SEQUESTERING CARBON DIOXIDE IN COALBEDS [report]

K.A.M. Gasem, Jr. R.L. Robinson, L.R. Radovic
2001 unpublished
Experimental Work: A second adsorption apparatus, utilizing equipment donated by BP Amoco, was assembled. Having confirmed the reliability of this additional experimental apparatus and procedures, adsorption isotherms for pure CO 2 , methane, ethane, and nitrogen on wet Fruitland coal and on dry activated carbon were measured at 319.3 K (115 °F) and pressures to 12.4 MPa (1800 psia). The addition of this new facility has allowed us to essentially double our rate of data production. In addition,
more » ... ction. In addition, adsorption isotherms for pure CO 2 , methane, and nitrogen on wet Illinois-6 coal and on activated carbon were measured at 319.3 K (115 °F) and pressures to 12.4 MPa (1800 psia) on our first apparatus. The activated carbon measurements showed good agreement with literature data and with measurements obtained on our second apparatus The Illinois-6 adsorption measurements are a new addition to the existing database. All the pure-fluid adsorption data show an expected uncertainty of about 3%. Adsorption from binary mixtures of methane, nitrogen and carbon dioxide at a series of compositions was also measured on the wet Fruitland coal at 319.3 K (115 °F), using our first apparatus. The nominal feed compositions of these mixtures are 20%/80%, 40%/60%, 60%/40%, and 80%/20%. The experiments were conducted at pressures from 0.69 MPa (100 psia) to 12.4 MPa (1800 psia). The expected uncertainty for these binary mixture data varies from 2 to 9%. A study addressing the previously-reported rise in the CO 2 absolute adsorption on wet Fruitland coal at 319.3 K (115 °F) was completed. Our additional adsorption 4 measurements on Fruitland coal and on activated carbon show that: (a) the Gibbs adsorption isotherm for CO 2 under study exhibits typical adsorption behavior for supercritical gas adsorption, and (b) a slight variation from Type I absolute adsorption may be observed for CO 2 , but the variation is sensitive to t he estimates used for adsorbed phase density. Model Development: The experimental data were used to evaluate the predictive capabilities of various adsorption models, including the Langmuir/loading ratio correlation, a two-dimensional cubic equation of state (EOS), a new two-dimensional (2-D) segment-segment interactions equation of state, and the simplified local density model (SLD). Our model development efforts have focused on developing the 2-D analog to the Park-Gasem-Robinson (PGR) EOS and an improved form of the SLD model. The new PGR EOS offers two advantages: (a) it has a more accurate repulsive term, which is important for reliable adsorption predictions, and (b) it is a segment-segment interactions model, which should more closely describe the gas-coal interactions during the adsorption process. In addition, a slit form of the SLD model was refined to account more precisely for heterogeneity of the coal surface and matrix swelling. In general, all models performed well for the Type I adsorption exhibited by methane, nitrogen, and carbon dioxide up to 8.3 MPa (average deviations within 2%). In comparison, the SLD model represented the adsorption behavior of all fluids considered within 5% average deviations, including the near-critical behavior of carbon dioxide beyond 8.3 MPa (1200 psia). Work is in progress to (a) derive and implement the micropore form of the SLD model, which would expand the number of structural geometries used to represent the heterogeneity of coal surface; and (b) extend the SLD model to mixture predictions. Accurate gas-phase compressibility (Z) factors are required for methane, ethane, nitrogen and carbon dioxide and their mixtures to properly analyze our experimental adsorption data. A careful evaluation of the current literature, led us to conclude that an adequate predictive capability for the mixture Z factors does not exist. Therefore, we have elected to develop such a capability using the Benedict-Webb-Rubin (BWR) equation of state. Specifically, we have used the available pure-fluid and binary mixture data to refit the BWR equation and improve its accuracy significantly; in general, the new BWR EOS parameters yield deviations in the Z factor within 0.2%. Coal Characterization: At Pennsylvania State University, we have completed determining CO 2 and methane adsorption properties for six coals of different rank. The coals used in this study are from the Argonne Premium sample bank, covering the rank range from lignite to low-volatile bituminous, including Beulah (lignite), Smith Roland (subbituminous), Illinois-6 (high-volatile bituminous), Pittsburg-8 (high-volatile bituminous), Stockton-Lewiston (medium-volatile bituminous), and Pocahontas (lowvolatile bituminous). Significant differences in CO 2 sequestration ability have been observed for different coals. Furthermore, when these differences are compared to the relative affinities of coals for CO 2 vs. methane, it is concluded that they are mostly due to differences in CO 2 uptakes on different coals. DE-FC26-98FT40426 DE-FC26-98FT40426 DE-FC26-98FT40426
doi:10.2172/812558 fatcat:yuseaqrhqzbpjcekhok6m7x42m