naar top
Logo Print


In membrane separation of hydrogen from other gases, hydrogen dissolves in the polymeric membrane and then diffuses to the other side. The separation is partial-pressure driven; so the pressure drop across the membrane is considerable. This technology is used in oil refineries to produce hydrogen from hydrogen production units such as naphtha reforming or natural gas steam reforming.

PSA versus membrane separationComparison with alternatives

Alternative conventional separation technologies are Pressure Swing Adsorption (PSA) and cryogenic distillation. The preferred method depends on the available pressure of the feed and the required hydrogen purity. Cryogenic distillation is chosen when hydrogen is to be separated from hydrocarbons only [1]. Otherwise PSA or membranes are considered. PSA is selected when a high purity (>96 %) is required, while for lower purities membrane separation is usually preferred [1]. Table 1 shows a comparison between PSA and membranes. It shows that membrane separation is preferred for purities of 96% and less, as it is more attractive with regards to recovery, energy, reliability [2] and cost [3]. The compression cost for the membrane system is in general higher than for the PSA system and depending on the local refinery requirements can overrule the other cost [1, 2].

Technology providers

Installations for hydrogen membrane separations are provided by technology providers Air Liquide, UOP and Air Products. The first commercial implementation was in 1977, provided by Air Liquide. They have installed over 90 membrane systems with their MEDAL membrane technology [4].

UOP installed over 80 membrane systems, their first one in 1985 [5]. Air products with their PRISM technology installed their first unit in 1977 [6]. They did not disclose how many installations they provided. Their total number of gas plants installed is 2.500. If they have the same membrane/PSA ratio as UOP (0.08), then they have 200 membrane systems installed. They started installing membrane systems 10 years earlier than UOP, so it is indeed likely that they have more systems installed. The total number of hydrogen membrane systems installed is then likely to be over 370. 

Total cost reduction and reliability will be the driving forces for choosing membranes over PSA when purity is less important and compression cost does not dominate. The total cost reduction is obtained by higher yield, lower energy cost and lower capital cost. Conceptual design of this separation method is described in text books [2, 7, and 8].

The technology providers have the knowledge for both concept design and detailed design for commercial-scale hydrogen membrane applications. These companies also have PSA design knowledge in-house and thus can also evaluate PSA versus membranes for any specific refinery case.

It was very hard to find information about scale-up methods for the commercialization of hydrogen membranes. Most articles do not mention anything about pilot plant tests or scale-up. Baker mentions in a single sentence in his review article that for a particular application pilot plant tests have been performed [9]. Pabby in his handbook on membrane applications does not treat scale-up or pilot plant tests in general. He only mentions that for a specific hydrogen sulphide removal from crude gasoline fractions by membranes that lab scale, pilot plant scale and a demonstration plant have been employed leading finally to commercial scale implementation [7]. It is therefore likely that the scale-up method employed by the technology providers for hydrogen membrane separation has been 'brute force' [10] in combination with Pabby's step-by-step cautious innovation, meaning that a module of membrane tubes has been tested for a real feed in a refinery and then a small scale commercial scale application has been implemented with all design parameter values the same. Subsequently, larger scale membrane systems have been installed.

By Jan Harmsen (Harmsen Consultancy BV)


[1] Z. Rabiei, Hydrogen management in refineries, Petroleum & Coal, 54 (2012) 357-378.
[2] B. Freeman, et al., Membrane Gas Separation, J. Wiley, Hoboken, 2011.
[3] R.L. Schendel et al., Permeation; membrane separation, Hydrocarbon processing, Aug. 1983.
[5] UOP, Processing solutions refining hydrogen management,
[7] A.K. Pabby, S.H. Rizvi, A.M. Sastre Requena, Handbook of Membrane Separations: Chemical, Pharmaceutical, Food and Biotechnological Appli­cations, CRC Press, Boca Raton, USA, 2008.
[8] M.K. Turner, Effective Industrial Membrane Processes: Benefits and Opportunities, Springer, Netherlands, 2012.
[9] R. W. Baker, Review: Future Directions of Membrane Gas Separation Technology, Ind. Eng. Chem. Res. 41 (2002) 1393-1411.
[10] J. Harmsen, Process intensification: its drivers and hurdles for commercial implementation, Chemical Engineering and Processing 49 (2010) 70-73.