(154e) Surface Modification of Ultrafiltration Membranes : Where Are We Now ?

The surface modification of ultrafiltration membranes to increase their resistance to fouling started at the beginning of the 90â??. Lots of works have been published, with some of them turned into successful commercial products. If commercial membranes are nowadays more fouling resistant than ever, there is still room for improvement and we can try to understand how this could be achieved. Models borrowed from polymer physics can help understanding the behaviour of hydrophilic moieties attached by one end to a membrane surface. As for an example, a model known as the Alexanderâ??s model, further developed by de Gennes and co-workers, suggests that when the grafting density is is high enough, the polymer chains stretch towards the bulk solution and form a so called â??brushâ?. Then, the foulant molecules/particles can:
Diffuse through the brush, if their dimension is small compared to the average distance between the grafted polymers on a membrane surface
Compress the brush if they are too large to penetrate it.
This helps to define some criteria for the membrane modification to be efficient. In particular, the trade-off between length of the polymer chains and grafting density can be discussed. A comparison of the model predictions to data of adsorption of proteins on solid surfaces taken from the literature shows the relevance of this model and its limitation. The experimental minimum grafting density required to avoid the adsorption of proteins seems to be larger than the threshold density for which a â??brushâ? of polymerâ? would form at the surface of the membrane. We analyse, by FTIR microspectrometry, the surface of modified membranes to reveal
a) the heterogeneity of the surface modification at the scale of the millimeter, showing that parts of the surface of the membranes are not covered with the copolymer, even after several hours of contacting the membrane to the solution,
b) that the dynamics of the adsorption of the polymer to the surface is slower than expected.
In cases where the surface modification is unevenly distributed over the surface of the membrane, the existence of a protective brush at the surface of the membranes is also probably unevenly distributed. This may explain this discrepancy between the experimental threshold density and the one predicted by the theory.
In the presence of a filtration flux, this picture can be significantly changed. When comparing adsorption (no flux through) and fouling (in the presence of flux) experiments run with a membrane modified with a given co-polymer, we find that the optimum grafting density in terms of resistance to fouling is larger than the grafting density which gives a high resistance to adsorption.
This can be analysed by accounting for the contribution of the flux to the compression of the brush. By combining the transport equations at the surface of the membrane to the Alexander model, one can show that this contribution can be significant, especially for short polymer length. A higher flux would then require a higher grafting density.

Figure 1: Depending on the fouling molecule/particle radius and on the density of hydrophilic polymers grafted at the surface of a membrane, a molecule/particle can approach the surface either by insertion into the brush formed by the polymers or by compressing the brush. The "resistance" to the compression (V/kT) requires higher grafting densities for small molecules/particles. If the grafting density is lower than a threshold value (around 0.5% here) the grafted polymers are not expected to form a brush but may mask part of the surface of the membrane and reduce adsorption.

Some references:

[1] Surface modification of ultrafiltration membranes by low temperature plasma, II. Graft polymerization onto polyacrylonitrile and polysulfone, M. Ulbricht, G. Belfort, J Membr Sci, 111 (1996) 193-215

[2] S. Alexander, Adsorption of chain colloids with a polar head : a scaling description, J Physique 38(1977) 983-987

[3] S. I. Jeon, J. H. Lee, J. D. Andrade, P. G. de Gennes, Protein-Surface Interactions in the Presence of Polyethylene Oxide, J Coll Interface Sci 142(1991)149-158.