Quantifying osmotic membrane fouling to enable comparisons across diverse processes

Abstract

In this study, a method of in situ membrane fouling quantification is developed that enables comparisons of foulant accumulation between desalination processes with different membranes, driving forces, and feed solutions. Unlike the conventional metric of flux decline, which measures the response of a process to fouling, the proposed method quantifies the foulant accumulation. Foulant accumulation is parameterized by two variables, cake structural parameter and hydraulic diameter, that are calculated from flux measurements using a model for salt and water transport through fouled reverse osmosis (RO) and forward osmosis (FO) membranes, including dispersive mass transfer in the FO membrane support layer. Model results show that pressure declines through the foulant layer and can, in FO, reach negative absolute values at the membrane. Experimental alginate gel fouling rates are measured within a range of feed ionic compositions where cake hydraulic resistance is negligible. Using both flux decline and cake structural parameter as metrics, the effect of feed salinity on RO fouling is tested and RO is compared to FO. When RO is fouled with alginate, feed salinity and membrane permeability affect flux decline but not foulant accumulation rate. Between FO and RO, the initial rates of foulant accumulation are similar; however, FO exhibits slower flux decline, which causes greater foulant accumulation over time. The new methodology enables meaningful quantification and comparison of fouling rates with the aim of improving fundamental understanding of fouling processes.

In situ visualization of organic fouling and cleaning mechanisms in reverse osmosis and forward osmosis

“…The 95% confidence interval for flux decline for the present experimental apparatus operating at the same initial flux range is ±0.036 in RO and ±0.054 in FO [9].…”

“…Foulant-free predictions are made using the foulant-free models for RO and FO in [9], which are comparable to standard models except for the addition in [9] of dispersive salt transport in the FO support layer.…”

“…However, permeability varies between membrane samples, as evidenced by the different pressures needed to achieve the same initial flux (39 barg for the trial with dye and 27.3 barg for the trial without). Membrane permeability has been shown by the layered membrane model developed in our previous report [9] to affect flux decline for a given amount of foulant accumulation. In both cases, cleaning at atmospheric pressure with high cross- flow velocity (25 cm/s) for one hour partially restores the flux, but it is difficult to claim that the cleaning step was equally effective in the two cases.…”

“…Pressure, temperature, and flow rate are controlled while permeate mass (or draw mass, in FO) is recorded, as detailed in our previous report [9]. A camera records fouling and cleaning through a window in the membrane module.…”

“…Membranes are equilibrated and compacted as detailed in our previous report [9] cleaning. The camera is manually focused on the membrane at the beginning of each experiment.…”