COAL CHARACTERIZATION FOR CARBON DIOXIDE SEQUESTRATION PURPOSES G. Di FEDERICOI, I. CAMPLONEI, S. BRANDINII, J. BARKERII Centre for CO2 Technology, University College London I Department of Chemical Engineering, UCL, Torrington Place, WC1E 7JE, London, UK II Department of Earth Sciences, UCL, Gower Street, WC1E 6BT, London, UK 1. Introduction The greenhouse gas effect helps to regulate the temperature of our planet. It is the result of heat adsorption by a number of gases in the atmosphere called greenhouse gases because they effectively ‘trap’ heat in the lower atmosphere. Water vapour is the most abundant greenhouse gas, followed by carbon dioxide and methane. Without a natural greenhouse gas effect, the temperature of the Earth would be about –18°C instead of its present 14°C [1]. So, the concern is not with the fact that there is a greenhouse gas effect, but whether human activities are leading to an enhancement of this effect. Reduction in CO2 emissions is the most quoted solution for global warming in the short time. At the third Conference of Parties of the Framework Convention on Climate Change (FCCC) in Kyoto (1997), developed countries agreed to stabilise emissions of greenhouse gases through a 5.2% reduction from 1990 levels by 2008-12. In order to stabilise atmospheric carbon dioxide concentrations it will be necessary to reach more than 60% reduction by 2050 [2]. TABLE 1: CO2 emissions in 1996 and reductions agreed in Kyoto [3] Nation USA Russian Federation Germany UK Canada Poland Total emission (1996) million tonnes/year 5043 1500 924 571.4 507 373 Reduction required % 7 0 21 12.5 6 6 1 2. CO2 Sequestration Carbon dioxide sequestration is one of the options that are being considered for the reduction of greenhouse gas emissions. It would enable the world to continue to use fossil fuels and at the same time reduce emissions from their combustion. It involves two main stages that are the gas capture and its storage in suitable geological structures. In the capture operation CO2 is separated from flue gases produced in combustion processes using adsorption, absorption, cryogenic or membrane systems. The storage is realised by injecting the carbon dioxide in the coal seams, depleted gas reservoirs, depleted oil reservoirs, etc [4]. The present work is focused on the characterisation of coal for CO2 sequestration and simultaneous production of methane. The anthropogenic methane contained in coal can be regarded as a significant source of energy. This methane is usually referred to as coal-bed methane (CBM) in the literature. CBM is retained in coal in three ways: first, as free gas within the pore space or fractures in coal; second, as adsorbed molecules on the organic surface of coal; and third, dissolved in groundwater within the coal. The depletion techniques used to recover CBM allow the recovery of 50% of the total amount contained in the reservoir. Carbon dioxide is more strongly adsorbed in coal than methane therefore the injection of carbon dioxide will enhance the production of methane with recoveries of up to 100% [5]. This process is often referred to as enhanced-CBM (ECBM). 2.1. GAZ TRANSPORT IN COAL The gas transport through the coal reservoir is governed by the permeabilitycontrolled flow in the macropore system and by the diffusion-controlled flow in the micropore system. The diffusion-controlled flow obeys Fick’s law. Clarkson and Bustin [6] estimated that the inverse of the CO2 effective diffusion time constant, D` D l 2 , for Cretaceous Gates Formation is between 6∙10–3 s–1 and 3∙10–2 s–1 in the macropores and between 2∙10–3 s–1 and 4∙10–3 s–1 in the micropores. The permeability-controlled flow is generally considered to be laminar, so it obeys D’Arcy’s law [7]. m k P (1) with the gas density, k the permeability, P the pressure. Permeability is a measure of coal ability to transmit fluids therefore knowing the coal permeability the injection pressure can be estimated. Typical values for the permeability are between 0.1 and 250 md for US coals [8]. 2 The permeability of a coal seam varies with effective external stress, which can be defined as the difference between the external stresses applied on a rock and the internal pressure within the pore space of the rock. Increasing the external stress the permeability decreases because the system shrinks and the pressure drops in the matrix increase. Coal matrix shrinkage and the resulting change in cleat or fracture system porosity can have a profound effect on reservoir permeability and thus on production (or injection) performance. In fact, gas desorption causes a decrease in internal pressure, so it is expected that this would cause an increase in the effective stress and hence a decrease in permeability. Another important effect that occurs in the sequestration of CO2 in coal is the swelling of the structure. Coal swells upon exposure to carbon dioxide and the extent of expansion increases with increasing carbon dioxide pressure. The application of the external stress limits the extent of coal swelling. In this project the swelling is evaluated measuring the radial and the axial variation for the sample size under various stress conditions. It will be useful to develop a physical model to predict the swelling and the shrinkage in a coal reservoir and the associated changes in permeability in order to predict the maximum amount of CO2 that can be sequestered. Most available laboratory data, show that CO2 adsorption causes more strain and swelling than CH4 and N2 because it is adsorbed in larger concentration by coal and also suggest that CO2 causes more swelling on a unit of concentration basis [9]. 2.2. COAL ADSORPTION The sequestration process is controlled by the absorptive capacity of coal in terms of the maximum amount adsorbed and its rate of sorption. Data available in the literature is sparse but provides insight into the applicability of simple isotherm models. Experimental data for Fruitland Coal A [10] show favourable shaped isotherms with carbon dioxide more strongly adsorbed than methane and nitrogen (Figure 1). This behaviour is the driving force for the ECBM process in which CO2 replaces CH4. Figure 1 shows that the experimental data can be correlated using the Langmuir (eq. 2) and the truncated virial (eq. 3) isotherms. bP 1 bP (2) bp exp 1 A1q q (3) q qs where q is the quantity sorbed, P the pressure. qs, b, A1 are the models’ parameters determined by regression of the single component data. 3 Amount adsorbed (mol/g-coal) 0.0012 Virial Langmuir 0.001 carbon dioxide 0.0008 methane 0.0006 0.0004 nitrogen 0.0002 0 0 20 40 60 80 100 120 140 Pressure (bar) Figure 1. Single component sorption isotherms for Fruitland Coal A at 115ºF. Amount adsorbed (mol/g-coal) 0.0012 0.001 0.0008 0.0006 methane carbon dioxide 0.0004 0.0002 0 0 0.1 0.2 0.3 0.4 0.5 0.6 EVP-scaled Figure 2: Single component sorption isotherms for Fruitland Coal A at 115ºF with scaled pressures. The Langmuir parameters obtained from the Fruitland Coal A experimental data show a molecule dependent saturation capacity which is thermodynamically inconsistent [11]. For nitrogen qs = 7.416·10–4 mol/g-coal, methane qs = 1.014·10–3 mol/g-coal and carbon dioxide qs = 1.327·10–3 mol/gcoal, therefore the Langmuir model should be considered only as a useful fitting function. It is interesting to note [10] that by plotting the data on a reduced 4 pressure basis, p/p*, where p* is the extrapolated vapour pressure (EVP) using a simple two constant Clapeyron form, CO2 and CH4 isotherms are in reasonable agreement as shown in figure 2. Using the Ideal Adsorbed Solution theory it is possible to obtain a reasonable agreement with the experimental binary adsorption data as shown in Table 2. TABLE 2. CH4-N2 equilibrium data [10]. Pressure (bar) 25.77 Gas Total sorption Composition Experimental ymethane (mol/g) 0.6865 4.21E-04 Total sorption Total sorption IAS-Langmuir IAS-Virial (mol/g) (mol/g) 4.56E-04 4.52E-04 55.44 0.7085 6.20E-04 6.34E-04 6.23E-04 83.87 0.7203 7.07E-04 7.16E-04 7.23E-04 27.48 0.0832 2.55E-04 2.43E-04 2.52E-04 57.26 0.0906 3.84E-04 3.72E-04 3.76E-04 82.56 0.0954 4.46E-04 4.38E-04 4.49E-04 25.57 0.1478 2.63E-04 2.62E-04 2.75E-04 58.62 0.1633 4.32E-04 4.21E-04 4.27E-04 83.94 0.1713 5.22E-04 4.90E-04 5.03E-04 3. Design of test cell With the aim of characterizing coal for sequestration purposes, a test cell has been designed and assembled. What makes this a challenging problem is that coal will undergo swelling as a consequence of CO2 adsorption, thus modifying the solid properties, in particular permeation [12]. The coal can swell by 1-4% in volume and it is function of the mechanical stress state [13,14] and this swelling can have a large effect on the coal permeability. Therefore it is necessary to reproduce, as close as possible, the stress conditions of the coal sample in the coal reservoir. For typical UK coal the pressure along the bedding plane is lower than the overburden pressure (i.e. along the vertical direction) by a factor of 0.8 to 1 [15]. A schematic diagram of the test cell is shown in Figure 3. The different stresses are obtained by flowing water at different pressures in the two compartments, 5 thus obtaining the required pressure ratios. The water is also used to maintain the cell at isothermal conditions, which may vary from 25 to 60 ºC. Two HPLC pumps are used to maintain the water pressure constant and a schematic diagram of the system is shown in Figure 4. Figure 3. Test cell. Figure 4. Experimental set-up. To investigate the carbon dioxide adsorption process for a given coal sample, the gas is injected from 1 (Fig.3) and the outlet stream 2 is analysed. The gas injected can be pure carbon dioxide or its mixtures with nitrogen to reproduce streams coming from a power plant. The gas pressure can be varied from atmospheric to 120 atm. In order to quantify the extent of the swelling a measurement of the piston 5 displacement is used to evaluate the variation in the longitudinal length during the experiment. It is also possible to measure the diameter changes of the sample 6 using a strain gauge belt. The use of multiple strain gauge belts to follow the propagation of the swelling along the axis is being considered. When combined with unconstrained adsorption isotherm measurements this test cell should provide the necessary experimental information to characterize coal at the coal matrix scale. The results will be included in a large-scale reservoir simulator currently being developed at UCL’s Earth Sciences Department. 6 4. Conclusions Experiments will be performed in order to evaluate the kinetics of adsorption/desorption, the effect of the stress state and swelling on permeability and how it affects the coal structure. They will be carried out with nitrogen, carbon dioxide and methane as pure components or mixtures. From the analysis of the results it will be possible to understand and predict the swelling phenomenon and lead to a physical model for the sequestration process that will consider adsorption, permeability and swelling. 5. Notation CBM Coal Bed Methane ECBM Enhanced Coal Bed Methane EVP Extrapolated Vapour Pressure FCCC Framework Convention on Climate Change SCF Standard Cubic Feet A area of cross-section of flow D` effective diffusivity D diffusivity l diffusion path length m mass flow rate 2 k permeability (1 md = 9.8692 10–4 ) fluid viscosity P pressure pout outlet pressure p pressure drops along the coal sample pav the average pressure between the injection pressure and outlet pressure gas density Acknowledgements Financial support from the Leverhulme Trust (Philip Leverhulme Prize), Royal Society (Royal Society Wolfson Research Merit Award) and EPSRC is gratefully acknowledged. 7 6. References 1. National Data Climate Center web site: http://www.ncdc.noaa.gov 2. Royal Commission on Environmental Pollution Report 22, June 2000. Energy – The Change in Climate 3. European Monitoring and Evaluating Programme web site: http://www.emep.no 4. National Energy Technology Laboratory website: www.netl.doe.gov 5. Pashin J. C., Groshong Jr. R.H., Carrol R.E., Enhanced coalbed methane recovery through sequestration of carbon dioxide: potential for a marketbased environmental solution in the Black Warrior Basin of Alabama, in www.netl.doe.gov. 6. Clarkson C.R. and Bustin R.M (1999) The effect of pore structure and gas pressure upon the transport properties of coal: a laboratory and modeling study. 2. Adsorption rate modeling, FUEL 78 1345-1362. 7. Harpalani S. and Chen G. (1997) Influence of gas production induced volumetric strain on permeability of coal, Geotechnical and Geological Engineering 15, 303-325. 8. Jones, S.C. (1992) The profile permeameter; a new fast accurate minipermeameter , Paper SPE 24757 presented at the 1992 SPE Technical Conference and Exhibition, Washington, Oct 4-7. 9. Pekot L.J. and Reeves S.R. Modelling (November 2002) Coal matrix shrinkage and differential swelling with CO2 injection for enhanced coalbed methane recovery and carbon dioxide sequestration applications, U.S. Department of Energy. 10. Chaback J.J., Morgan W.D. and Yee D. (1996) Sorption of nitrogen, methane, carbon dioxide and their mixtures on bituminous coals at in-situ conditions, Fluid Phase Equilibria 117, 289-296. 11. D.B. Broughton(1948) Ind. And Eng. Chem. 40 (8) 1506. 12. Harpalani S. and Schraufnagel R.A.(Sept 1990) Influence of matrix shrinkage and compressibility on gas production from coalbed methane reservoirs SPE 20729, Procs. 65th Ann. Tech. Conf., New Orleans, LA,. 13. Reucroft P.J. and Patel H. (1986) Gas-induced swelling in coal, FUEL, 65, June. 14. Reucroft P.J. and Sethuraman A. R. (1987) Effect of pressure on carbon dioxide induced coal swelling, Energy & Fuels, 1,72-75. 15. Sinka I.C. (1997) An investigation into rock and reservoir properties of coal with special reference to stimulated coalbed methane well performance, PhD thesis. 8