.

242

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

2. The nature of gas in coal Gas contained in deep, high rank coals is generally of thermogenic origin, i.e. generated during the process of organic metamorphism. For shallow coals, gas is mainly a byproduct of anaerobic biogenic activity. Worldwide, seam gas mainly consists of methane with lesser amount of carbon dioxide. Nitrogen is present in some areas, but only in small quantities (< 10% of the total gas volume). Other higher hydrocarbons (C2+) may also be present in some deeper (> 600 m) Australian coals, but they generally do not exceed 12% of the total gas (Faiz et al., 1999). 2.1. Gas storage in coal Coal contains a mixture of micro-, meso- and macropores. In this paper reference to micro-pores implies pore diameters of less than 2 nm (nanometer) and reference to macro-pores implies pores of diameters larger than 50 nm. The intermediate pore sizes are referred to as meso-pores (van Krevelen, 1993). Although coal is a reservoir rock for gas, it differs markedly from conventional petroleum reservoirs in that the volume of gas, which it can store, exceeds its open pore volume by an order of magnitude. The volume of pores in coal is small (less than 10% volume for coals in this study) and the majority of gas in coal consists of adsorbed gas which covers the surfaces of micro-pores. Coal micro-pore surface area can reach several hundreds of square meters per gram of solid, making large areas available for gas adsorption. This has been well established since the early work on coal pore structure; for instance Griffith and Hirst (1944) measured pore surface area using the heat of wetting of coal in methanol. This study showed that the surface area of coal is in the order of 20 to 200 m2/g. The surface area measured depends on the type of sorbate used; for example CO2 gas yields higher surface areas than N2 as it has more accessibility to coal micro-pores. Levine (1993) has quoted Thomas and Damberger (1976), who measured surface areas of coals from Illinois Basin and found that CO2 surface area for these coals decreases from over 250 m2/g for high volatile C bituminous to less than 50 m2/g for high volatile A bituminous coal. However, Reucroft and Patel (1986) have reported an overestimation of 10–20% in surface area by using CO2 adsorption techniques. They attribute this error to excessive swelling of the coal when exposed to CO2. The amount of adsorbed phase in coal depends not only on available surface area, but also on the equilibrium state of surface attraction and repulsion forces (van der Waals forces). Equilibrium is reached when the total gas– solid, surface potential energy is minimised. Adsorption

of gas onto coal is a long-range, weak interaction and the phenomenon is therefore a physisorption or physical adsorption. In the process of physisorption, molecules of gas lose kinetic energy and adhere to the coal surface. The amount of energy released is an indication of the strength of adsorption; generally it is less than half of the enthalpy for condensation of methane (Yee and Seidle, 1993) and, also, for the condensation of carbon dioxide (Stevenson et al., 1991). The magnitude of energy release in physisorption is in the order of 20 kJ mol− 1 (Zwietering and van Krevelen, 1954), whereas in chemisorption gas molecules form a covalent chemical bond with the solid surface molecules, and the energy release is much higher, typically in the order of 200 kJ mol− 1 (Atkins, 1978). As mentioned above, the surface areas of pore systems in coal greatly exceed the volume of pores, therefore, most gas stored is in an adsorbed phase. This is true particularly at low pressures where nearly all gas in coal is in an adsorbed phase. The potential volume of gas that can be adsorbed onto a coal can be estimated from knowledge of its internal surface area. For example, according to the current study, the volume of CO2 that can be adsorbed onto a coal of 200 m2/g pore surface, is more than 60 m3/t assuming that the CO2 molecular kinematic diameter is 3.3 Å (Mahajan, 1991) and adsorption occurs in a single layer. 2.1.1. Mathematical expression of gas storage in coal Adsorbed volume of gas in coal depends on the energy of the free gas, and therefore on temperature and pressure. At a given temperature and pressure, and after sufficient time, the adsorbed phase and free phase are in kinetic equilibrium, i.e. the rate of adsorption and desorption from coal pore surfaces is equal. One way of expressing the adsorbed volume kinetic equilibrium is by using a Langmuir type equation of the form, ca ¼

VL p p þ PL

ð1Þ

where ca is adsorbed gas content (gas volume per unit mass of coal), p is gas pressure, and VL and PL are experimental coefficients. The coefficient VL represents the maximum gas storage capacity of the coal and is termed the ‘Langmuir volume parameter’. The coefficient PL is the ‘Langmuir pressure parameter’ and represents the gas pressure at which coal adsorbs a volume of gas equal to half of its maximum capacity, i.e. c =VL / 2 when p =PL. At low pressures the Langmuir equation reduces to a linear equation of pressure (Henry's Law) with a linearity coefficient of VL /PL. Thus low PL values indicate a steeper or more convex adsorption curve. This means that coal at low

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

PL values can reach the full storage capacity at lower gas pressures compared to coal with high PL values. As the Langmuir equation is based on a mono-layer adsorption mechanism (Type I isotherm) it is applicable for low pressures (<6 MPa for CO2) and for gases where the molecule sizes and coal pore diameter are of the same order of magnitude. Currently in Australia, the Langmuir equation is the most commonly employed mathematical expression used in industry to describe adsorption of gas onto coal (Crosdale et al., 2005). 2.1.2. Gas content of coal In-situ coal contains gas both on micro-pore surfaces, in an adsorbed phase, and as a free phase compressed within macro-pores. Adsorbed gas content of coal can be directly measured from drill cores using standard techniques (e.g. Williams et al., 1992). Free gas content can be expressed as, cf ¼

ep qPa

ð2Þ

where cf is the free gas content of coal expressed in terms of volume of gas per unit mass of coal; p, ε and ρ are gas pressure, porosity and density of the coal, respectively and Pa is atmospheric pressure. As mentioned above, the adsorbed phase volume (ca) is always higher than the free gas volume (cf) for typical coal seams. However, at very high pressures, i.e., deep coal seams, or low rank coals, free gas mass may become comparable with the adsorbed phase mass if porosity is not reduced significantly. Theoretically, by comparing Eqs. (1) and (2), it can be deduced that at pressures equal or greater than pressure Pc, the density of free gas would be so high that its mass becomes comparable to the adsorbed phase mass: p z Pc ¼ qPa

VL −PL Z cf z ca e

ð3Þ

As the above equation suggests, for high values of PL and coal porosity, the free gas content of coal can be equal to or higher than the adsorbed gas content. 2.2. Gas flow in coal Gas flow in coal pores occurs as a result of both pressure and concentration gradients. Coal has a heterogeneous pore structure and is a dual porosity medium; it can be assumed that in micro-pores gas is displaced due to diffusive flow, whereas in macro-pores and fractures viscous flow is dominant (Saghafi et al., 1987). The two

243

regimes of gas flow can be combined and presented by a single equation expressing conservation of mass: Y Y Ac ¼ − j : wv þs At

ð4Þ

In this equation, c represents gas content and t repreY sents time; (j: ) is the divergence operator, which is applied to the gas flux to obtain the divergence of viscous gas flux ψv (the first term on the right-hand side of the equation). The second term on the right-hand side, s, is the source (or sink) term which corresponds to the rate of desorption (source) or adsorption (sink). Diffusive flux is embedded in this term as the rate of desorption/adsorption would depend on the diffusivity of coal. The source/sink term in this equation would be the rate of adsorption/ desorption of gas into the fracture system. The diffusion of gas from micro-pores into macro-pores is a combination of various diffusion transport regimes. Migration from pores and microfractures into macro-pores and fractures may be expressed by a generalized Fick's equation of the form, Y

Y

wd ¼ −D j c

ð5Þ

where ψd is the diffusion flux, D is the diffusion coefY ficient, j is the gradient operator, c is the gas concentration (content). Diffusion coefficient is expressed in terms of cm2 s− 1. Note that the diffusive flux ψd is caused by gas concentration gradient whereas the viscous flux ψv in Eq. (4) is a result of gas pressure gradient.

3. Measurement methods 3.1. Adsorption isotherm The technique developed to measure adsorption isotherms is based on a gravimetric method (Saghafi, 2003). This method involves direct measurement of the increase in weight of coal as it is saturated with gas at increasing gas pressures. The equation of state is derived from the density of the free gas phase and is measured simultaneously in a reference empty container held at the same pressure and temperature as the coal container. Approximately 300 g of coal is used for the experiments and particle size ranges from 90 to 150 μm. A schematic of the gravimetric system developed and used for this study is shown in Fig. 1. 3.1.1. Determination of volume of gas adsorbed In the gravimetric method, the increase in the mass of an empty reference canister and a canister containing

244

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

coal is measured while both canisters are connected to a supply gas at the same pressure. The empty canister is used to directly determine the density of gas at measurement pressure and temperature (equation of state for gas phase) as follows, dg ¼

wref vref

ð6Þ

where dg is the gas phase density, wref is the mass of gas compressed in the reference canister and vref is the volume of the reference canister. The ‘void uncorrected’ isotherm (‘Excess’ or ‘Gipps isotherm’) is calculated by using the following expression, wG ¼ w−Vvoid dg

ð7Þ

where w is the mass of gas in the coal canister, Vvoid is the uncorrected volume of void, which is the volume of the empty canister minus the volume of coal. In order to obtain the true mass of adsorbed gas, the void volume in the coal canister must be corrected; the correction is for the volume of adsorbed phase, which occupies part of the initial void. The net volume of the void is, ðVvoid Þt ¼ Vvoid −

wads dads

ð8Þ

where dads is the density of the adsorbed phase and wads is the mass of adsorbed phase. The adsorbed phase mass

is, wads ¼ w−ðVvoid Þt dg

ð9Þ

If the corrected void volume from Eq. (8) is used in (9) the adsorbed phase mass would be, wads ¼

wG d

1− dadsg

ð10Þ

The adsorption isotherm is expressed in terms of volume, as a gas phase, at standard conditions. Hence the adsorbed volume is, vads ¼

wads dgst

ð11Þ

where vads is the volume of gas adsorbed, normalized to standard conditions and dgst is the gas density at standard conditions. Note that the Australian standard conditions are gas pressure of 1 atm (101.325 kPa) and a temperature of 20 °C. 3.1.2. Adsorption isotherm equation For sub-critical gas conditions adsorptive behaviour is assumed to follow a mono-layer type I mechanism and is modeled by a two parameter, hyperbolic, Langmuir type equation as described above (Eq. (1)). The Langmuir isotherm can be written both in terms of gauge or absolute pressures (Fig. 2); the gauge pressure is the absolute

Fig. 1. Schematic diagram of gravimetric adsorption measurement apparatus; wc and wr are the mass of gas in coal and reference cells respectively, p is the gas pressure.

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

245

Fig. 2. Calculated total CO2 gas content (Qt) and CO2 desorbable gas content (Qd) for coal3; based on adsorption isotherm.

pressure less than the atmospheric pressure. When the equation is written in terms of gauge pressure, the value of ca in Eq. (1) represents desorbable gas content, Qd. Equally, when the equation is written in terms of absolute pressure, the value of ca represents total gas content, Qt. The relationship between the Langmuir absolute parameters (VLa, PLa) and gauge parameters (VLg, PLg) can be obtained by moving the origin from point (p =Pa, ca =C0) to point (0,0). The origin then corresponds to the zero absolute pressure in the (ca, p) space. It yields, VLa ¼

VLg ; PLa ¼ PLg −Pa 1− PPLga

ð12Þ

where Pa is the atmospheric pressure and C0, is the gas adsorbed at atmospheric pressure. 3.2. Measurement method for gas diffusivity Gas diffusivity of coal indicates the ease with which gas migrates from micro-pores into macro-pores and fractures.

Diffusive flux is proportional to the gas concentration gradient, with the coefficient of proportionality termed the diffusivity and, it is a fundamental property of a coal-gas system. Diffusive flow can be a combination of various flow mechanisms depending on coal pore structure and gas type; molecular, Knudsen and surface diffusion mechanisms may all contribute depending on the pore structure of coal and type of gas. For strongly adsorptive gases such as CO2, surface diffusion may be the dominant regime. However, for the purpose of this study we assume that all mechanisms of diffusion can be combined and be represented by the equation described in Eq. (5). In a one dimensional space it is written as Fick's equation: wd ¼ −De

Ac Ax

ð13Þ

where De is the effective diffusion coefficient combining various diffusive flow mechanisms, c is the gas concentration, x is the space dimension and ψd is the gas diffusive flux. Diffusion coefficient has dimensions of length squared divided by time (cm2 s− 1).

Fig. 3. Schematic diagram of measurement apparatus for gas diffusivity in coal, used in the present study; gas flows from chamber 1 to chamber 2 through the coal disk.

246

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

The technique commonly used for measurement of diffusivity is to saturate powdered coal with gas and then release the pressure and measure the desorption rate. A ‘single porosity’ model is usually used to describe the gas diffusion out of coal particles and determine the diffusion coefficient (Nandi and Walker, 1970). This ‘indirect method’ would yield diffusivity which depends on assumptions made concerning the average particle size and, therefore, cannot represent true diffusion in coal matrices. More recently Harpalani and Ouyang (1999) introduced a new method, whereby, solid coal samples instead of powdered coal were saturated with methane and then the coal was allowed to release its gas by injecting helium into the coal container. This method also requires numerous assumptions on the transient values of methane concentration in solid coal. It also requires significant lengths of time for coal to be saturated with methane. A new technique has been developed at CSIRO laboratories to measure the gas flux through specially prepared solid coal discs taken from core samples (Saghafi, 1996; Saghafi, 2003). A schematic diagram of the apparatus used to measure diffusivity is presented in Fig. 3. The system consists of a coal disk holder and two gas chambers at each side of the disk holder. The solid coal disk is 6 to 12 mm in diameter and 2–6 mm in thickness. Initially each chamber is filled with a different type of gas (e.g. CO2 and N2). Gas 1 at concentration C1,1 in chamber 1 would diffuse through the coal disk into chamber 2 where its concentration is lower (C2,1). Gas 2 would diffuse in the opposite direction from a concentration C2,2 in chamber 2 to a lower concentration of C1,2 in chamber 1. The evolution of each gas concentration in each of the two chambers is measured over a period of up to one week. Based on Fick's equation (Eq. (13)) and the law of conservation of mass, the evolution of concentration is theoretically determined. The diffusion coefficient is then determined by matching the measured flux curve against the theoretical curve and carrying out an optimization calculation for best value of De. 3.3. Coal porosity In the present study a mercury intrusion technique was used to measure porosity. The porosity measured includes pores with diameters larger than about 4 nm (i.e. mainly meso- and macro-pores). The method involves crushing the coal to 0.5 to 1.0 mm diameter particles and then forcing mercury to intrude the pores at pressures up to 420 MPa. By comparing this uptake to the uptake at low pressure (200 kPa) which defines the gross volume of the sample, the total pore volume can be calculated. This methodology, however, will not measure porosity of the cleat system, due to very small particles.

4. Results and discussion 4.1. Coal sampling and measuring conditions In the present study gas adsorption isotherms for 27 samples from 17 Permian coal seams from the Sydney Basin were measured. These samples were obtained from depths between 27 m and 723 m. Tables 1 and 2 summarise sample details, including results of proximate analyses and coal seam depths. Fig. 4 shows the relationship of volatile matter of the coals with depth. The coals are mostly of high to medium volatile bituminous rank with mean maximum vitrinite reflectance in oil (Ro max) ranging from 0.66% to 1.45%. Two of the coal samples have been affected by contact metamorphism and yield Ro max values over 11%. The inherent moisture content of the coals varies from 0.4 to 7.9%, whereas the ash yield for all samples, except one, varies from 5 to 34%. One high rank sample has an ash yield of 57%. The daf (dry ash free) volatile matter content ranges from 21 to 43%, except for the two contact metamorphosed coals for which it is about 15%. 4.2. Adsorption isotherm measurements All samples were measured at sub-critical CO2 pressure conditions (<6 MPa). In this study, the most commonly used measurement temperature for isotherms is 27 °C, which corresponds to the average in-situ temperature of the coals. As indicated in Table 1, four measurements were made at temperatures ranging from 30 to 39 °C for samples from deep seams. The adsorption isotherms were measured on coal ‘as received’ and the inherent moisture content was measured on some coals before and after measurement to investigate the change in moisture level during adsorption measurement. The change was, for all coals, below 0.1% absolute. After the completion of each measurement the adsorbed volumes were normalised to the Australian standard P–T conditions (i.e. 20 °C and 101.325 kPa). The measurements of CO2 adsorption for the coals in terms of Langmuir parameters as described by Eq. (1) are given in Table 3. Langmuir parameters are expressed both in terms of gas gauge and absolute pressure as described in Section 3.1.2. The most adsorptive coal analysed (coal13) is from the deepest seam (723 m) which has the highest Ro max value (1.45%). The least adsorptive coal (coal14) is from 362 m which is not the shallowest nor the lowest rank coal. The vitrinite reflectance value of this coal was not measured, but would be ∼1% based on measurements on other coals from the same seam in neighboring locations. In Fig. 5, the adsorption isotherms of the most adsorptive and least

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

247

Table 1 Details of Sydney Basin coal samples studied for adsorption properties Coal samples

Coal proximate and density analysis a

Sample code Seam code Location code Seam depth Vitrinite reflectance Moisture Volatile matter Fixed carbon Ash yield Coal density (m) (%) (%) (%) (%) (%) (g/cm3) Coal1b Coal2b Coal3 Coal4 Coal5 Coal6 Coal7 Coal8 Coal9 Coal10 Coal11 Coal12 Coal13 Coal14 Coal15 Coal16 Coal17 Coal18 Coal19 Coal20 Coal21 Coal22 Coal23 Coal24 Coal25 Coal26 Coal27

Seam1 Seam1 Seam1 Seam5 Seam6 Seam2 Seam2 Seam7 Seam8 Seam15 Seam14 Seam1 Seam3 Seam1 Seam1 Seam1 Seam4 Seam9 Seam10 Seam11 Seam12 Seam13 Seam16 Seam14 Seam1 Seam1 Seam17

Loc1 Loc1 Loc1 Loc3 Loc3 Loc4 Loc4 Loc5 Loc5 Loc5 Loc5 Loc6 Loc6 Loc7 Loc8 Loc8 Loc2 Loc9 Loc9 Loc9 Loc10 Loc10 Loc11 Loc11 Loc1 Loc12 Loc13

264.0 264.0 430.0 250.0 207.0 365.6 332.0 293.9 348.9 377.9 410.9 609.0 722.5 362.0 434.0 450.0 93.5 57.3 60.5 75.5 431.0 557.2 26.8 37.5 450.0 485.0 41.0

11.64 11.21 1.39

4.6 4.0 0.4 4.7 7.9 0.6 0.6 5.1 3.9 3.4 2.9 0.7 0.8 1.5 1.2 1.2 2.2 3.0 2.9 2.8 1.6 1.4 4.8 2.2 1.0 0.9 3.2

0.73 0.68 0.68 0.72 1.23 1.45 1.04 0.99 0.76 0.81 0.83 0.75 0.91 0.88 0.66 0.67 1.30 0.81

10.4 11.9 21.0 28.4 27.1 24.2 35.1 33.9 32.1 30.2 31.8 25.9 22.1 28.9 33.0 25.7 38.2 26.0 30.0 32.6 27.6 26.8 34.2 30.4 18.8 12.4 24.2

60.0 62.7 66.4 57.3 50.4 66.9 52.7 53.7 59.0 46.6 53.8 65.3 62.3 64.3 59.2 64.0 51.4 36.8 55.8 53.6 62.0 44.6 54.9 53.5 71.4 29.7 51.0

25.0 21.4 12.2 9.6 14.7 8.3 11.6 7.3 5.0 19.8 11.8 8.2 14.8 5.3 6.6 9.1 8.2 34.2 11.3 11.0 8.4 27.2 6.1 13.9 8.8 57.0 21.6

1.89 1.77 1.49 1.37 1.38 1.35 1.38 1.36 1.33 1.45 1.38 1.38 1.45 1.34 1.38 1.44 1.38 1.55 1.35 1.32 1.43 1.56 1.36 1.43 1.48 1.87 1.55

a

Mean maximum reflectance in oil (Ro max %), bcontact metamorphosed coal.

adsorptive coal are shown which indicate a twofold adsorption capacity variation across the Sydney Basin studied.

4.2.1. Suitability of Langmuir model to represent CO2 isotherms The Langmuir isotherm is the most commonly used expression for adsorption of methane and carbon dioxide

Table 2 Coal seams from which samples were obtained Code

Coal seam

Seam1 Seam2 Seam3 Seam4 Seam5 Seam6 Seam7 Seam8 Seam9 Seam10 Seam11 Seam12 Seam13 Seam14 Seam15 Seam16 Seam17

Bulli Wongawilli Tongarra Vaux Vynn Kayuga Woodlands Hill Arrowfield Warkworth Mt Arthur Piercefield Vales Point Borehole Glen Munro Bowfield Blakefield Bayswater

Fig. 4. Volatile matter according to depth for coals from Sydney Basin; coal samples obtained from 23 to 723 m depth.

248

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

Table 3 Measured CO2 adsorption isotherm parameters for coals from Sydney Basin Coal samples and isotherm temperature

Desorbable isotherm

Sample

Temperature (°C)

Langmuir volume, VL (m3/t)

Coal1a Coal2a Coal3 Coal4 Coal5 Coal6 Coal7 Coal8 Coal9 Coal10 Coal11 Coal12 Coal13 Coal14 Coal15 Coal16 Coal17 Coal18 Coal19 Coal20 Coal21 Coal22 Coal23 Coal24 Coal25 Coal26 Coal27

27 27 27 27 27 27 27 27 27 27 27 33 39 27 27 27 27 27 27 27 30 35 27 27 27 27 27

45.9 52.7 44.2 36.5 37.1 38.1 37.2 43.4 47.1 39.0 41.4 48.3 57.2 35.4 36.5 41.3 43.9 30.2 50.1 47.0 62.2 47.8 43.6 46.8 40.5 29.2 40.08

Error using Langmuir equation

Absolute isotherm

Langmuir pressure, PL (kPa)

Standard deviation (m3/t)

Relative error (%)

Langmuir volume, VL (m3/t)

Langmuir pressure, PL (kPa)

4570 3473 1790 1877 2161 1179 1176 2297 2429 2220 2311 1776 1584 1422 2119 2096 2334 1415 2308 2265 2781 3257 2512 2824 2216 2768 2135

0.39 1.08 1.64 0.40 0.56 0.87 0.84 0.54 0.70 0.45 0.52 0.88 0.99 0.64 0.27 0.23 0.76 0.29 0.91 0.83 1.50 0.80 0.88 0.78 1.07 0.25 0.28

0.85 2.04 3.72 1.11 1.51 2.27 2.25 1.26 1.48 1.16 1.26 1.83 1.74 1.82 0.75 0.57 1.74 0.96 1.82 1.76 2.41 1.67 2.02 1.67 2.64 0.86 0.70

46.9 54.2 46.9 38.6 38.9 41.7 40.7 45.4 49.1 40.8 43.3 51.2 61.1 38.1 38.3 43.4 45.9 32.6 52.4 49.2 64.5 49.3 45.4 48.6 42.5 30.3 42.1

4469 3372 1689 1775 2060 1078 1075 2196 2328 2119 2210 1675 1483 1321 2018 1995 2233 1314 2207 2164 2680 3156 2411 2723 2115 2667 2034

a

Contact metamorphosed coal.

into coal but some researchers have suggested introducing other terms into the equation for carbon dioxide adsorption (e.g. Ozdemir et al., 2003). Although the introduction of new terms would reduce the errors of curve fitting, this would also introduce more complexity to the equation and could lose its acceptance in industry. One of the aims of the present study is to assess the suitability of the Langmuir equation to represent CO2 storage behaviour of coal. This assessment is based on relative error, defined as the standard deviation of measured and fitted values normalized to the Langmuir volume: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u n uX u ðci −cL Þ2 u t 1 r¼ ð14Þ nVL2 where r is the relative error, ci is the volume of gas adsorbed at measurement point i, cL is the volume predicted by the Langmuir equation, VL is the Langmuir volume parameter and n is the total number of measurement points

for each isotherm. In Fig. 6, the measured values and the Langmuir predicted values for all isotherms are illustrated. The results show that the points of the curve all lie very close to the bisector of the coordinate axes indicating only small errors in using a Langmuir type equation to describe the CO2 isotherm. In Fig. 7, the relative errors of deviation of Langmuir-predicted values are given, indicating errors of below 2.7% in all except one case. These levels of error are acceptable as the errors in measurements can be higher than the errors in using Langmuir-predicted values. For example, Australian Standards (1999) guide to determination of gas content recommends that gas content errors should be less than 10%. 4.2.2. CO2 adsorption properties of Sydney Basin coals Langmuir volumes for coals represent maximum sorption capacity and for the coals studied they range between 40 and 80 m3/t (daf). Fig. 8 indicates that the CO2 adsorption capacity of the coals overall decreases with volatile matter (VM; dry ash-free basis) though this correlation is weak (r2 = 0.13).

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

249

Fig. 7. Relative error (r) in using a Langmuir type equation for determining CO2 adsorption for all coals studied.

Fig. 5. CO2 adsorption isotherms for the highest and lowest adsorptive coals from Sydney Basin; the data are corrected for ash yield and moisture; all coals except one have relative errors below 2.7%.

The parameter PL (Langmuir pressure parameter), determined from the adsorption isotherm, gives an indication of the rate of gas adsorption onto coal relative to pressure. A low PL indicates that a coal would reach its maximum potential capacity at low gas pressures. On the other hand, for gas desorption the pressure needs to be significantly reduced for gas to start desorbing. In Fig. 9 the parameter PL is plotted against volatile matter content. The pressure parameters vary between 1000 to 3000 kPa except for the contact metamorphosed coals

Fig. 8. CO2 Langmuir volume VL for the Sydney Basin coals in relation to volatile matter content (VM).

where PL reaches values up to 4500 kPa. No correlation with volatile matter content is evident from the PL data. The adsorption capacity in terms of coal rank (Ro max) for some of the studied coals is presented in Fig. 10

Fig. 6. Measured adsorbed volume versus Langmuir estimated volume for CO2 adsorption for coals studied.

Fig. 9. Langmuir pressure (PL) in relation to volatile matter content.

250

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

muir volume, and therefore in adsorption capacity, with rank.

Fig. 10. CO2 adsorption capacity of the Sydney Basin coals according to the rank, northern and southern Sydney Basin coals both show increases in adsorption capacity with increases in rank; Langmuir volume is corrected for ash yield and moisture content.

(vitrinite reflectance data on all samples are not available). Two groups of data can be distinguished, the lower rank coals are from northern part of Sydney Basin and the higher rank coals belong to southern part of this basin. Within each group an overall increase may occur in Lang-

Fig. 11. CO2 storage capacity of the Sydney Basin coals at in-situ pressure; C is the adsorption capacity at coal depth h and C0, h0 and α are curve fitting parameters.

4.2.3. Potential for CO2 sequestration in Sydney Basin In the State of New South Wales (NSW) about 95% of electricity is generated from coal-fired power plants. Options for mitigation of greenhouse gas emissions from power plants are limited and sequestration of CO2 into coal seams is an option regarded with growing interest from local government and power producers. Coal resources in NSW are abundant and are located close to power plants. Estimation by Geoscience Australia, quoted by Australian Black Coal Statistics (2004), indicates a total of demonstrated (34 Bt) and inferred (57 Bt; for deep coals) resources of 91.0 billion tonnes of coal in NSW. For evaluation of CO2 sequestration potential of coal seams of the Sydney Basin, the storage capacities can be plotted according to coal depth (Fig. 11). The data show that the adsorption capacity varies from 6 to 51 m3/t for seam depths of between 27 and 723 m. The total volume of stored gas may be larger if the volume of free, compressed gas is also considered. Table 4 CO2 diffusivity (D), coal porosity, Langmuir volume (VL) and volatile matter content (VM): results for samples from Sydney Basin Sample code

Depth (m)

Porosity (%)

Coal1 Coal2 Coal3 Coal4 Coal5 Coal6 Coal7 Coal8 Coal9 Coal10 Coal11 Coal12 Coal13 Coal14 Coal15 Coal16 Coal17 Coal18 Coal19 Coal20 Coal21 Coal22 Coal23 Coal24 Coal25 Coal26 Coal27

264.0 264.0 430.0 250.0 207.0 365.6 332.0 293.9 348.9 377.9 410.9 609.0 722.5 362.0 434.0 450.0 93.5 57.3 60.5 75.5 431.0 557.2 26.8 37.5 450.0 485.0 41.0

9.86 9.88 4.23

D × 106 (cm2/s) 7.73 5.83 7.83 1.85

5.30 4.80 5.26 5.15 5.07 8.65 6.62 6.71

8.13 7.41 2.25 1.18 1.48 2.34

4.95 7.22 3.95 3.12 6.10

8.03 7.70 9.06 10.20 8.76

VL, daf (m3/t)

VM, daf (%)

66.65 72.71 53.64 45.02 50.26 45.75 46.36 51.77 53.89 53.16 50.77 56.23 72.34 40.90 41.53 48.33 51.17 51.87 61.04 57.11 71.70 69.11 50.98 57.91 47.10 71.90 55.95

14.77 15.95 24.03 33.19 34.95 26.56 39.98 38.70 35.24 39.32 37.28 28.43 26.18 31.00 35.79 28.65 42.63 41.40 34.97 37.82 30.67 37.54 38.38 36.23 20.84 29.45 32.14

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

251

Fig. 12. CO2 diffusivity and coal porosity according to depth; for the coals of the Sydney Basin listed in Table 4.

Fig. 14. CO2 diffusivity (diffusion coefficient) and porosity according to rank for coals listed in Table 4.

The adsorption capacity according to depth can be also expressed mathematically by a power equation:  a h C ¼ C0 ð15Þ h0

coal depth of h0. For the data shown in Fig. 11 the best fit values are: C0 = 30.1 m3/t, h0 = 400 m and α = 0.56. Assuming that CO2 could be potentially sequestered in all ‘inferred’ resources of NSW, and that the adsorption capacity of the coal is 37.8 m3/t (an underestimation for coal > 600 m deep) about 4.2 giga tonnes of CO2 can be sequestered.

where C is the adsorbed capacity (m3/t) at a coal depth of h (m) and C0 is the adsorbed capacity at a reference

4.3. Diffusivity and porosity measurements Diffusivities of CO2 for 15 coals from 9 seams were measured using the diffusivity measurement system described in Section 3.2. The porosity of these coals using the mercury intrusion technique described in Section 3.3 was also measured; these values relate to the meso- and macropores. As summarised in Table 4, the porosity for these coals varies from 3.95 to 9.88% whereas the diffusion coefficient varies from 1.18 × 10− 6 to 10.20 × 10− 6 cm2/s.

Table 5 Physical properties of the main gases occurring in coal CO2 a

Kinetic diameter (A°) Gas phase density (g/cm3)b Adsorbed phase density (g/cm3) Boiling point (°K)c Fig. 13. CO2 diffusivity (diffusion coefficient) and porosity according to volatile matter content for coals listed in Table 4.

a

CH4

N2

3.30 3.80 3.65 1.830 × 10− 3 0.665 × 10− 3 1.165 × 10− 3 0.915 0.415 0.804 194.5

113.3

77

After Mahajan (1991) and Walker and Mahjan (1978), bat Australian standard conditions (20°, 101.325 kPa), cat atmospheric pressure.

252

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

Table 6 Adsorption isotherms according to gas type for coals from Sydney Basin; VL is the Langmuir volume and PL the Langmuir pressure parameters, ‘coal as received’ CO2

CH4

N2

Adsorption capacity ratio

Coal sample

PL VL PL VL PL CO2/ CO2/ VL (m3/t) (kPa) (m3/t) (kPa) (m3/t) (kPa) CH4 N2

Coal1a Coal15 Coal17 Coal26

46.92 38.29 45.85 30.27

4469 2018 2233 2667

16.62 24.65 25.07 18.43

3240 2667 3061 2164

5.83 9.17 7.82 8.45

3297 2055 2399 2656

2.8 1.6 1.8 1.6

8.0 4.2 5.9 3.6

a

Contact metamorphosed coal.

As illustrated in Fig. 12, the diffusion coefficient and porosity have no clear relationship with depth. These data are also presented according to volatile matter contents (Fig. 13) where that the porosity appears to generally decrease with decreases in volatile matter content. The two contact metamorphosed coals have much higher porosity for their VM contents. Diffusion coefficient and porosity plotted against coal rank (Fig. 14) show that porosity appears to generally decrease with coal rank, but there is no clear relationship between diffusivity and rank for the coals for which both rank and diffusivity data are available. 5. Comparison of properties of main seam gases Australian coal seams mainly contain methane, carbon dioxide and lesser amounts of nitrogen. The physical

Table 7 Gas diffusivity in coal according to gas type for coals from Sydney Basin; D is gas diffusivity Gas type

CO2

CH4 6

2

CO2/CH4 6

2

Coal sample

D × 10 (cm /s)

D × 10 (cm /s)

Diffusivity ratio

Coal1 Coal15 Coal17 Coal26

7.73 8.13 2.25 10.24

3.63 4.46 1.23 4.88

2.13 1.82 1.84 2.10

properties of these gases are shown in Table 5. Measurements of adsorption isotherms using all the gases were undertaken on four of the coals including one of the contact metamorphosed coals (Table 6). Gas adsorption capacity of these coals, represented by their Langmuir volumes, show that, for the non-metamorphosed coals, the adsorption capacity is 1.6 to 1.8 times higher for CO2 than for CH4 and the adsorption capacity of CO2 compared to N2 is 3.6 to 5.9 times higher. The enhanced adsorption capacity of coal for CO2 is in accordance with the stronger sorption of this gas onto coal surfaces and with its higher boiling point (Table 5). For the contact metamorphosed coal these ratios increase significantly to 2.8 for CO2:CH4 and 8.0 for CO2:N2. In Fig. 15 the graphs of adsorption curves for coal17 is illustrated showing the difference in adsorption capacity according to gas type. In order to study preferential gas diffusion in coal, diffusivities of methane and carbon dioxide were also measured for these four coals (Table 7). The results show that the diffusion of CO2 takes place twice as quickly than for CH4. The higher diffusion coefficient may be related to the smaller kinetic diameter of CO2 gas molecules. In addition the polarity of CO2 molecules and their elongated shape would facilitate transport through small pores. 6. Summary and conclusions

Fig. 15. Adsorption of pure CO2, CH4 and N2 on coal (coal17), coal ‘as received’.

CO2 storage and diffusivity properties of 27 coal samples from 17 coal seams in Sydney Basin, NSW, Australia were measured using a gravimetric method and a new diffusivity system based on Fick's diffusion law. The measurements were undertaken at CO2 subcritical pressure and temperature conditions, namely at gas pressures below 6 MPa and at a temperature of 27 °C for most of the coals. The coals analysed are of high volatile bituminous to low volatile bituminous rank (Ro max = 0.66 to 1.45%) and they were collected from Permian coal seams located at depths between 27 m and 723 m. The highest adsorptive capacity corresponds to the deepest, highest rank coal, the lowest corresponds to a shallower coal but not the lowest rank coal.

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

Langmuir isotherms satisfactorily represent the adsorption data, with a relative error of less than 2.7% for all except one result. Therefore, the Langmuir mono-layer mechanism can represent CO2 adsorption for pressures of up to 6 MPa. The maximum CO2 adsorption capacity of these coals varies from about 40 to 80 m3/t (daf). The projected CO2 storage capacity for these coals at ‘as received’ conditions for the depths of up to 723 m is 6 to 51 m3/t. The CO2 diffusion coefficients for 15 samples from 9 coal seams and from 8 locations vary from 1.2 × 10− 6 to 10.2 × 10− 6 cm2/s. For these samples no relation between the diffusivity and the depth or rank of coals is apparent. For four coals, measurement of adsorption and diffusion was undertaken using CH4 and N2 in addition to CO2. The results show that the CO2 storage capacity is about twice that of CH4 and six times that of N2. The diffusivity of CO2 in coal is about twice that of CH4. Acknowledgment The authors wish to thank the coal mining industry in Sydney Basin, Australia for providing the coal samples used in this study. Also our thanks are extended to Neil Sherwood and Nigel Russell of CSIRO Petroleum for their invaluable editorial comments. References Australian Black Coal Statistics, 2004. Published by Coal Services Pty Ltd and Queensland Department of Natural Resources and Mines. Atkins, P.W., 1978. Physical Chemistry. Oxford University Press. ISBN: 0 19 855148 7. Clayton, J.L., 1998. Geochemistry of coalbed gas — a review. International Journal of Coal Geology 35, 159–173. Crosdale, P., Saghafi, A., Williams, R., Yurakov, E., 2005. Interlaboratory comparative CH4 isotherm measurement on Australian coals. In: Beeston, J.W. (Ed.), Bowen Basin Symposium 2005, The Future for Coal: Fuel for Thought, Yeppoon, Qld., Geological Society of Australia Coal Geology Group and the Bowen Basin Geologists Group, pp. 273–277. Faiz, M.M., Saghafi, A., Sherwood, N.R., 1999. Higher hydrocarbon in southern Sydney Basin coals. In: Mastalerz, M., Glikson, M., Golding, S.D. (Eds.), Coal Bed Methane: Scientific, Environmental and Economic Evaluation. Kluwer Acadamic Publishers. ISBN: 0-7923-5698-5, 233–255. Griffith, M., Hirst, W., 1944. Proceedings of a Conference on the Ultra Fine Structure of Coals and coke. The British Coal Utilisation Research Association, London, p. 80. Harpalani, S., Ouyang, S., 1999. A new laboratory technique to estimate gas diffusion characteristics of coal. Int. Coalbed Methane Symposium, May 3–7, Tuscaloosa, Alabama, pp. 141–152. James, A.T., Burns, B.J., 1984. Microbial alteration of subsurface natural gas accumulations. Bulletin of the American Association of Petroleum Geologists 68, 957–960.

253

Kotarba, M.J., Rice, D.D., 1995a. Composition and origin of coalbed gases in the Lower Silesian coal basin, NW Poland. Report from Research Cooperation within the U.S.-Polish Maria SklodowskaCurie Joint Fund II, Origin and Habitat of Coal Gases in Polish Basins, Isotopic and Geological, pp. 69–78. Kotarba, M.J., Rice, D.D., 1995b. Composition and origin of coalbed gases in the Upper Silesian and Lublin coal basins, Poland. Report from Research Cooperation within the U.S.-Polish Maria Sklodowska-Curie Joint Fund II, Origin and Habitat of Coal Gases in Polish Basins, Isotopic and Geological Approach, pp. 79–89. Lama, R.D., Bodzinoy, J., 1996. Outburst of Gas, Coal and Rock in Underground Coal Mines. R.D. Lama and Associates, Wollongong, NSW, Australia. Lama, R., Saghafi, A., 2002. Overview of gas outbursts and unusual emissions. The Proceedings of Coal 2002, 3rd Australian Coal Operators' Conference, The Australasian Institute of Mining and Metallurgy, 6–8 February 2002, pp. 74–88. Levine, J.R., 1993. Coalification: the evolution of coal as source rock and reservoir rock for oil and gas. In: Law, B.E., Rice, D.D. (Eds.), Hydrocarbon from Coal. American Association of Petroleum Geologists Studies in Geology, vol. 38, pp. 39–77. Mahajan, O.M.P., 1991. CO2 surface area of coals: the 25-year paradox. Carbon 29, 735–742. Moore, R.D., Hanes, J., 1980. Bursts in coal at Leichhardt colliery, central Queensland and the apparent benefits of Mining by short firing. The Proceedings of the Symposium on the Occurrence, Prediction and Control of Outbursts in Coal Mines. The AusIMM Southern Queensland Branch, pp. 72–83. Nandi, S.P., Walker, P.L., 1970. Activated diffusion of methane in coal. Fuel 49, 309–323. Ozdemir, E., Morsi, B.I., Schroeder, K., 2003. Importance of volume effects to adsorption isotherms of carbon dioxide on coals. Langmuir 19, 9764–9773. Reucroft, P.J., Patel, H., 1986. Gas-induced swelling in coal. Fuel 65, 816–820. Saghafi, A., 1996. Numerical modelling of gas emission rate from a longwall goaf in a thick seam. The 26th International Symposium on Application of Computers and Operations Research in the Mineral Industry. The Pennsylvania State University, USA, pp. 209–211. Saghafi, A., 2003. Aspects of gas storage and flow properties of Australian coals. The 2nd Annual Australian Coal Seam and Mine Methane Conference, 19–20 February 2003, Brisbane, Australia. Saghafi, A., Williams, D.J., 1998. Safe mining in outburst conditions and accuracy of gas content measurements. Proceedings of the Int. Mining Tech. Symposium, Coal Mine Safety and Health, Chongqing, China, 14–16 Oct. 1998, pp. 93–104. Saghafi, A., Jeger, C., Tauziede, C., Williams, R.J., 1987. A new computer program for gas flow into drainage boreholes. 22nd International Conference of Safety in Mines Research Institutes, Beijing, China, pp. 147–158. Smith, J.W., Gould, K.W., 1980. An isotopic study of the role of carbon dioxide in outbursts in coal mine. Geochemical Journal 14, 27–32. Smith, J.W., Pallasser, R.J., 1996. Microbial origin of Australian coalbed methane. Bulletin of the American Association of Petroleum Geologists 80, 891–897. Smith, J.W., Gould, K.W., Hart, G.H., Rigby, D., 1985. Isotopic studies of Australian natural and coal seam gas. Bulletin of Australasian Institute of Mining and Metallurgy 290, 43–51. Standards Association of Australia, 1999. Guide to the Determination of Gas Content of Coal-direct Desorption Method. Australian Standards AS 3980-1999.

254

A. Saghafi et al. / International Journal of Coal Geology 70 (2007) 240–254

Stevenson, M.D., Pinczewski, W.V., Somers, M.L., 1991. Adsorption/ desorption of multi-component gas mixtures at in-seam conditions. Society of Petroleum Engineers 23026, Society of Petroleum Engineers Asia Pacific Conference, Nov. 4–7, Perth, Australia. Thomas Jr., J., Damberger, H.H., 1976. Internal Surface Area, Moisture Content and Porosity in Illinois Coal: Variation with Coal Rank. Illinois State Geological Survey Circ., vol. 493. 38 pp. van Krevelen, D.W., 1993. Coal: Typology–Physics–Chemistry– Constitution. Elsevier, Amsterdam. 979 pp. Walker Jr., P.L., Mahjan, O.P., 1978. Methane diffusion in coals and chars. In: Karr, C. (Ed.), Analytical Methods for Coal and Coal Products, vol. 1. Academic Press, pp. 163–188. Williams, D.J., Saghafi, A., Drummond, D.B., Roberts, D.B., 1992. Development of new equipment for rapid determination of coal gas content. Symposium on Coalbed Methane Research and Development in Australia, Townsville, Australia, vol. 4, pp. 21–30. Williams, D.J., Saghafi, A., Roberts, D., 1995. Some observations on the mixed gas desorption from coal. The Proceedings of the

International Symposium-cum-Workshop on Management and Control of High Gas Emissions and Outbursts in Underground Coal Mines. National Organising Committee of the Symposium, Wollongong, Australia, p. 387. Whiticar, M.J., Faber, E., Schoell, M., 1986. Biogenic methane formation in marine and freshwater environments, CO reduction vs. acetate fermentation–isotopic evidence. Geochimica et Cosmochimica Acta 50, 693–709. Yee, D., Seidle, J.P., 1993. Gas sorption on coal and measurement of gas content. In: Law, B.E., Rice, D.D. (Eds.), Hydrocarbons from coal. American Association of Petroleum Geologists Studies in Geology, Bulletin, vol. 38, pp. 203–218. Zwietering, P., van Krevelen, D.W., 1954. Chemical structure and properties of coal, IV-pore structure. Fuel 33, 331–337.