Removal of Transparent Exopolymer Particles (TEP) in integrated membrane systems

Integrated membrane systems (IMS) had been successful in controlling particulate fouling in reverseosmosis (RO) treatment. However, it fell short in dealing with a more persistent problem of organic andbiological fouling. The discovery of a formerly unknown but abundant form of polysaccharide calledtransparent exopolymer particles (TEP), led to a better understanding of the role of polysaccharides inbiofouling of RO systems. The main objectives of this research were: (i) to develop a suitable method tomeasure TEP and its colloidal fraction in fresh and sea water; (ii) assess TEP removal by pre-treatment;and (iii) employ TEP as indicator of biofouling potential of RO feedwater in IMS.Modifications on the spectrophotometric method (Passow and Alldredge, 1995b) were proposed,evaluated and verified based on its applications to fresh and marine water samples. The majormodification was the introduction of smaller pore size filters (0.4ìm) and colloidal (0.05-0.4ìm) TEP. The TEP retained on the filters were stained withAlcian Blue (dye specific for acid polysaccharides), dissolved in acid and then absorbance at 787nm weremeasured. Based on the absorbance, TEP was semi-quantified by relating it with the absorbance of astandard. A commercially available polysaccharide (Xanthan gum) was used to standardize the weightof polysaccharides binding with Alcian Blue. TOC measurements were employed to estimate the dryweight of Xanthan, relating TOC removal by filtration to the amount of Xanthan (ìg) retained on thefilter. TEP concentrations were expressed in terms of milligrams Xanthan equivalent per liter (mg Xeq.L-1).Sample measurements showed significant TEP variations for different types of water. Total TEPconcentrations ranged from 1.13 mg Xeq.L-1 in canal water to 8.12 mg Xeq.L-1 in seawater, with highconcentrations recorded during warm seasons. Sequential filtration of water samples revealed thatcolloidal TEP (0.05-0.40ìm) was more abundant than particulate TEP (>0.40 ìm), highlighting theimportance of measuring the colloidal/dissolved TEP fraction.TEP measurements performed on 8 IMS plants recorded TEP removal of 80% by coagulationflocculation-sedimentation, 42-81% by ultrafiltration, 57-70% by in-line coagulation (FeCl3) + UF, 48-80%by microfiltration, 34-71% by coagulation (FeCl3) + rapid sand filtration and 53% by rapid sand filtration.The RO concentrate to feed ratio were analysed to predict TEP accumulation or depletion in the ROsystem and relate it to the extent of biofouling based on the RO cleaning frequency of 5 IMS plants. HighTEP accumulation coincided with high cleaning frequency for 3 plants (A, B, C) while it was not the casefor 2 plants (D and E). Bioufouling limitations for Plant D were attributed to nightly chlorine disinfectionin the MF and/or possible osmotic shock and osmotic backflush cause by the recycling of RO permeate,while the good performance of Plant E was probably due to the fact that it was running with a newlyinstalled RO system (3 months).LC-OCD results for Plant B showed that only 50% of biopolymers in the RO feedwater wererecovered in the concentrate as compared to >70% recovery of other NOM fractions, an indication thatbiopolymer accumulation likely occurred in the RO system. Biopolymer analysis revealed that more than95% of proteins were removed by ultrafiltration while just 60% removal of polysaccharides, suggestingthat the latter is the likely cause of biopolymer fouling in the RO system downstream. Comparison of theLC-OCD polysaccharide and colloidal TEP results in the RO system of 4 IMS plants showed closecorrelation for Plants B, D and E (±13%) but not for Plant A (± 96%). The latter could be a result of the sizelimitations of the two methods (proposed TEP method>0.05 ìm; LC-OCD< 0.45 ìm).The specific deposition rate (SDR) of TEP was determined on 5 IMS based on mass balancecomputations in the RO system. SDR of TEP ranged from -2.6ìg Xeq/m2h in April to +22.9ìg Xeq/m2h inMay. However, some of the SDR results did not coincided with the hypothesis that the higher the SDR,the higher the extent of biofouling and thus frequent RO cleaning. This was attributed to flowassumptions, which probably did not matched with the actual conditions during sampling.Improvements of data inputs for future studies are necessary to make SDR an efffective diagnosis tool offouling in RO systems