by Ronan K. McGovern

Inspired by the thinking of Randy Truby and Jim Birkett to find a single figure of merit for seawater RO membranes, I thought I’d see if I could compare the performance of a selection of membranes currently on the market.

To begin I plotted the nominal flux and rejection of 8” membranes from Toray, Filmtec, Hydranautics and NanoH2O – note the membranes that compete directly (e.g. at 99.8% rejection and a flux of about 38 lmhb). The problem I have with data in this form is that neither flux nor rejection are properties of the membranes. Furthermore, rejection depends upon flux – with higher fluxes tending to increase the relative amount of water passing compared to salts and thus increasing rejection. Therefore, I took the data from Fig. 1 and computed the water permeability and salt permeation coefficient for each membrane*.  I then took this data and generated Fig. 2.

test

Fig. 1: Nominal flux and rejection of NanoH2O Qfx SW 400 ES, Qfx SW 400 R; Filmtec SW30HR-380, SW30HR LE-400, SW30XLE-400I; Toray TM820C-400, TM820M-400, TM820E-400, TM820R-400, TM820V-400, TM820K-400; Hydranautics SWC4+, SWC5, SWC6. Overlapping points are displaced for them to be visible.

Fig. 2 has a number of remarkable points:

  • We typically regard both membrane permeability and the salt permeation coefficient to scale linearly with the thickness of the active layer (the solution diffusion model). For this reason, any straight line drawn from the origin represents a membrane with the same active layer chemical composition but of decreasing thickness as one moves away from the origin.
  • Following the above logic, the three Hydranautics membranes appear to have similar active layer chemistry but different thicknesses. The Toray membranes seem to include membranes of two different chemical compositions, as do the NanoH2O membranes.
  • Hydranautics, Toray and NanoH2O membranes show the expected trade-off between water permeability and salt permeation coefficient. An improvement in one leads to a dis-improvement in the other. By contrast, the Filmtec membranes do not illustrate this trend. This is surprising and begs the question as to why one would produce the membranes with lower permeability (it would be better to produce a thinner version of the most permeable active layer).
  • Clearly both membrane permeability and salt diffusion coefficient are both important to an RO system. For example, the former influences energy consumption and the latter influences the number of membrane passes required. However, in terms of characterizing the performance of the active layer, it seems that only the ratio of the permeability (A) to the salt permeation coefficient (B) is of importance.  I say this because, in solution-diffusion theory, if a membrane with a given ratio of A/B exists, then any membrane along a line joining it to the origin could be produced by varying the thickness of the active layer. For the purpose of this article, let us call this ratio the active layer performance ratio, with units of inverse pressure.
  • Interestingly, all four of the membrane manufacturers considered have membranes with an active layer performance ratio of about 17 inverse bar. Also interesting is the fact that the high rejection NanoH2O membrane was not just the high flux membrane adjusted to a higher thickness – it appears to have a fundamentally different chemistry that pushed out the technology line.
test

Fig. 2: Water permeability and salt permeation coefficients of the membranes shown in Fig. 1

In conclusion, the ratio of permeability to salt permeation coefficient seems to be a relevant metric for active layer performance. To some extent, the best membranes offered by manufacturers are not fundamentally different in terms of chemistry. It appears that membranes are differentiated more by the thickness of the active layer than by any fundamental difference in chemistry.

*Water permeability was computed by dividing the flux by the driving pressure difference (the applied hydraulic pressure minus the osmotic pressure of NaCl at 32,000 ppm – computed using activity coefficients from ). The salt permeation coefficient was computed by multiplying the salt passage (1 minus rejection) by the flux.

Note: This article first appeared as a post on the IDA Young Leaders Program Website.

Ronan K. McGovern is a postdoctoral associate in the Center for Clean Water and Clean Energy at MIT and serves as co-chair of the International Desalination Association Young Leaders Program.

References

Birkett, J., and R. Truby. “A Figure of Merit for Appreciating Improvements in RO Membrane Performance.” International Desalination Association News 16 (2007).

Robinson, Robert Anthony, and Robert Harold Stokes. Electrolyte solutions. Courier Dover Publications, 2002.