In order to determine the material fracture resistance necessary to provide adequate control of the propagation of ductile fracture in a pipeline, a knowledge of the decompression wave speed (W) following the quasi‐instantaneous formation of an unstable, full‐bore rupture is necessary. The thermodynamic and fluid dynamics background of such calculations is understood, but predictions based on specific equations of state need to be validated against experimental measurements. A program of tests has been conducted using a specially constructed shock tube to determine the impact of impurities on W in carbon dioxide (CO2), so that the results can be compared to two existing theoretical models. In this paper, data and analysis results are presented for four tests involving simulated anthropogenic CO2 mixtures containing Argon (Ar) as the primary impurity. These mixtures represent typical oxy‐fuel CO2 capture technology with or without a purification train. Comparisons of the experimentally obtained W with predictions by two commonly used equation of state (EOS) models were also made. Generally, the GERG‐2008 EOS exhibits better overall performance than the Peng‐Robinson (PR) EOS when compared to the experimental results. The bubble point curve as predicted by GERG‐2008 was always at a higher pressure than that predicted by PR for these mixtures. This resulted in the plateau pressure prediction by GERG‐2008 being higher in cases where the isentropes intersect the phase envelope on the bubble point side. An example of pipeline material toughness required to arrest ductile fracture is presented which shows that prediction by GERG‐2008 is more conservative, and is therefore recommended.