Fouling of nanofiltration membranes in full- and bench-scale systems treating groundwater containing silica

• Fouling at bench- and full-scale by a silica-containing groundwater was compared.
• First- and second-stage full-scale foulant layers had physico-chemical differences.
• Efficiency of chemical cleaning was similar for both full-scale foulant layers.
• Bench-scale foulant layers were not representative of full-scale foulant layers.
• RBS was used for elemental and structural characterization of the foulant layers.

To better understand full-scale fouling of nanofiltration and reverse osmosis membranes, it is common to autopsy a membrane element from a specific location in the full-scale system of interest, or conduct fouling tests at bench and pilot scale. However, several factors that may impact fouling at full scale vary along full-scale membrane systems and are difficult to recreate at bench and pilot scale. Thus, we performed an in-depth comparison of fouling in membranes at the lead and tail ends of a full-scale nanofiltration system treating a groundwater containing silica, as well as in a membrane fouled at bench scale with concentrate water from the full-scale system. We characterized the elemental composition, chemistry and structure of the foulant layers, and water permeability recovery upon chemical cleaning. Results showed that while there were apparent physico-chemical differences between the foulant layers of membranes fouled at the lead and tail ends of the full-scale system, water permeability recovery upon cleaning was similar for the two membranes. Bench-scale fouling tests produced foulant layers that were different in composition from those generated at full scale and led to the incorrect conclusion in terms of which was the optimum cleaning solution at full-scale. The dissimilarities between bench and full scale results were likely caused by the various challenges in reproducing full-scale conditions at bench-scale, particularly the inability to maintain anoxic conditions and use a one-pass filtration scheme. Thus, to determine the optimum cleaning solutions for the full-scale system, it was preferable to conduct bench-scale cleaning tests on a membrane fouled at full scale than to attempt recreating fouling at bench scale; this finding may hold true for other anoxic waters. In this study, we also introduce Rutherford backscattering spectrometry (RBS) for elemental and structural characterization of foulant layers. Our results showed that RBS is able to depth profile the elemental composition of the first few microns from the surface of foulant layers, which is information that the more commonly used X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) spectroscopy analyses cannot provide.

graphical-highlightsFigure optionsDownload high-quality image (191 K)Download as PowerPoint slide