(640f) Benchmarking Of Hydrogen Selective Membranes Through Experiments and Modelling

Palladium (-alloy) membranes are a promising option for hydrogen separation in industrial hydrogen production and in pre-combustion CO₂ capture [1,2]. Given the state of the art, benchmarking of membranes from different vendors is a crucial step in the demonstration of the maturity and performance under industrially relevant conditions. A multitude of membrane manufacturers and developers exists, yet at the moment, performance figures have been published at different conditions, and including the impact of externals, e.g. different membrane module designs. The present contribution provides an approach to assess the performance of H₂-selective membranes under industrially relevant conditions, independent from the impact of module design. The strategy presented here, consisting of experiments and model development, will be a valuable tool in benchmarking membrane performance. In order to evaluate membrane performance without artifacts induced by the experimental module, a membrane model has been devised. Systematic experimental tests were performed (extended for up to 1500 hours on stream) with membranes that were delivered by a number of different suppliers, allowing for model development and experimental validation. The staged experimental approach allows for investigating systematically the effects of intrinsic permeation, hydrodynamics, and concentration polarization. Membranes are tested at a scale of 300 to 800 cm² in a pilot setup [3] capable of accommodating up to 6 membranes of 100 cm length. Experiments have been performed that started with pure H₂ feed, without sweep, subsequently followed by introducing N₂ on the feed side, and N₂ sweep gas. Using a phenomenological description for the palladium layer and the dusty gas model for the membrane support, coupled to a 2D Navier–Stokes solver with a convection-diffusion equation to account for possible concentration polarisation, all relevant mass transfer resistances are adequately modeled [4]. For the conditions investigated, the main resistances to mass transfer are concentration polarisation in the retentate, hydrogen permeation through the metallic palladium layer, and a diffusional resistance in the support layer. Finally, the effect of inhibition by syngas components was measured for representative gas compositions and incorporated in the permeation equation. The 2D model was then used for evaluating and comparing the intrinsic membrane performance and for predicting membrane performance at commercial scale. In parallel, membrane and membrane-module robustness data extracted from the long term stability experimental testing program provided valuable insight in the technological status.

References

[1]. N.W. Ockwig and T.M. Nenoff, Chemical Reviews 107 (2007) 4078-4110.

[2]. G.Q. Lu, J.C Diniz da Costa, M. Duke, S. Giessler, R. Socolow, R.H. Williams, T. Kreutz, Journal of Colloid and Interface Science 314 (2007) 589.

[3]. D. Jansen, J.W. Dijkstra, Energy Procedia, 1 (2009) 253.

[4]. J. Boon, J.A.Z. Pieterse, J.W. Dijkstra, M. van Sint Annaland, Int. J. of Greenhouse Gas Control 11 (2012) 122.