In this work, a Maxwell-Stefan model for high flux tubular silicalite-1 membranes for separation of CO2 from a CO2/H-2 mixture was developed. The model concerns tubular membranes operating in a counter flow module and includes transport through flow-through defects in the silicalite-1 film and pressure drop over the graded alumina support. Adsorption and diffusion parameters for perfect silicalite-1 crystals were taken from the literature. The flux and selectivity predicted by the model were in reasonably good agreement with experimentally observed data for a ZSM-5 membrane without any fitting of the model. However, the CO2 flux and selectivity measured experimentally for the ZSM-5 membrane were higher than that predicted by the model for a silicalite-1 membrane. The model was used to investigate a case with a 20,000 Nm(3)/d feed comprised of a 50/50 mixture of CO2/H-2 at pressure of 25 bar and a membrane temperature of 296 K. The permeate pressure was 1 bar and 90% of the CO2 permeated the membrane. In this case, the membrane permselectivity and CO2 flux varied along the length of the tubes between 20-26 and 950-396 kg/(m(2) h), respectively. Further, both defects and pressure drop over the support were shown to have an adverse effect on the selectivity, which indicates that membrane selectivity can be improved by reducing the flow-through defects and/or by preparing supports with less flow resistance. For a one-stage process, the required membrane area is as small as ca. 0.85 m(2) and the hydrogen loss through the membrane was 124%. For a two-stage process the required membrane area almost doubled to 1.6 m(2), however the hydrogen loss through the second membrane is reduced to as little as 2.5%. In summary, this work shows that high flux zeolite membranes may be an interesting option for CO2 removal from synthesis gas. (C) 2014 The Authors. Published by Elsevier B.V.