The permeability of coal samples from Pittsburgh Seam was determined using carbon dioxide as the flowing fluid. The confining pressure was varied to cover a wide range of depths. The permeability was determined as a function of exposure time of carbon dioxide while the confining stress was kept constant. The porosities of the coal samples were found to be very low and most of the samples had porosities less than 1%. The permeability of these coal samples was very low-less than 1 mu D. Since the objective of this study was to investigate the influence of CO(2) exposure on coal permeability, it was necessary to increase the initial permeability of the coal samples by introducing a fracture. A longitudinal fracture was induced mechanically, and CT scans were taken to ensure that the fracture was present throughout the sample and that the sample was not damaged otherwise during the process. In this study, the permeability of coal was determined by using pressure transient methods. Two types of pressure pulses were used: A-spike and Sine-6 pressure transients. It was first established that the permeability of fractured coal samples did not change with exposure time when an inert gas (Argon) was used as the fluid medium in the experiments. However, the permeability of coal samples decreased significantly when carbon dioxide was used as the fluid medium. This reduction can be attributed to the coal swelling phenomenon. The results show that the permeability reduction in fractured coal samples can be over 90% of the original value, and the exposure time for such reductions can range from 1.5 days up to a week, typically about 2 days under laboratory conditions. The permeability decreased significantly with the increase in confining pressure. The higher confining pressure appears to close internal fractures causing a reduction in permeability. (C) 2008 Elsevier B.V. All rights reserved.